sensemaking 0.16.0 → 0.17.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +1 -1
- package/dist/cjs/chunk/extract.d.cts +2 -0
- package/dist/cjs/chunk/extract.d.ts +2 -0
- package/dist/cjs/chunk/extract.js +139 -0
- package/dist/cjs/chunk/extract.js.map +1 -0
- package/dist/cjs/chunk/group.d.cts +4 -0
- package/dist/cjs/chunk/group.d.ts +4 -0
- package/dist/cjs/chunk/group.js +489 -0
- package/dist/cjs/chunk/group.js.map +1 -0
- package/dist/cjs/chunk/index.d.cts +8 -0
- package/dist/cjs/chunk/index.d.ts +8 -0
- package/dist/cjs/chunk/index.js +44 -0
- package/dist/cjs/chunk/index.js.map +1 -0
- package/dist/cjs/chunk/parse.d.cts +2 -0
- package/dist/cjs/chunk/parse.d.ts +2 -0
- package/dist/cjs/chunk/parse.js +71 -0
- package/dist/cjs/chunk/parse.js.map +1 -0
- package/dist/cjs/chunk/types.d.cts +19 -0
- package/dist/cjs/chunk/types.d.ts +19 -0
- package/dist/cjs/chunk/types.js +5 -0
- package/dist/cjs/chunk/types.js.map +1 -0
- package/dist/cjs/chunk/version.d.cts +1 -0
- package/dist/cjs/chunk/version.d.ts +1 -0
- package/dist/cjs/chunk/version.js +14 -0
- package/dist/cjs/chunk/version.js.map +1 -0
- package/dist/cjs/cli/map.js +1 -1
- package/dist/cjs/cli/map.js.map +1 -1
- package/dist/cjs/cli/named.js +1 -1
- package/dist/cjs/cli/named.js.map +1 -1
- package/dist/cjs/cli/path.js +2 -2
- package/dist/cjs/cli/path.js.map +1 -1
- package/dist/cjs/cli/peek.js +1 -1
- package/dist/cjs/cli/peek.js.map +1 -1
- package/dist/cjs/cli/related.js +1 -1
- package/dist/cjs/cli/related.js.map +1 -1
- package/dist/cjs/cli/search.js +1 -1
- package/dist/cjs/cli/search.js.map +1 -1
- package/dist/cjs/cli/shared.d.cts +1 -1
- package/dist/cjs/cli/shared.d.ts +1 -1
- package/dist/cjs/cli/shared.js +3 -3
- package/dist/cjs/cli/shared.js.map +1 -1
- package/dist/cjs/cli/status.js +1 -1
- package/dist/cjs/cli/status.js.map +1 -1
- package/dist/cjs/commands/map.d.cts +1 -1
- package/dist/cjs/commands/map.d.ts +1 -1
- package/dist/cjs/commands/map.js.map +1 -1
- package/dist/cjs/commands/peek.d.cts +1 -1
- package/dist/cjs/commands/peek.d.ts +1 -1
- package/dist/cjs/commands/peek.js +13 -6
- package/dist/cjs/commands/peek.js.map +1 -1
- package/dist/cjs/commands/scope.js +1 -1
- package/dist/cjs/commands/scope.js.map +1 -1
- package/dist/cjs/commands/search.d.cts +1 -1
- package/dist/cjs/commands/search.d.ts +1 -1
- package/dist/cjs/commands/search.js +3 -3
- package/dist/cjs/commands/search.js.map +1 -1
- package/dist/cjs/commands/signals.js +1 -1
- package/dist/cjs/commands/signals.js.map +1 -1
- package/dist/cjs/config/access.d.cts +3 -1
- package/dist/cjs/config/access.d.ts +3 -1
- package/dist/cjs/config/access.js +11 -11
- package/dist/cjs/config/access.js.map +1 -1
- package/dist/cjs/config/index.d.cts +1 -0
- package/dist/cjs/config/index.d.ts +1 -0
- package/dist/cjs/config/index.js +4 -0
- package/dist/cjs/config/index.js.map +1 -1
- package/dist/cjs/config/types.d.cts +1 -0
- package/dist/cjs/config/types.d.ts +1 -0
- package/dist/cjs/config/types.js.map +1 -1
- package/dist/cjs/config/validate.d.cts +1 -0
- package/dist/cjs/config/validate.d.ts +1 -0
- package/dist/cjs/config/validate.js +8 -1
- package/dist/cjs/config/validate.js.map +1 -1
- package/dist/cjs/db/open.d.cts +1 -1
- package/dist/cjs/db/open.d.ts +1 -1
- package/dist/cjs/db/open.js +4 -4
- package/dist/cjs/db/open.js.map +1 -1
- package/dist/cjs/db/reconcile.js +8 -8
- package/dist/cjs/db/reconcile.js.map +1 -1
- package/dist/cjs/{sql-functions.js → db/sql-functions.js} +1 -1
- package/dist/cjs/db/sql-functions.js.map +1 -0
- package/dist/cjs/embed/langfit.js +45 -85
- package/dist/cjs/embed/langfit.js.map +1 -1
- package/dist/cjs/embed/languages.js.map +1 -1
- package/dist/cjs/embed/query.js +6 -6
- package/dist/cjs/embed/query.js.map +1 -1
- package/dist/cjs/embed/static.js +3 -40
- package/dist/cjs/embed/static.js.map +1 -1
- package/dist/cjs/features/embed.d.cts +2 -5
- package/dist/cjs/features/embed.d.ts +2 -5
- package/dist/cjs/features/embed.js +36 -58
- package/dist/cjs/features/embed.js.map +1 -1
- package/dist/cjs/{fences.js → features/fences.js} +1 -1
- package/dist/cjs/features/fences.js.map +1 -0
- package/dist/cjs/features/links.js +1 -1
- package/dist/cjs/features/links.js.map +1 -1
- package/dist/cjs/features/rank.js +1 -1
- package/dist/cjs/features/rank.js.map +1 -1
- package/dist/cjs/features/sections.js +4 -3
- package/dist/cjs/features/sections.js.map +1 -1
- package/dist/cjs/features/tags.js +1 -1
- package/dist/cjs/features/tags.js.map +1 -1
- package/dist/cjs/features/types.d.cts +4 -2
- package/dist/cjs/features/types.d.ts +4 -2
- package/dist/cjs/graph/graph.js.map +1 -0
- package/dist/cjs/graph/traverse.js.map +1 -0
- package/dist/cjs/index.d.cts +4 -4
- package/dist/cjs/index.d.ts +4 -4
- package/dist/cjs/index.js +4 -1
- package/dist/cjs/index.js.map +1 -1
- package/dist/cjs/output/column-hint.js.map +1 -0
- package/dist/cjs/output/output.js.map +1 -0
- package/dist/cjs/output/progress.js.map +1 -0
- package/dist/cjs/{search-error.js → output/search-error.js} +1 -1
- package/dist/cjs/output/search-error.js.map +1 -0
- package/dist/cjs/scan/frontmatter.d.cts +13 -0
- package/dist/cjs/scan/frontmatter.d.ts +13 -0
- package/dist/cjs/scan/frontmatter.js +229 -0
- package/dist/cjs/scan/frontmatter.js.map +1 -0
- package/dist/cjs/scan/index.d.cts +25 -0
- package/dist/cjs/scan/index.d.ts +25 -0
- package/dist/cjs/scan/index.js +107 -0
- package/dist/cjs/scan/index.js.map +1 -0
- package/dist/cjs/scan/list.d.cts +12 -0
- package/dist/cjs/scan/list.d.ts +12 -0
- package/dist/cjs/scan/list.js +154 -0
- package/dist/cjs/scan/list.js.map +1 -0
- package/dist/cjs/text/segment.d.cts +3 -0
- package/dist/cjs/text/segment.d.ts +3 -0
- package/dist/cjs/{segment.js → text/segment.js} +17 -14
- package/dist/cjs/text/segment.js.map +1 -0
- package/dist/cjs/text/strip.d.cts +3 -0
- package/dist/cjs/text/strip.d.ts +3 -0
- package/dist/cjs/text/strip.js +32 -0
- package/dist/cjs/text/strip.js.map +1 -0
- package/dist/esm/chunk/extract.d.ts +2 -0
- package/dist/esm/chunk/extract.js +114 -0
- package/dist/esm/chunk/extract.js.map +1 -0
- package/dist/esm/chunk/group.d.ts +4 -0
- package/dist/esm/chunk/group.js +253 -0
- package/dist/esm/chunk/group.js.map +1 -0
- package/dist/esm/chunk/index.d.ts +8 -0
- package/dist/esm/chunk/index.js +16 -0
- package/dist/esm/chunk/index.js.map +1 -0
- package/dist/esm/chunk/parse.d.ts +2 -0
- package/dist/esm/chunk/parse.js +63 -0
- package/dist/esm/chunk/parse.js.map +1 -0
- package/dist/esm/chunk/types.d.ts +19 -0
- package/dist/esm/chunk/types.js +2 -0
- package/dist/esm/chunk/types.js.map +1 -0
- package/dist/esm/chunk/version.d.ts +1 -0
- package/dist/esm/chunk/version.js +3 -0
- package/dist/esm/chunk/version.js.map +1 -0
- package/dist/esm/cli/map.js +1 -1
- package/dist/esm/cli/map.js.map +1 -1
- package/dist/esm/cli/named.js +1 -1
- package/dist/esm/cli/named.js.map +1 -1
- package/dist/esm/cli/path.js +2 -2
- package/dist/esm/cli/path.js.map +1 -1
- package/dist/esm/cli/peek.js +1 -1
- package/dist/esm/cli/peek.js.map +1 -1
- package/dist/esm/cli/related.js +1 -1
- package/dist/esm/cli/related.js.map +1 -1
- package/dist/esm/cli/search.js +1 -1
- package/dist/esm/cli/search.js.map +1 -1
- package/dist/esm/cli/shared.d.ts +1 -1
- package/dist/esm/cli/shared.js +3 -3
- package/dist/esm/cli/shared.js.map +1 -1
- package/dist/esm/cli/status.js +1 -1
- package/dist/esm/cli/status.js.map +1 -1
- package/dist/esm/commands/map.d.ts +1 -1
- package/dist/esm/commands/map.js.map +1 -1
- package/dist/esm/commands/peek.d.ts +1 -1
- package/dist/esm/commands/peek.js +8 -1
- package/dist/esm/commands/peek.js.map +1 -1
- package/dist/esm/commands/scope.js +1 -1
- package/dist/esm/commands/scope.js.map +1 -1
- package/dist/esm/commands/search.d.ts +1 -1
- package/dist/esm/commands/search.js +3 -3
- package/dist/esm/commands/search.js.map +1 -1
- package/dist/esm/commands/signals.js +1 -1
- package/dist/esm/commands/signals.js.map +1 -1
- package/dist/esm/config/access.d.ts +3 -1
- package/dist/esm/config/access.js +9 -9
- package/dist/esm/config/access.js.map +1 -1
- package/dist/esm/config/index.d.ts +1 -0
- package/dist/esm/config/index.js +1 -0
- package/dist/esm/config/index.js.map +1 -1
- package/dist/esm/config/types.d.ts +1 -0
- package/dist/esm/config/types.js.map +1 -1
- package/dist/esm/config/validate.d.ts +1 -0
- package/dist/esm/config/validate.js +8 -2
- package/dist/esm/config/validate.js.map +1 -1
- package/dist/esm/db/open.d.ts +1 -1
- package/dist/esm/db/open.js +5 -5
- package/dist/esm/db/open.js.map +1 -1
- package/dist/esm/db/reconcile.js +4 -4
- package/dist/esm/db/reconcile.js.map +1 -1
- package/dist/esm/{sql-functions.js → db/sql-functions.js} +1 -1
- package/dist/esm/db/sql-functions.js.map +1 -0
- package/dist/esm/embed/langfit.js +3 -3
- package/dist/esm/embed/langfit.js.map +1 -1
- package/dist/esm/embed/languages.js.map +1 -1
- package/dist/esm/embed/query.js +5 -5
- package/dist/esm/embed/query.js.map +1 -1
- package/dist/esm/embed/static.js +1 -2
- package/dist/esm/embed/static.js.map +1 -1
- package/dist/esm/features/embed.d.ts +2 -5
- package/dist/esm/features/embed.js +34 -32
- package/dist/esm/features/embed.js.map +1 -1
- package/dist/esm/{fences.js → features/fences.js} +1 -1
- package/dist/esm/features/fences.js.map +1 -0
- package/dist/esm/features/links.js +1 -1
- package/dist/esm/features/links.js.map +1 -1
- package/dist/esm/features/rank.js +1 -1
- package/dist/esm/features/rank.js.map +1 -1
- package/dist/esm/features/sections.js +4 -3
- package/dist/esm/features/sections.js.map +1 -1
- package/dist/esm/features/tags.js +1 -1
- package/dist/esm/features/tags.js.map +1 -1
- package/dist/esm/features/types.d.ts +4 -2
- package/dist/esm/features/types.js.map +1 -1
- package/dist/esm/graph/graph.js.map +1 -0
- package/dist/esm/graph/traverse.js.map +1 -0
- package/dist/esm/index.d.ts +4 -4
- package/dist/esm/index.js +2 -2
- package/dist/esm/index.js.map +1 -1
- package/dist/esm/output/column-hint.js.map +1 -0
- package/dist/esm/output/output.js.map +1 -0
- package/dist/esm/output/progress.js.map +1 -0
- package/dist/esm/{search-error.js → output/search-error.js} +1 -1
- package/dist/esm/output/search-error.js.map +1 -0
- package/dist/esm/scan/frontmatter.d.ts +13 -0
- package/dist/esm/{scan.js → scan/frontmatter.js} +4 -123
- package/dist/esm/scan/frontmatter.js.map +1 -0
- package/dist/esm/scan/index.d.ts +25 -0
- package/dist/esm/scan/index.js +58 -0
- package/dist/esm/scan/index.js.map +1 -0
- package/dist/esm/scan/list.d.ts +12 -0
- package/dist/esm/scan/list.js +58 -0
- package/dist/esm/scan/list.js.map +1 -0
- package/dist/esm/text/segment.d.ts +3 -0
- package/dist/esm/{segment.js → text/segment.js} +3 -3
- package/dist/esm/text/segment.js.map +1 -0
- package/dist/esm/text/strip.d.ts +3 -0
- package/dist/esm/text/strip.js +12 -0
- package/dist/esm/text/strip.js.map +1 -0
- package/package.json +15 -2
- package/schema.json +8 -1
- package/skills/sense/SKILL.md +1 -1
- package/skills/sense-setup/SKILL.md +3 -3
- package/dist/cjs/column-hint.js.map +0 -1
- package/dist/cjs/fences.js.map +0 -1
- package/dist/cjs/graph.js.map +0 -1
- package/dist/cjs/output.js.map +0 -1
- package/dist/cjs/progress.js.map +0 -1
- package/dist/cjs/scan.d.cts +0 -35
- package/dist/cjs/scan.d.ts +0 -35
- package/dist/cjs/scan.js +0 -436
- package/dist/cjs/scan.js.map +0 -1
- package/dist/cjs/search-error.js.map +0 -1
- package/dist/cjs/segment.d.cts +0 -2
- package/dist/cjs/segment.d.ts +0 -2
- package/dist/cjs/segment.js.map +0 -1
- package/dist/cjs/sql-functions.js.map +0 -1
- package/dist/cjs/traverse.js.map +0 -1
- package/dist/esm/column-hint.js.map +0 -1
- package/dist/esm/fences.js.map +0 -1
- package/dist/esm/graph.js.map +0 -1
- package/dist/esm/output.js.map +0 -1
- package/dist/esm/progress.js.map +0 -1
- package/dist/esm/scan.d.ts +0 -35
- package/dist/esm/scan.js.map +0 -1
- package/dist/esm/search-error.js.map +0 -1
- package/dist/esm/segment.d.ts +0 -2
- package/dist/esm/segment.js.map +0 -1
- package/dist/esm/sql-functions.js.map +0 -1
- package/dist/esm/traverse.js.map +0 -1
- /package/dist/cjs/{sql-functions.d.cts → db/sql-functions.d.cts} +0 -0
- /package/dist/cjs/{sql-functions.d.ts → db/sql-functions.d.ts} +0 -0
- /package/dist/cjs/{fences.d.cts → features/fences.d.cts} +0 -0
- /package/dist/cjs/{fences.d.ts → features/fences.d.ts} +0 -0
- /package/dist/cjs/{graph.d.cts → graph/graph.d.cts} +0 -0
- /package/dist/cjs/{graph.d.ts → graph/graph.d.ts} +0 -0
- /package/dist/cjs/{graph.js → graph/graph.js} +0 -0
- /package/dist/cjs/{traverse.d.cts → graph/traverse.d.cts} +0 -0
- /package/dist/cjs/{traverse.d.ts → graph/traverse.d.ts} +0 -0
- /package/dist/cjs/{traverse.js → graph/traverse.js} +0 -0
- /package/dist/cjs/{column-hint.d.cts → output/column-hint.d.cts} +0 -0
- /package/dist/cjs/{column-hint.d.ts → output/column-hint.d.ts} +0 -0
- /package/dist/cjs/{column-hint.js → output/column-hint.js} +0 -0
- /package/dist/cjs/{output.d.cts → output/output.d.cts} +0 -0
- /package/dist/cjs/{output.d.ts → output/output.d.ts} +0 -0
- /package/dist/cjs/{output.js → output/output.js} +0 -0
- /package/dist/cjs/{progress.d.cts → output/progress.d.cts} +0 -0
- /package/dist/cjs/{progress.d.ts → output/progress.d.ts} +0 -0
- /package/dist/cjs/{progress.js → output/progress.js} +0 -0
- /package/dist/cjs/{search-error.d.cts → output/search-error.d.cts} +0 -0
- /package/dist/cjs/{search-error.d.ts → output/search-error.d.ts} +0 -0
- /package/dist/esm/{sql-functions.d.ts → db/sql-functions.d.ts} +0 -0
- /package/dist/esm/{fences.d.ts → features/fences.d.ts} +0 -0
- /package/dist/esm/{graph.d.ts → graph/graph.d.ts} +0 -0
- /package/dist/esm/{graph.js → graph/graph.js} +0 -0
- /package/dist/esm/{traverse.d.ts → graph/traverse.d.ts} +0 -0
- /package/dist/esm/{traverse.js → graph/traverse.js} +0 -0
- /package/dist/esm/{column-hint.d.ts → output/column-hint.d.ts} +0 -0
- /package/dist/esm/{column-hint.js → output/column-hint.js} +0 -0
- /package/dist/esm/{output.d.ts → output/output.d.ts} +0 -0
- /package/dist/esm/{output.js → output/output.js} +0 -0
- /package/dist/esm/{progress.d.ts → output/progress.d.ts} +0 -0
- /package/dist/esm/{progress.js → output/progress.js} +0 -0
- /package/dist/esm/{search-error.d.ts → output/search-error.d.ts} +0 -0
package/dist/esm/cli/shared.d.ts
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
import type { ParseArgsOptionsConfig } from 'node:util';
|
|
2
2
|
import type { ResolvedConfig, SearchOverrides } from '../config/index.js';
|
|
3
3
|
import type { OpenResult } from '../db/index.js';
|
|
4
|
-
import type { RowFormat } from '../output.js';
|
|
4
|
+
import type { RowFormat } from '../output/output.js';
|
|
5
5
|
import type { Ctx } from './types.js';
|
|
6
6
|
export declare const FORMAT: ParseArgsOptionsConfig;
|
|
7
7
|
export declare const CONFIG: ParseArgsOptionsConfig;
|
package/dist/esm/cli/shared.js
CHANGED
|
@@ -1,9 +1,9 @@
|
|
|
1
1
|
import { parseArgs } from 'node:util';
|
|
2
|
-
import { columnHint } from '../column-hint.js';
|
|
3
2
|
import { resolvePreset } from '../config/index.js';
|
|
4
3
|
import { open } from '../db/index.js';
|
|
5
|
-
import {
|
|
6
|
-
import {
|
|
4
|
+
import { columnHint } from '../output/column-hint.js';
|
|
5
|
+
import { printRowStream } from '../output/output.js';
|
|
6
|
+
import { searchError } from '../output/search-error.js';
|
|
7
7
|
// Spreadable option fragments -- one flag name keeps one meaning across every command's table.
|
|
8
8
|
export const FORMAT = {
|
|
9
9
|
format: {
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/cli/shared.ts"],"sourcesContent":["import type { ParseArgsOptionsConfig } from 'node:util';\nimport { parseArgs } from 'node:util';\nimport { columnHint } from '../column-hint.ts';\nimport type { ResolvedConfig, SearchOverrides } from '../config/index.ts';\nimport { resolvePreset } from '../config/index.ts';\nimport type { OpenResult } from '../db/index.ts';\nimport { open } from '../db/index.ts';\nimport type { Row, RowFormat } from '../output.ts';\nimport { printRowStream } from '../output.ts';\nimport { searchError } from '../search-error.ts';\nimport type { Ctx } from './types.ts';\n\n// Spreadable option fragments -- one flag name keeps one meaning across every command's table.\nexport const FORMAT: ParseArgsOptionsConfig = { format: { type: 'string', default: 'table' } };\nexport const CONFIG: ParseArgsOptionsConfig = { config: { type: 'string' } };\n// The scope vocabulary every scoped command shares: named preset, ad hoc include/exclude\n// globs, a where SQL condition. search adds k on top.\nexport const SCOPE: ParseArgsOptionsConfig = {\n where: { type: 'string' },\n preset: { type: 'string' },\n include: { type: 'string', multiple: true },\n exclude: { type: 'string', multiple: true },\n // Widening, which no other flag can express: --include and --exclude each override their\n // own side of the preset, so neither can drop an exclusion the preset declares.\n 'no-exclude': { type: 'boolean', default: false },\n};\nexport const SEARCH_FLAGS: ParseArgsOptionsConfig = { ...SCOPE, k: { type: 'string' } };\n\ntype Values = Record<string, string | boolean | string[] | undefined>;\n\n// The SCOPE fragment's parsed values as the overrides every scoped command passes down. One\n// place to read them, so a new scope field is added to the fragment and here, not in each\n// command's call.\nexport function scopeOf(values: Values): SearchOverrides {\n return {\n preset: values.preset as string | undefined,\n include: values.include as string[] | undefined,\n exclude: values.exclude as string[] | undefined,\n where: values.where as string | undefined,\n noExclude: values['no-exclude'] === true,\n };\n}\n\n// An unrecognised value used to fall through to `table`, so a typo looked like it worked.\n// parseArgs is strict about flag names; this is the same strictness for the value.\nfunction pickFormat<T extends string>(values: Values, allowed: readonly T[]): T {\n const format = String(values.format);\n if ((allowed as readonly string[]).includes(format)) return format as T;\n console.error(`unknown --format \"${format}\"; expected ${allowed.join(', ')}`);\n process.exit(2);\n}\n\nexport function formatOf(values: Values): 'table' | 'json' {\n return pickFormat(values, ['table', 'json'] as const);\n}\n\n// csv is a row set rendered as rows; map, peek, status, and path render structures instead,\n// and take formatOf.\nexport function rowFormatOf(values: Values): RowFormat {\n return pickFormat(values, ['table', 'json', 'csv'] as const);\n}\n\n// Per-command parseArgs: strict (a foreign flag exits 2), and every table gets --help for free.\nexport function parse(argv: string[], usage: string, options: ParseArgsOptionsConfig): { values: Values; positionals: string[] } {\n let values: Values;\n let positionals: string[];\n try {\n ({ values, positionals } = parseArgs({\n args: argv,\n options: { ...options, help: { type: 'boolean', default: false, short: 'h' } },\n strict: true,\n allowPositionals: true,\n }));\n } catch (err) {\n console.error((err as Error).message);\n console.error(usage);\n process.exit(2);\n }\n if (values.help) {\n console.log(usage);\n process.exit(0);\n }\n return { values, positionals };\n}\n\n// --k must be a positive integer: SQLite reads a bound LIMIT of -1 as \"unlimited\" and 0 as\n// \"nothing\", and parseInt would silently truncate \"5.9\" -- all three are caller mistakes\n// worth a usage error, matching the config-level SavedSearch validation.\nexport function parseK(k: string | undefined, usageError: (message: string) => never): number | undefined {\n if (k === undefined) return undefined;\n const parsed = Number(k);\n if (!Number.isInteger(parsed) || parsed <= 0) usageError(`--k expects a positive integer, got \"${k}\"`);\n return parsed;\n}\n\n// Shared open-query-close envelope for commands that touch the tree.\n\nexport function printWarnings(warnings: string[]): void {\n for (const w of warnings) console.warn(w);\n}\n\nexport async function withDb(ctx: Ctx, configPath: string | undefined, fn: (db: OpenResult['db'], cfg: ResolvedConfig) => void | Promise<void>): Promise<void> {\n const cfg = ctx.resolveConfig(configPath);\n const { db, warnings } = open(cfg);\n printWarnings(warnings);\n try {\n await fn(db, cfg);\n } finally {\n db.close();\n }\n}\n\n// `--preset` on SQL binds the scope rather than applying it: the statement joins `scope`\n// itself. Filtering behind the query's back is not available -- the obvious version, temp views\n// shadowing the base tables, cannot cover `content`, because FTS5 `MATCH` uses the table name\n// as a hidden column and a view has none. A flag that silently scoped three tables of four\n// would look scoped and not be.\nfunction bindScope(db: OpenResult['db'], cfg: ResolvedConfig, preset: string): void {\n const { name } = resolvePreset(cfg, preset); // unknown names throw, listing what is declared\n db.exec('DROP TABLE IF EXISTS temp.scope');\n db.exec('CREATE TEMP TABLE scope (\"path\" TEXT PRIMARY KEY)');\n db.prepare('INSERT INTO temp.scope (\"path\") SELECT \"path\" FROM preset_files WHERE preset = ?').run(name);\n}\n\n// An unbound `?` silently binds NULL, so mismatched param counts fail loudly instead.\nexport function runSql(cfg: ResolvedConfig, sql: string, params: string[], format: RowFormat, label: string, preset?: string): void {\n const placeholderCount = (sql.match(/\\?/g) ?? []).length;\n if (params.length !== placeholderCount) {\n console.error(`${label} expects ${placeholderCount} parameter(s), got ${params.length}`);\n process.exit(2);\n }\n // Naming a preset and never joining it would return the whole index while reading as scoped,\n // which is the one hazard of binding rather than applying. Refuse instead.\n if (preset !== undefined && !/\\bscope\\b/i.test(sql)) {\n console.error(`--preset binds a temporary \"scope\" table of the preset's paths, and ${label} never joins it, so the preset would have no effect`);\n console.error(`add: JOIN scope ON scope.\"path\" = <table>.\"path\"`);\n process.exit(2);\n }\n const { db, warnings } = open(cfg);\n printWarnings(warnings);\n if (preset !== undefined) bindScope(db, cfg, preset);\n // Streamed rather than collected: `sql` is the one caller whose result size is the\n // statement's business, not this package's, so the rows are never held here in bulk.\n // columns() reads the statement's own metadata, so a 0-row csv still has its header.\n // A mid-stream SQLite error (e.g. SQLITE_BUSY outlasting busy_timeout) can still leave\n // partial output; the nonzero exit code is the caller's signal, output completeness is not\n // guaranteed on failure.\n try {\n const statement = db.prepare(sql);\n statement.setReadBigInts(true); // int64 past 2^53 arrives as BigInt instead of throwing at step time\n const columns = statement.columns().map((c) => c.name);\n printRowStream(statement.iterate(...params) as Iterable<Row>, format, columns);\n } catch (err) {\n const hinted = columnHint(db, err as Error); // pragma needs the db still open\n db.close();\n // A saved query or ad-hoc SQL can carry `content MATCH ?` too, so the same FTS5\n // punctuation trap applies -- the bound parameters are the search terms there. SQL\n // without MATCH gets its error verbatim; search advice on a plain typo would mislead.\n if (/\\bMATCH\\b/i.test(sql)) throw searchError(hinted, params.join(' '));\n throw hinted;\n }\n db.close();\n}\n"],"names":["parseArgs","columnHint","resolvePreset","open","printRowStream","searchError","FORMAT","format","type","default","CONFIG","config","SCOPE","where","preset","include","multiple","exclude","SEARCH_FLAGS","k","scopeOf","values","noExclude","pickFormat","allowed","String","includes","console","error","join","process","exit","formatOf","rowFormatOf","parse","argv","usage","options","positionals","args","help","short","strict","allowPositionals","err","message","log","parseK","usageError","undefined","parsed","Number","isInteger","printWarnings","warnings","w","warn","withDb","ctx","configPath","fn","cfg","resolveConfig","db","close","bindScope","name","exec","prepare","run","runSql","sql","params","label","placeholderCount","match","length","test","statement","setReadBigInts","columns","map","c","iterate","hinted"],"mappings":"AACA,SAASA,SAAS,QAAQ,YAAY;AACtC,SAASC,UAAU,QAAQ,oBAAoB;AAE/C,SAASC,aAAa,QAAQ,qBAAqB;AAEnD,SAASC,IAAI,QAAQ,iBAAiB;AAEtC,SAASC,cAAc,QAAQ,eAAe;AAC9C,SAASC,WAAW,QAAQ,qBAAqB;AAGjD,+FAA+F;AAC/F,OAAO,MAAMC,SAAiC;IAAEC,QAAQ;QAAEC,MAAM;QAAUC,SAAS;IAAQ;AAAE,EAAE;AAC/F,OAAO,MAAMC,SAAiC;IAAEC,QAAQ;QAAEH,MAAM;IAAS;AAAE,EAAE;AAC7E,yFAAyF;AACzF,sDAAsD;AACtD,OAAO,MAAMI,QAAgC;IAC3CC,OAAO;QAAEL,MAAM;IAAS;IACxBM,QAAQ;QAAEN,MAAM;IAAS;IACzBO,SAAS;QAAEP,MAAM;QAAUQ,UAAU;IAAK;IAC1CC,SAAS;QAAET,MAAM;QAAUQ,UAAU;IAAK;IAC1C,yFAAyF;IACzF,gFAAgF;IAChF,cAAc;QAAER,MAAM;QAAWC,SAAS;IAAM;AAClD,EAAE;AACF,OAAO,MAAMS,eAAuC;IAAE,GAAGN,KAAK;IAAEO,GAAG;QAAEX,MAAM;IAAS;AAAE,EAAE;AAIxF,4FAA4F;AAC5F,0FAA0F;AAC1F,kBAAkB;AAClB,OAAO,SAASY,QAAQC,MAAc;IACpC,OAAO;QACLP,QAAQO,OAAOP,MAAM;QACrBC,SAASM,OAAON,OAAO;QACvBE,SAASI,OAAOJ,OAAO;QACvBJ,OAAOQ,OAAOR,KAAK;QACnBS,WAAWD,MAAM,CAAC,aAAa,KAAK;IACtC;AACF;AAEA,0FAA0F;AAC1F,mFAAmF;AACnF,SAASE,WAA6BF,MAAc,EAAEG,OAAqB;IACzE,MAAMjB,SAASkB,OAAOJ,OAAOd,MAAM;IACnC,IAAI,AAACiB,QAA8BE,QAAQ,CAACnB,SAAS,OAAOA;IAC5DoB,QAAQC,KAAK,CAAC,CAAC,kBAAkB,EAAErB,OAAO,YAAY,EAAEiB,QAAQK,IAAI,CAAC,OAAO;IAC5EC,QAAQC,IAAI,CAAC;AACf;AAEA,OAAO,SAASC,SAASX,MAAc;IACrC,OAAOE,WAAWF,QAAQ;QAAC;QAAS;KAAO;AAC7C;AAEA,4FAA4F;AAC5F,qBAAqB;AACrB,OAAO,SAASY,YAAYZ,MAAc;IACxC,OAAOE,WAAWF,QAAQ;QAAC;QAAS;QAAQ;KAAM;AACpD;AAEA,gGAAgG;AAChG,OAAO,SAASa,MAAMC,IAAc,EAAEC,KAAa,EAAEC,OAA+B;IAClF,IAAIhB;IACJ,IAAIiB;IACJ,IAAI;QACD,CAAA,EAAEjB,MAAM,EAAEiB,WAAW,EAAE,GAAGtC,UAAU;YACnCuC,MAAMJ;YACNE,SAAS;gBAAE,GAAGA,OAAO;gBAAEG,MAAM;oBAAEhC,MAAM;oBAAWC,SAAS;oBAAOgC,OAAO;gBAAI;YAAE;YAC7EC,QAAQ;YACRC,kBAAkB;QACpB,EAAC;IACH,EAAE,OAAOC,KAAK;QACZjB,QAAQC,KAAK,CAAC,AAACgB,IAAcC,OAAO;QACpClB,QAAQC,KAAK,CAACQ;QACdN,QAAQC,IAAI,CAAC;IACf;IACA,IAAIV,OAAOmB,IAAI,EAAE;QACfb,QAAQmB,GAAG,CAACV;QACZN,QAAQC,IAAI,CAAC;IACf;IACA,OAAO;QAAEV;QAAQiB;IAAY;AAC/B;AAEA,2FAA2F;AAC3F,yFAAyF;AACzF,yEAAyE;AACzE,OAAO,SAASS,OAAO5B,CAAqB,EAAE6B,UAAsC;IAClF,IAAI7B,MAAM8B,WAAW,OAAOA;IAC5B,MAAMC,SAASC,OAAOhC;IACtB,IAAI,CAACgC,OAAOC,SAAS,CAACF,WAAWA,UAAU,GAAGF,WAAW,CAAC,qCAAqC,EAAE7B,EAAE,CAAC,CAAC;IACrG,OAAO+B;AACT;AAEA,qEAAqE;AAErE,OAAO,SAASG,cAAcC,QAAkB;IAC9C,KAAK,MAAMC,KAAKD,SAAU3B,QAAQ6B,IAAI,CAACD;AACzC;AAEA,OAAO,eAAeE,OAAOC,GAAQ,EAAEC,UAA8B,EAAEC,EAAuE;IAC5I,MAAMC,MAAMH,IAAII,aAAa,CAACH;IAC9B,MAAM,EAAEI,EAAE,EAAET,QAAQ,EAAE,GAAGnD,KAAK0D;IAC9BR,cAAcC;IACd,IAAI;QACF,MAAMM,GAAGG,IAAIF;IACf,SAAU;QACRE,GAAGC,KAAK;IACV;AACF;AAEA,yFAAyF;AACzF,gGAAgG;AAChG,8FAA8F;AAC9F,2FAA2F;AAC3F,gCAAgC;AAChC,SAASC,UAAUF,EAAoB,EAAEF,GAAmB,EAAE/C,MAAc;IAC1E,MAAM,EAAEoD,IAAI,EAAE,GAAGhE,cAAc2D,KAAK/C,SAAS,gDAAgD;IAC7FiD,GAAGI,IAAI,CAAC;IACRJ,GAAGI,IAAI,CAAC;IACRJ,GAAGK,OAAO,CAAC,oFAAoFC,GAAG,CAACH;AACrG;AAEA,sFAAsF;AACtF,OAAO,SAASI,OAAOT,GAAmB,EAAEU,GAAW,EAAEC,MAAgB,EAAEjE,MAAiB,EAAEkE,KAAa,EAAE3D,MAAe;QAChGyD;IAA1B,MAAMG,mBAAmB,EAACH,aAAAA,IAAII,KAAK,CAAC,oBAAVJ,wBAAAA,aAAoB,EAAE,EAAEK,MAAM;IACxD,IAAIJ,OAAOI,MAAM,KAAKF,kBAAkB;QACtC/C,QAAQC,KAAK,CAAC,GAAG6C,MAAM,SAAS,EAAEC,iBAAiB,mBAAmB,EAAEF,OAAOI,MAAM,EAAE;QACvF9C,QAAQC,IAAI,CAAC;IACf;IACA,6FAA6F;IAC7F,2EAA2E;IAC3E,IAAIjB,WAAWmC,aAAa,CAAC,aAAa4B,IAAI,CAACN,MAAM;QACnD5C,QAAQC,KAAK,CAAC,CAAC,oEAAoE,EAAE6C,MAAM,mDAAmD,CAAC;QAC/I9C,QAAQC,KAAK,CAAC,CAAC,gDAAgD,CAAC;QAChEE,QAAQC,IAAI,CAAC;IACf;IACA,MAAM,EAAEgC,EAAE,EAAET,QAAQ,EAAE,GAAGnD,KAAK0D;IAC9BR,cAAcC;IACd,IAAIxC,WAAWmC,WAAWgB,UAAUF,IAAIF,KAAK/C;IAC7C,mFAAmF;IACnF,qFAAqF;IACrF,qFAAqF;IACrF,uFAAuF;IACvF,2FAA2F;IAC3F,yBAAyB;IACzB,IAAI;QACF,MAAMgE,YAAYf,GAAGK,OAAO,CAACG;QAC7BO,UAAUC,cAAc,CAAC,OAAO,qEAAqE;QACrG,MAAMC,UAAUF,UAAUE,OAAO,GAAGC,GAAG,CAAC,CAACC,IAAMA,EAAEhB,IAAI;QACrD9D,eAAe0E,UAAUK,OAAO,IAAIX,SAA0BjE,QAAQyE;IACxE,EAAE,OAAOpC,KAAK;QACZ,MAAMwC,SAASnF,WAAW8D,IAAInB,MAAe,iCAAiC;QAC9EmB,GAAGC,KAAK;QACR,gFAAgF;QAChF,mFAAmF;QACnF,sFAAsF;QACtF,IAAI,aAAaa,IAAI,CAACN,MAAM,MAAMlE,YAAY+E,QAAQZ,OAAO3C,IAAI,CAAC;QAClE,MAAMuD;IACR;IACArB,GAAGC,KAAK;AACV"}
|
|
1
|
+
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/cli/shared.ts"],"sourcesContent":["import type { ParseArgsOptionsConfig } from 'node:util';\nimport { parseArgs } from 'node:util';\nimport type { ResolvedConfig, SearchOverrides } from '../config/index.ts';\nimport { resolvePreset } from '../config/index.ts';\nimport type { OpenResult } from '../db/index.ts';\nimport { open } from '../db/index.ts';\nimport { columnHint } from '../output/column-hint.ts';\nimport type { Row, RowFormat } from '../output/output.ts';\nimport { printRowStream } from '../output/output.ts';\nimport { searchError } from '../output/search-error.ts';\nimport type { Ctx } from './types.ts';\n\n// Spreadable option fragments -- one flag name keeps one meaning across every command's table.\nexport const FORMAT: ParseArgsOptionsConfig = { format: { type: 'string', default: 'table' } };\nexport const CONFIG: ParseArgsOptionsConfig = { config: { type: 'string' } };\n// The scope vocabulary every scoped command shares: named preset, ad hoc include/exclude\n// globs, a where SQL condition. search adds k on top.\nexport const SCOPE: ParseArgsOptionsConfig = {\n where: { type: 'string' },\n preset: { type: 'string' },\n include: { type: 'string', multiple: true },\n exclude: { type: 'string', multiple: true },\n // Widening, which no other flag can express: --include and --exclude each override their\n // own side of the preset, so neither can drop an exclusion the preset declares.\n 'no-exclude': { type: 'boolean', default: false },\n};\nexport const SEARCH_FLAGS: ParseArgsOptionsConfig = { ...SCOPE, k: { type: 'string' } };\n\ntype Values = Record<string, string | boolean | string[] | undefined>;\n\n// The SCOPE fragment's parsed values as the overrides every scoped command passes down. One\n// place to read them, so a new scope field is added to the fragment and here, not in each\n// command's call.\nexport function scopeOf(values: Values): SearchOverrides {\n return {\n preset: values.preset as string | undefined,\n include: values.include as string[] | undefined,\n exclude: values.exclude as string[] | undefined,\n where: values.where as string | undefined,\n noExclude: values['no-exclude'] === true,\n };\n}\n\n// An unrecognised value used to fall through to `table`, so a typo looked like it worked.\n// parseArgs is strict about flag names; this is the same strictness for the value.\nfunction pickFormat<T extends string>(values: Values, allowed: readonly T[]): T {\n const format = String(values.format);\n if ((allowed as readonly string[]).includes(format)) return format as T;\n console.error(`unknown --format \"${format}\"; expected ${allowed.join(', ')}`);\n process.exit(2);\n}\n\nexport function formatOf(values: Values): 'table' | 'json' {\n return pickFormat(values, ['table', 'json'] as const);\n}\n\n// csv is a row set rendered as rows; map, peek, status, and path render structures instead,\n// and take formatOf.\nexport function rowFormatOf(values: Values): RowFormat {\n return pickFormat(values, ['table', 'json', 'csv'] as const);\n}\n\n// Per-command parseArgs: strict (a foreign flag exits 2), and every table gets --help for free.\nexport function parse(argv: string[], usage: string, options: ParseArgsOptionsConfig): { values: Values; positionals: string[] } {\n let values: Values;\n let positionals: string[];\n try {\n ({ values, positionals } = parseArgs({\n args: argv,\n options: { ...options, help: { type: 'boolean', default: false, short: 'h' } },\n strict: true,\n allowPositionals: true,\n }));\n } catch (err) {\n console.error((err as Error).message);\n console.error(usage);\n process.exit(2);\n }\n if (values.help) {\n console.log(usage);\n process.exit(0);\n }\n return { values, positionals };\n}\n\n// --k must be a positive integer: SQLite reads a bound LIMIT of -1 as \"unlimited\" and 0 as\n// \"nothing\", and parseInt would silently truncate \"5.9\" -- all three are caller mistakes\n// worth a usage error, matching the config-level SavedSearch validation.\nexport function parseK(k: string | undefined, usageError: (message: string) => never): number | undefined {\n if (k === undefined) return undefined;\n const parsed = Number(k);\n if (!Number.isInteger(parsed) || parsed <= 0) usageError(`--k expects a positive integer, got \"${k}\"`);\n return parsed;\n}\n\n// Shared open-query-close envelope for commands that touch the tree.\n\nexport function printWarnings(warnings: string[]): void {\n for (const w of warnings) console.warn(w);\n}\n\nexport async function withDb(ctx: Ctx, configPath: string | undefined, fn: (db: OpenResult['db'], cfg: ResolvedConfig) => void | Promise<void>): Promise<void> {\n const cfg = ctx.resolveConfig(configPath);\n const { db, warnings } = open(cfg);\n printWarnings(warnings);\n try {\n await fn(db, cfg);\n } finally {\n db.close();\n }\n}\n\n// `--preset` on SQL binds the scope rather than applying it: the statement joins `scope`\n// itself. Filtering behind the query's back is not available -- the obvious version, temp views\n// shadowing the base tables, cannot cover `content`, because FTS5 `MATCH` uses the table name\n// as a hidden column and a view has none. A flag that silently scoped three tables of four\n// would look scoped and not be.\nfunction bindScope(db: OpenResult['db'], cfg: ResolvedConfig, preset: string): void {\n const { name } = resolvePreset(cfg, preset); // unknown names throw, listing what is declared\n db.exec('DROP TABLE IF EXISTS temp.scope');\n db.exec('CREATE TEMP TABLE scope (\"path\" TEXT PRIMARY KEY)');\n db.prepare('INSERT INTO temp.scope (\"path\") SELECT \"path\" FROM preset_files WHERE preset = ?').run(name);\n}\n\n// An unbound `?` silently binds NULL, so mismatched param counts fail loudly instead.\nexport function runSql(cfg: ResolvedConfig, sql: string, params: string[], format: RowFormat, label: string, preset?: string): void {\n const placeholderCount = (sql.match(/\\?/g) ?? []).length;\n if (params.length !== placeholderCount) {\n console.error(`${label} expects ${placeholderCount} parameter(s), got ${params.length}`);\n process.exit(2);\n }\n // Naming a preset and never joining it would return the whole index while reading as scoped,\n // which is the one hazard of binding rather than applying. Refuse instead.\n if (preset !== undefined && !/\\bscope\\b/i.test(sql)) {\n console.error(`--preset binds a temporary \"scope\" table of the preset's paths, and ${label} never joins it, so the preset would have no effect`);\n console.error(`add: JOIN scope ON scope.\"path\" = <table>.\"path\"`);\n process.exit(2);\n }\n const { db, warnings } = open(cfg);\n printWarnings(warnings);\n if (preset !== undefined) bindScope(db, cfg, preset);\n // Streamed rather than collected: `sql` is the one caller whose result size is the\n // statement's business, not this package's, so the rows are never held here in bulk.\n // columns() reads the statement's own metadata, so a 0-row csv still has its header.\n // A mid-stream SQLite error (e.g. SQLITE_BUSY outlasting busy_timeout) can still leave\n // partial output; the nonzero exit code is the caller's signal, output completeness is not\n // guaranteed on failure.\n try {\n const statement = db.prepare(sql);\n statement.setReadBigInts(true); // int64 past 2^53 arrives as BigInt instead of throwing at step time\n const columns = statement.columns().map((c) => c.name);\n printRowStream(statement.iterate(...params) as Iterable<Row>, format, columns);\n } catch (err) {\n const hinted = columnHint(db, err as Error); // pragma needs the db still open\n db.close();\n // A saved query or ad-hoc SQL can carry `content MATCH ?` too, so the same FTS5\n // punctuation trap applies -- the bound parameters are the search terms there. SQL\n // without MATCH gets its error verbatim; search advice on a plain typo would mislead.\n if (/\\bMATCH\\b/i.test(sql)) throw searchError(hinted, params.join(' '));\n throw hinted;\n }\n db.close();\n}\n"],"names":["parseArgs","resolvePreset","open","columnHint","printRowStream","searchError","FORMAT","format","type","default","CONFIG","config","SCOPE","where","preset","include","multiple","exclude","SEARCH_FLAGS","k","scopeOf","values","noExclude","pickFormat","allowed","String","includes","console","error","join","process","exit","formatOf","rowFormatOf","parse","argv","usage","options","positionals","args","help","short","strict","allowPositionals","err","message","log","parseK","usageError","undefined","parsed","Number","isInteger","printWarnings","warnings","w","warn","withDb","ctx","configPath","fn","cfg","resolveConfig","db","close","bindScope","name","exec","prepare","run","runSql","sql","params","label","placeholderCount","match","length","test","statement","setReadBigInts","columns","map","c","iterate","hinted"],"mappings":"AACA,SAASA,SAAS,QAAQ,YAAY;AAEtC,SAASC,aAAa,QAAQ,qBAAqB;AAEnD,SAASC,IAAI,QAAQ,iBAAiB;AACtC,SAASC,UAAU,QAAQ,2BAA2B;AAEtD,SAASC,cAAc,QAAQ,sBAAsB;AACrD,SAASC,WAAW,QAAQ,4BAA4B;AAGxD,+FAA+F;AAC/F,OAAO,MAAMC,SAAiC;IAAEC,QAAQ;QAAEC,MAAM;QAAUC,SAAS;IAAQ;AAAE,EAAE;AAC/F,OAAO,MAAMC,SAAiC;IAAEC,QAAQ;QAAEH,MAAM;IAAS;AAAE,EAAE;AAC7E,yFAAyF;AACzF,sDAAsD;AACtD,OAAO,MAAMI,QAAgC;IAC3CC,OAAO;QAAEL,MAAM;IAAS;IACxBM,QAAQ;QAAEN,MAAM;IAAS;IACzBO,SAAS;QAAEP,MAAM;QAAUQ,UAAU;IAAK;IAC1CC,SAAS;QAAET,MAAM;QAAUQ,UAAU;IAAK;IAC1C,yFAAyF;IACzF,gFAAgF;IAChF,cAAc;QAAER,MAAM;QAAWC,SAAS;IAAM;AAClD,EAAE;AACF,OAAO,MAAMS,eAAuC;IAAE,GAAGN,KAAK;IAAEO,GAAG;QAAEX,MAAM;IAAS;AAAE,EAAE;AAIxF,4FAA4F;AAC5F,0FAA0F;AAC1F,kBAAkB;AAClB,OAAO,SAASY,QAAQC,MAAc;IACpC,OAAO;QACLP,QAAQO,OAAOP,MAAM;QACrBC,SAASM,OAAON,OAAO;QACvBE,SAASI,OAAOJ,OAAO;QACvBJ,OAAOQ,OAAOR,KAAK;QACnBS,WAAWD,MAAM,CAAC,aAAa,KAAK;IACtC;AACF;AAEA,0FAA0F;AAC1F,mFAAmF;AACnF,SAASE,WAA6BF,MAAc,EAAEG,OAAqB;IACzE,MAAMjB,SAASkB,OAAOJ,OAAOd,MAAM;IACnC,IAAI,AAACiB,QAA8BE,QAAQ,CAACnB,SAAS,OAAOA;IAC5DoB,QAAQC,KAAK,CAAC,CAAC,kBAAkB,EAAErB,OAAO,YAAY,EAAEiB,QAAQK,IAAI,CAAC,OAAO;IAC5EC,QAAQC,IAAI,CAAC;AACf;AAEA,OAAO,SAASC,SAASX,MAAc;IACrC,OAAOE,WAAWF,QAAQ;QAAC;QAAS;KAAO;AAC7C;AAEA,4FAA4F;AAC5F,qBAAqB;AACrB,OAAO,SAASY,YAAYZ,MAAc;IACxC,OAAOE,WAAWF,QAAQ;QAAC;QAAS;QAAQ;KAAM;AACpD;AAEA,gGAAgG;AAChG,OAAO,SAASa,MAAMC,IAAc,EAAEC,KAAa,EAAEC,OAA+B;IAClF,IAAIhB;IACJ,IAAIiB;IACJ,IAAI;QACD,CAAA,EAAEjB,MAAM,EAAEiB,WAAW,EAAE,GAAGtC,UAAU;YACnCuC,MAAMJ;YACNE,SAAS;gBAAE,GAAGA,OAAO;gBAAEG,MAAM;oBAAEhC,MAAM;oBAAWC,SAAS;oBAAOgC,OAAO;gBAAI;YAAE;YAC7EC,QAAQ;YACRC,kBAAkB;QACpB,EAAC;IACH,EAAE,OAAOC,KAAK;QACZjB,QAAQC,KAAK,CAAC,AAACgB,IAAcC,OAAO;QACpClB,QAAQC,KAAK,CAACQ;QACdN,QAAQC,IAAI,CAAC;IACf;IACA,IAAIV,OAAOmB,IAAI,EAAE;QACfb,QAAQmB,GAAG,CAACV;QACZN,QAAQC,IAAI,CAAC;IACf;IACA,OAAO;QAAEV;QAAQiB;IAAY;AAC/B;AAEA,2FAA2F;AAC3F,yFAAyF;AACzF,yEAAyE;AACzE,OAAO,SAASS,OAAO5B,CAAqB,EAAE6B,UAAsC;IAClF,IAAI7B,MAAM8B,WAAW,OAAOA;IAC5B,MAAMC,SAASC,OAAOhC;IACtB,IAAI,CAACgC,OAAOC,SAAS,CAACF,WAAWA,UAAU,GAAGF,WAAW,CAAC,qCAAqC,EAAE7B,EAAE,CAAC,CAAC;IACrG,OAAO+B;AACT;AAEA,qEAAqE;AAErE,OAAO,SAASG,cAAcC,QAAkB;IAC9C,KAAK,MAAMC,KAAKD,SAAU3B,QAAQ6B,IAAI,CAACD;AACzC;AAEA,OAAO,eAAeE,OAAOC,GAAQ,EAAEC,UAA8B,EAAEC,EAAuE;IAC5I,MAAMC,MAAMH,IAAII,aAAa,CAACH;IAC9B,MAAM,EAAEI,EAAE,EAAET,QAAQ,EAAE,GAAGpD,KAAK2D;IAC9BR,cAAcC;IACd,IAAI;QACF,MAAMM,GAAGG,IAAIF;IACf,SAAU;QACRE,GAAGC,KAAK;IACV;AACF;AAEA,yFAAyF;AACzF,gGAAgG;AAChG,8FAA8F;AAC9F,2FAA2F;AAC3F,gCAAgC;AAChC,SAASC,UAAUF,EAAoB,EAAEF,GAAmB,EAAE/C,MAAc;IAC1E,MAAM,EAAEoD,IAAI,EAAE,GAAGjE,cAAc4D,KAAK/C,SAAS,gDAAgD;IAC7FiD,GAAGI,IAAI,CAAC;IACRJ,GAAGI,IAAI,CAAC;IACRJ,GAAGK,OAAO,CAAC,oFAAoFC,GAAG,CAACH;AACrG;AAEA,sFAAsF;AACtF,OAAO,SAASI,OAAOT,GAAmB,EAAEU,GAAW,EAAEC,MAAgB,EAAEjE,MAAiB,EAAEkE,KAAa,EAAE3D,MAAe;QAChGyD;IAA1B,MAAMG,mBAAmB,EAACH,aAAAA,IAAII,KAAK,CAAC,oBAAVJ,wBAAAA,aAAoB,EAAE,EAAEK,MAAM;IACxD,IAAIJ,OAAOI,MAAM,KAAKF,kBAAkB;QACtC/C,QAAQC,KAAK,CAAC,GAAG6C,MAAM,SAAS,EAAEC,iBAAiB,mBAAmB,EAAEF,OAAOI,MAAM,EAAE;QACvF9C,QAAQC,IAAI,CAAC;IACf;IACA,6FAA6F;IAC7F,2EAA2E;IAC3E,IAAIjB,WAAWmC,aAAa,CAAC,aAAa4B,IAAI,CAACN,MAAM;QACnD5C,QAAQC,KAAK,CAAC,CAAC,oEAAoE,EAAE6C,MAAM,mDAAmD,CAAC;QAC/I9C,QAAQC,KAAK,CAAC,CAAC,gDAAgD,CAAC;QAChEE,QAAQC,IAAI,CAAC;IACf;IACA,MAAM,EAAEgC,EAAE,EAAET,QAAQ,EAAE,GAAGpD,KAAK2D;IAC9BR,cAAcC;IACd,IAAIxC,WAAWmC,WAAWgB,UAAUF,IAAIF,KAAK/C;IAC7C,mFAAmF;IACnF,qFAAqF;IACrF,qFAAqF;IACrF,uFAAuF;IACvF,2FAA2F;IAC3F,yBAAyB;IACzB,IAAI;QACF,MAAMgE,YAAYf,GAAGK,OAAO,CAACG;QAC7BO,UAAUC,cAAc,CAAC,OAAO,qEAAqE;QACrG,MAAMC,UAAUF,UAAUE,OAAO,GAAGC,GAAG,CAAC,CAACC,IAAMA,EAAEhB,IAAI;QACrD9D,eAAe0E,UAAUK,OAAO,IAAIX,SAA0BjE,QAAQyE;IACxE,EAAE,OAAOpC,KAAK;QACZ,MAAMwC,SAASjF,WAAW4D,IAAInB,MAAe,iCAAiC;QAC9EmB,GAAGC,KAAK;QACR,gFAAgF;QAChF,mFAAmF;QACnF,sFAAsF;QACtF,IAAI,aAAaa,IAAI,CAACN,MAAM,MAAMlE,YAAY+E,QAAQZ,OAAO3C,IAAI,CAAC;QAClE,MAAMuD;IACR;IACArB,GAAGC,KAAK;AACV"}
|
package/dist/esm/cli/status.js
CHANGED
|
@@ -5,7 +5,7 @@ import { docCount, getMeta, open, SCHEMA_VERSION } from '../db/index.js';
|
|
|
5
5
|
import { probeReachable } from '../embed/http.js';
|
|
6
6
|
import { languageDistribution } from '../embed/langfit.js';
|
|
7
7
|
import { isDownloadable, MODEL_FILENAMES, modelDir, modelPresent, readLanguages } from '../embed/store.js';
|
|
8
|
-
import { featuresLine, presetsLines } from '../output.js';
|
|
8
|
+
import { featuresLine, presetsLines } from '../output/output.js';
|
|
9
9
|
import { USAGE } from './index.js';
|
|
10
10
|
import { CONFIG, FORMAT, formatOf, parse, printWarnings } from './shared.js';
|
|
11
11
|
// status names what you cannot read off the config file: locations sense chose, values it
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/cli/status.ts"],"sourcesContent":["import { dirname } from 'node:path';\nimport { presetCoverage } from '../commands/index.ts';\nimport { embedConfig, featureStates, SUPPORTED_CONFIG_VERSION } from '../config/index.ts';\nimport { docCount, getMeta, open, SCHEMA_VERSION } from '../db/index.ts';\nimport { probeReachable } from '../embed/http.ts';\nimport { languageDistribution } from '../embed/langfit.ts';\nimport { isDownloadable, MODEL_FILENAMES, modelDir, modelPresent, readLanguages } from '../embed/store.ts';\nimport { featuresLine, presetsLines } from '../output.ts';\nimport { USAGE } from './index.ts';\nimport { CONFIG, FORMAT, formatOf, parse, printWarnings } from './shared.ts';\nimport type { Command } from './types.ts';\n\n// status names what you cannot read off the config file: locations sense chose, values it\n// derived, state it stored. Config content is not echoed -- the config line says which file to\n// read, and duplicating its globs here would create two places that disagree after an edit.\n// The exception is the embed block's url and key, whose effect is remote and otherwise invisible.\n\n// Vectors are computed on the first semantic query, not at reconcile, so a tree can have a\n// semantic-on preset with nothing embedded yet -- reported here because a fast reconcile on\n// such a tree otherwise reads as \"the feature isn't on\".\nfunction vectorState(db: ReturnType<typeof open>['db']): { embedded: number; pending: number } | null {\n try {\n const row = db.prepare('SELECT COUNT(vector) AS embedded, COUNT(*) - COUNT(vector) AS pending FROM embeddings').get() as { embedded: number; pending: number };\n return row;\n } catch {\n return null;\n }\n}\n\nfunction parseErrorCount(db: ReturnType<typeof open>['db']): number {\n const row = db.prepare(`SELECT COUNT(*) AS n FROM frontmatter WHERE \"_parse_error\" IS NOT NULL`).get() as { n: number };\n return row.n;\n}\n\n// A URL can carry credentials in its userinfo (https://user:pass@host/v1). status output gets\n// pasted into issues and transcripts, so it is always safe to paste.\nexport function redactUrl(url: string): string {\n try {\n const parsed = new URL(url);\n if (!parsed.username && !parsed.password) return url;\n parsed.username = '';\n parsed.password = '';\n return `${parsed.toString().replace(/\\/$/, '')} (credentials redacted)`;\n } catch {\n return url;\n }\n}\n\n// A stopped Ollama/LM Studio/Cohere endpoint prints identically to a running one otherwise; any\n// HTTP response counts as reachable (probeReachable), so this only distinguishes up from down.\nasync function embedReachable(e: { provider: string; url?: string }): Promise<boolean | null> {\n if (e.provider === 'static' || (e.provider === 'openai' && !e.url)) return null;\n const base = (e.url ?? 'https://api.cohere.com').replace(/\\/+$/, '');\n return probeReachable(e.provider === 'openai' ? `${base}/models` : base);\n}\n\nconst status: Command = async (ctx) => {\n const { values } = parse(ctx.argv, `usage: ${ctx.name} ${USAGE.status}`, { ...FORMAT, ...CONFIG });\n const format = formatOf(values);\n const cfg = ctx.resolveConfig(values.config as string | undefined);\n const { db, dbPath, warnings } = open(cfg);\n printWarnings(warnings);\n\n const features = featureStates(cfg);\n const e = embedConfig(cfg);\n const hasModel = e ? modelPresent(cfg) : false;\n const reachable = e ? await embedReachable(e) : null;\n const vectors = e ? vectorState(db) : null;\n const presets = presetCoverage(db, cfg);\n const staticLanguages = e && e.provider === 'static' ? readLanguages(e.model) : undefined;\n const declaredLanguages = staticLanguages && staticLanguages.length > 0 ? staticLanguages : null;\n const detectedLanguages = languageDistribution(db) ?? null;\n const heartbeat = getMeta(db, 'watch_heartbeat');\n // Derived at open() (F): 3x the largest reconcile this cache has ever held its write\n // transaction for, floored at 30s -- read back from the connection rather than\n // recomputed here, so this always reports what open() actually set.\n const busyTimeoutMs = (db.prepare('PRAGMA busy_timeout').get() as { timeout: number }).timeout;\n // The env var holds the token; only its name and whether it is set are ever reported.\n const keyEnv = e?.key ? { name: e.key, set: (process.env[e.key] ?? '') !== '' } : null;\n const result = {\n config: cfg.configPath,\n configVersion: SUPPORTED_CONFIG_VERSION,\n migratedFrom: cfg.migratedFrom ?? null,\n unknownConfigKeys: cfg.unknownKeys ?? null,\n tree: cfg.baseDir,\n cache: dirname(dbPath),\n db: dbPath,\n cacheSchema: SCHEMA_VERSION,\n docs: docCount(db),\n unparseableFrontmatter: parseErrorCount(db),\n features: features.on,\n featuresOff: features.off,\n embed: e ? { provider: e.provider, model: e.model, dir: e.provider === 'static' ? modelDir(e.model) : null, url: e.url ? redactUrl(e.url) : null, reachable, keyEnv, downloaded: hasModel, languages: declaredLanguages, detectedLanguages, ...(vectors ?? {}) } : null,\n presets,\n queries: Object.keys(cfg.queries ?? {}).length,\n watcherPid: getMeta(db, 'watch_pid'),\n watcherHeartbeatSecondsAgo: heartbeat ? Math.round((Date.now() - Date.parse(heartbeat)) / 1000) : null,\n busyTimeoutMs,\n };\n\n if (format === 'json') {\n console.log(JSON.stringify(result, null, 2));\n } else {\n const migrated = cfg.migratedFrom !== undefined ? `, migrated from v${cfg.migratedFrom} on this run` : '';\n console.log(`config: ${result.config} (v${SUPPORTED_CONFIG_VERSION}${migrated})`);\n // A load-time stderr warning has scrolled away by the time anyone looks here.\n if (cfg.unknownKeys) console.log(` unknown keys, ignored: ${cfg.unknownKeys.join(', ')}`);\n console.log(`tree: ${result.tree}`);\n console.log(`cache: ${result.cache} (schema ${SCHEMA_VERSION}; delete this directory to reset)`);\n console.log(`docs: ${result.docs}${result.unparseableFrontmatter > 0 ? ` (${result.unparseableFrontmatter} with unparseable frontmatter: WHERE _parse_error IS NOT NULL)` : ''}`);\n console.log('');\n console.log(featuresLine(features));\n // One authoritative block, off states included: an absent model reads as a degraded search,\n // not as an empty corpus.\n if (!e) console.log('embed: off (no preset asks for vectors)');\n else {\n console.log(`embed: ${e.provider} ${e.model}`);\n if (e.provider === 'static') console.log(` model: ${modelDir(e.model)} (${hasModel ? 'present' : `missing, ${isDownloadable(e.model) ? `fetches automatically at first search, or run \\`${ctx.name} download\\` to prefetch` : `no ${MODEL_FILENAMES} in that directory`}`})`);\n if (declaredLanguages) console.log(` languages: ${declaredLanguages.join(', ')} (declared by the model card)`);\n // Cohere defaults to api.cohere.com with no url in the config; show it anyway so the\n // reachability probe below has something to name.\n if (e.url || reachable !== null) {\n const shown = e.url ? redactUrl(e.url) : 'https://api.cohere.com (default)';\n console.log(` url: ${shown}${reachable === null ? '' : reachable ? ' (reachable)' : ' (unreachable)'}`);\n }\n if (keyEnv) console.log(` key: env ${keyEnv.name} (${keyEnv.set ? 'set' : 'NOT SET'})`);\n if (!hasModel) console.log(' vectors: off until the model is present; search runs on words and links');\n else if (vectors) console.log(` vectors: ${vectors.embedded} embedded, ${vectors.pending} pending (embedded on the first semantic search)`);\n if (detectedLanguages) {\n const total = Object.values(detectedLanguages).reduce((a, b) => a + b, 0);\n const breakdown = Object.entries(detectedLanguages)\n .sort((a, b) => b[1] - a[1])\n .map(([code, n]) => `${code} ${Math.round((n / total) * 100)}%`)\n .join(', ');\n console.log(` detected: ${breakdown} (${total} chunks classified)`);\n }\n }\n console.log('');\n for (const line of presetsLines(presets)) console.log(line);\n console.log('');\n console.log(`queries: ${result.queries} saved (${ctx.name} --list)`);\n const pid = result.watcherPid ? `pid ${result.watcherPid}, ` : '';\n console.log(result.watcherHeartbeatSecondsAgo === null ? 'watcher: none' : `watcher: ${pid}last heartbeat ${result.watcherHeartbeatSecondsAgo}s ago`);\n console.log(`sqlite: busy_timeout ${result.busyTimeoutMs}ms (derived: 3x the largest reconcile this cache has recorded, floored at 30000ms)`);\n }\n db.close();\n};\nexport default status;\n"],"names":["dirname","presetCoverage","embedConfig","featureStates","SUPPORTED_CONFIG_VERSION","docCount","getMeta","open","SCHEMA_VERSION","probeReachable","languageDistribution","isDownloadable","MODEL_FILENAMES","modelDir","modelPresent","readLanguages","featuresLine","presetsLines","USAGE","CONFIG","FORMAT","formatOf","parse","printWarnings","vectorState","db","row","prepare","get","parseErrorCount","n","redactUrl","url","parsed","URL","username","password","toString","replace","embedReachable","e","provider","base","status","ctx","process","cfg","values","argv","name","format","resolveConfig","config","dbPath","warnings","features","hasModel","reachable","vectors","presets","staticLanguages","model","undefined","declaredLanguages","length","detectedLanguages","heartbeat","busyTimeoutMs","timeout","keyEnv","key","set","env","result","configPath","configVersion","migratedFrom","unknownConfigKeys","unknownKeys","tree","baseDir","cache","cacheSchema","docs","unparseableFrontmatter","on","featuresOff","off","embed","dir","downloaded","languages","queries","Object","keys","watcherPid","watcherHeartbeatSecondsAgo","Math","round","Date","now","console","log","JSON","stringify","migrated","join","shown","embedded","pending","total","reduce","a","b","breakdown","entries","sort","map","code","line","pid","close"],"mappings":"AAAA,SAASA,OAAO,QAAQ,YAAY;AACpC,SAASC,cAAc,QAAQ,uBAAuB;AACtD,SAASC,WAAW,EAAEC,aAAa,EAAEC,wBAAwB,QAAQ,qBAAqB;AAC1F,SAASC,QAAQ,EAAEC,OAAO,EAAEC,IAAI,EAAEC,cAAc,QAAQ,iBAAiB;AACzE,SAASC,cAAc,QAAQ,mBAAmB;AAClD,SAASC,oBAAoB,QAAQ,sBAAsB;AAC3D,SAASC,cAAc,EAAEC,eAAe,EAAEC,QAAQ,EAAEC,YAAY,EAAEC,aAAa,QAAQ,oBAAoB;AAC3G,SAASC,YAAY,EAAEC,YAAY,QAAQ,eAAe;AAC1D,SAASC,KAAK,QAAQ,aAAa;AACnC,SAASC,MAAM,EAAEC,MAAM,EAAEC,QAAQ,EAAEC,KAAK,EAAEC,aAAa,QAAQ,cAAc;AAG7E,0FAA0F;AAC1F,+FAA+F;AAC/F,4FAA4F;AAC5F,kGAAkG;AAElG,2FAA2F;AAC3F,4FAA4F;AAC5F,yDAAyD;AACzD,SAASC,YAAYC,EAAiC;IACpD,IAAI;QACF,MAAMC,MAAMD,GAAGE,OAAO,CAAC,yFAAyFC,GAAG;QACnH,OAAOF;IACT,EAAE,OAAM;QACN,OAAO;IACT;AACF;AAEA,SAASG,gBAAgBJ,EAAiC;IACxD,MAAMC,MAAMD,GAAGE,OAAO,CAAC,CAAC,sEAAsE,CAAC,EAAEC,GAAG;IACpG,OAAOF,IAAII,CAAC;AACd;AAEA,8FAA8F;AAC9F,qEAAqE;AACrE,OAAO,SAASC,UAAUC,GAAW;IACnC,IAAI;QACF,MAAMC,SAAS,IAAIC,IAAIF;QACvB,IAAI,CAACC,OAAOE,QAAQ,IAAI,CAACF,OAAOG,QAAQ,EAAE,OAAOJ;QACjDC,OAAOE,QAAQ,GAAG;QAClBF,OAAOG,QAAQ,GAAG;QAClB,OAAO,GAAGH,OAAOI,QAAQ,GAAGC,OAAO,CAAC,OAAO,IAAI,uBAAuB,CAAC;IACzE,EAAE,OAAM;QACN,OAAON;IACT;AACF;AAEA,gGAAgG;AAChG,+FAA+F;AAC/F,eAAeO,eAAeC,CAAqC;QAEnDA;IADd,IAAIA,EAAEC,QAAQ,KAAK,YAAaD,EAAEC,QAAQ,KAAK,YAAY,CAACD,EAAER,GAAG,EAAG,OAAO;IAC3E,MAAMU,OAAO,EAACF,SAAAA,EAAER,GAAG,cAALQ,oBAAAA,SAAS,0BAA0BF,OAAO,CAAC,QAAQ;IACjE,OAAO7B,eAAe+B,EAAEC,QAAQ,KAAK,WAAW,GAAGC,KAAK,OAAO,CAAC,GAAGA;AACrE;AAEA,MAAMC,SAAkB,OAAOC;QAeHlC,uBAOmBmC,oBAI7BC,mBACKA,kBAWEA;IArCvB,MAAM,EAAEC,MAAM,EAAE,GAAGzB,MAAMsB,IAAII,IAAI,EAAE,CAAC,OAAO,EAAEJ,IAAIK,IAAI,CAAC,CAAC,EAAE/B,MAAMyB,MAAM,EAAE,EAAE;QAAE,GAAGvB,MAAM;QAAE,GAAGD,MAAM;IAAC;IAChG,MAAM+B,SAAS7B,SAAS0B;IACxB,MAAMD,MAAMF,IAAIO,aAAa,CAACJ,OAAOK,MAAM;IAC3C,MAAM,EAAE3B,EAAE,EAAE4B,MAAM,EAAEC,QAAQ,EAAE,GAAG/C,KAAKuC;IACtCvB,cAAc+B;IAEd,MAAMC,WAAWpD,cAAc2C;IAC/B,MAAMN,IAAItC,YAAY4C;IACtB,MAAMU,WAAWhB,IAAI1B,aAAagC,OAAO;IACzC,MAAMW,YAAYjB,IAAI,MAAMD,eAAeC,KAAK;IAChD,MAAMkB,UAAUlB,IAAIhB,YAAYC,MAAM;IACtC,MAAMkC,UAAU1D,eAAewB,IAAIqB;IACnC,MAAMc,kBAAkBpB,KAAKA,EAAEC,QAAQ,KAAK,WAAW1B,cAAcyB,EAAEqB,KAAK,IAAIC;IAChF,MAAMC,oBAAoBH,mBAAmBA,gBAAgBI,MAAM,GAAG,IAAIJ,kBAAkB;IAC5F,MAAMK,qBAAoBvD,wBAAAA,qBAAqBe,iBAArBf,mCAAAA,wBAA4B;IACtD,MAAMwD,YAAY5D,QAAQmB,IAAI;IAC9B,qFAAqF;IACrF,+EAA+E;IAC/E,oEAAoE;IACpE,MAAM0C,gBAAgB,AAAC1C,GAAGE,OAAO,CAAC,uBAAuBC,GAAG,GAA2BwC,OAAO;IAC9F,sFAAsF;IACtF,MAAMC,SAAS7B,CAAAA,cAAAA,wBAAAA,EAAG8B,GAAG,IAAG;QAAErB,MAAMT,EAAE8B,GAAG;QAAEC,KAAK,EAAC1B,qBAAAA,QAAQ2B,GAAG,CAAChC,EAAE8B,GAAG,CAAC,cAAlBzB,gCAAAA,qBAAsB,QAAQ;IAAG,IAAI;IAClF,MAAM4B,SAAS;QACbrB,QAAQN,IAAI4B,UAAU;QACtBC,eAAevE;QACfwE,YAAY,GAAE9B,oBAAAA,IAAI8B,YAAY,cAAhB9B,+BAAAA,oBAAoB;QAClC+B,iBAAiB,GAAE/B,mBAAAA,IAAIgC,WAAW,cAAfhC,8BAAAA,mBAAmB;QACtCiC,MAAMjC,IAAIkC,OAAO;QACjBC,OAAOjF,QAAQqD;QACf5B,IAAI4B;QACJ6B,aAAa1E;QACb2E,MAAM9E,SAASoB;QACf2D,wBAAwBvD,gBAAgBJ;QACxC8B,UAAUA,SAAS8B,EAAE;QACrBC,aAAa/B,SAASgC,GAAG;QACzBC,OAAOhD,IAAI;YAAEC,UAAUD,EAAEC,QAAQ;YAAEoB,OAAOrB,EAAEqB,KAAK;YAAE4B,KAAKjD,EAAEC,QAAQ,KAAK,WAAW5B,SAAS2B,EAAEqB,KAAK,IAAI;YAAM7B,KAAKQ,EAAER,GAAG,GAAGD,UAAUS,EAAER,GAAG,IAAI;YAAMyB;YAAWY;YAAQqB,YAAYlC;YAAUmC,WAAW5B;YAAmBE;eAAuBP,oBAAAA,qBAAAA,UAAW,CAAC,CAAhB;QAAmB,IAAI;QACnQC;QACAiC,SAASC,OAAOC,IAAI,EAAChD,eAAAA,IAAI8C,OAAO,cAAX9C,0BAAAA,eAAe,CAAC,GAAGkB,MAAM;QAC9C+B,YAAYzF,QAAQmB,IAAI;QACxBuE,4BAA4B9B,YAAY+B,KAAKC,KAAK,CAAC,AAACC,CAAAA,KAAKC,GAAG,KAAKD,KAAK7E,KAAK,CAAC4C,UAAS,IAAK,QAAQ;QAClGC;IACF;IAEA,IAAIjB,WAAW,QAAQ;QACrBmD,QAAQC,GAAG,CAACC,KAAKC,SAAS,CAAC/B,QAAQ,MAAM;IAC3C,OAAO;QACL,MAAMgC,WAAW3D,IAAI8B,YAAY,KAAKd,YAAY,CAAC,iBAAiB,EAAEhB,IAAI8B,YAAY,CAAC,YAAY,CAAC,GAAG;QACvGyB,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOrB,MAAM,CAAC,GAAG,EAAEhD,2BAA2BqG,SAAS,CAAC,CAAC;QAClF,8EAA8E;QAC9E,IAAI3D,IAAIgC,WAAW,EAAEuB,QAAQC,GAAG,CAAC,CAAC,iCAAiC,EAAExD,IAAIgC,WAAW,CAAC4B,IAAI,CAAC,OAAO;QACjGL,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOM,IAAI,EAAE;QACtCsB,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOQ,KAAK,CAAC,SAAS,EAAEzE,eAAe,iCAAiC,CAAC;QAClG6F,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOU,IAAI,GAAGV,OAAOW,sBAAsB,GAAG,IAAI,CAAC,GAAG,EAAEX,OAAOW,sBAAsB,CAAC,8DAA8D,CAAC,GAAG,IAAI;QACrLiB,QAAQC,GAAG,CAAC;QACZD,QAAQC,GAAG,CAACtF,aAAauC;QACzB,4FAA4F;QAC5F,0BAA0B;QAC1B,IAAI,CAACf,GAAG6D,QAAQC,GAAG,CAAC;aACf;YACHD,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE9D,EAAEC,QAAQ,CAAC,CAAC,EAAED,EAAEqB,KAAK,EAAE;YAChD,IAAIrB,EAAEC,QAAQ,KAAK,UAAU4D,QAAQC,GAAG,CAAC,CAAC,mBAAmB,EAAEzF,SAAS2B,EAAEqB,KAAK,EAAE,EAAE,EAAEL,WAAW,YAAY,CAAC,SAAS,EAAE7C,eAAe6B,EAAEqB,KAAK,IAAI,CAAC,gDAAgD,EAAEjB,IAAIK,IAAI,CAAC,uBAAuB,CAAC,GAAG,CAAC,GAAG,EAAErC,gBAAgB,kBAAkB,CAAC,EAAE,CAAC,CAAC,CAAC;YACvR,IAAImD,mBAAmBsC,QAAQC,GAAG,CAAC,CAAC,qBAAqB,EAAEvC,kBAAkB2C,IAAI,CAAC,MAAM,6BAA6B,CAAC;YACtH,qFAAqF;YACrF,kDAAkD;YAClD,IAAIlE,EAAER,GAAG,IAAIyB,cAAc,MAAM;gBAC/B,MAAMkD,QAAQnE,EAAER,GAAG,GAAGD,UAAUS,EAAER,GAAG,IAAI;gBACzCqE,QAAQC,GAAG,CAAC,CAAC,mBAAmB,EAAEK,QAAQlD,cAAc,OAAO,KAAKA,YAAY,iBAAiB,kBAAkB;YACrH;YACA,IAAIY,QAAQgC,QAAQC,GAAG,CAAC,CAAC,uBAAuB,EAAEjC,OAAOpB,IAAI,CAAC,EAAE,EAAEoB,OAAOE,GAAG,GAAG,QAAQ,UAAU,CAAC,CAAC;YACnG,IAAI,CAACf,UAAU6C,QAAQC,GAAG,CAAC;iBACtB,IAAI5C,SAAS2C,QAAQC,GAAG,CAAC,CAAC,mBAAmB,EAAE5C,QAAQkD,QAAQ,CAAC,WAAW,EAAElD,QAAQmD,OAAO,CAAC,gDAAgD,CAAC;YACnJ,IAAI5C,mBAAmB;gBACrB,MAAM6C,QAAQjB,OAAO9C,MAAM,CAACkB,mBAAmB8C,MAAM,CAAC,CAACC,GAAGC,IAAMD,IAAIC,GAAG;gBACvE,MAAMC,YAAYrB,OAAOsB,OAAO,CAAClD,mBAC9BmD,IAAI,CAAC,CAACJ,GAAGC,IAAMA,CAAC,CAAC,EAAE,GAAGD,CAAC,CAAC,EAAE,EAC1BK,GAAG,CAAC,CAAC,CAACC,MAAMxF,EAAE,GAAK,GAAGwF,KAAK,CAAC,EAAErB,KAAKC,KAAK,CAAC,AAACpE,IAAIgF,QAAS,KAAK,CAAC,CAAC,EAC9DJ,IAAI,CAAC;gBACRL,QAAQC,GAAG,CAAC,CAAC,oBAAoB,EAAEY,UAAU,EAAE,EAAEJ,MAAM,mBAAmB,CAAC;YAC7E;QACF;QACAT,QAAQC,GAAG,CAAC;QACZ,KAAK,MAAMiB,QAAQtG,aAAa0C,SAAU0C,QAAQC,GAAG,CAACiB;QACtDlB,QAAQC,GAAG,CAAC;QACZD,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOmB,OAAO,CAAC,QAAQ,EAAEhD,IAAIK,IAAI,CAAC,QAAQ,CAAC;QACpE,MAAMuE,MAAM/C,OAAOsB,UAAU,GAAG,CAAC,IAAI,EAAEtB,OAAOsB,UAAU,CAAC,EAAE,CAAC,GAAG;QAC/DM,QAAQC,GAAG,CAAC7B,OAAOuB,0BAA0B,KAAK,OAAO,mBAAmB,CAAC,UAAU,EAAEwB,IAAI,eAAe,EAAE/C,OAAOuB,0BAA0B,CAAC,KAAK,CAAC;QACtJK,QAAQC,GAAG,CAAC,CAAC,uBAAuB,EAAE7B,OAAON,aAAa,CAAC,kFAAkF,CAAC;IAChJ;IACA1C,GAAGgG,KAAK;AACV;AACA,eAAe9E,OAAO"}
|
|
1
|
+
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/cli/status.ts"],"sourcesContent":["import { dirname } from 'node:path';\nimport { presetCoverage } from '../commands/index.ts';\nimport { embedConfig, featureStates, SUPPORTED_CONFIG_VERSION } from '../config/index.ts';\nimport { docCount, getMeta, open, SCHEMA_VERSION } from '../db/index.ts';\nimport { probeReachable } from '../embed/http.ts';\nimport { languageDistribution } from '../embed/langfit.ts';\nimport { isDownloadable, MODEL_FILENAMES, modelDir, modelPresent, readLanguages } from '../embed/store.ts';\nimport { featuresLine, presetsLines } from '../output/output.ts';\nimport { USAGE } from './index.ts';\nimport { CONFIG, FORMAT, formatOf, parse, printWarnings } from './shared.ts';\nimport type { Command } from './types.ts';\n\n// status names what you cannot read off the config file: locations sense chose, values it\n// derived, state it stored. Config content is not echoed -- the config line says which file to\n// read, and duplicating its globs here would create two places that disagree after an edit.\n// The exception is the embed block's url and key, whose effect is remote and otherwise invisible.\n\n// Vectors are computed on the first semantic query, not at reconcile, so a tree can have a\n// semantic-on preset with nothing embedded yet -- reported here because a fast reconcile on\n// such a tree otherwise reads as \"the feature isn't on\".\nfunction vectorState(db: ReturnType<typeof open>['db']): { embedded: number; pending: number } | null {\n try {\n const row = db.prepare('SELECT COUNT(vector) AS embedded, COUNT(*) - COUNT(vector) AS pending FROM embeddings').get() as { embedded: number; pending: number };\n return row;\n } catch {\n return null;\n }\n}\n\nfunction parseErrorCount(db: ReturnType<typeof open>['db']): number {\n const row = db.prepare(`SELECT COUNT(*) AS n FROM frontmatter WHERE \"_parse_error\" IS NOT NULL`).get() as { n: number };\n return row.n;\n}\n\n// A URL can carry credentials in its userinfo (https://user:pass@host/v1). status output gets\n// pasted into issues and transcripts, so it is always safe to paste.\nexport function redactUrl(url: string): string {\n try {\n const parsed = new URL(url);\n if (!parsed.username && !parsed.password) return url;\n parsed.username = '';\n parsed.password = '';\n return `${parsed.toString().replace(/\\/$/, '')} (credentials redacted)`;\n } catch {\n return url;\n }\n}\n\n// A stopped Ollama/LM Studio/Cohere endpoint prints identically to a running one otherwise; any\n// HTTP response counts as reachable (probeReachable), so this only distinguishes up from down.\nasync function embedReachable(e: { provider: string; url?: string }): Promise<boolean | null> {\n if (e.provider === 'static' || (e.provider === 'openai' && !e.url)) return null;\n const base = (e.url ?? 'https://api.cohere.com').replace(/\\/+$/, '');\n return probeReachable(e.provider === 'openai' ? `${base}/models` : base);\n}\n\nconst status: Command = async (ctx) => {\n const { values } = parse(ctx.argv, `usage: ${ctx.name} ${USAGE.status}`, { ...FORMAT, ...CONFIG });\n const format = formatOf(values);\n const cfg = ctx.resolveConfig(values.config as string | undefined);\n const { db, dbPath, warnings } = open(cfg);\n printWarnings(warnings);\n\n const features = featureStates(cfg);\n const e = embedConfig(cfg);\n const hasModel = e ? modelPresent(cfg) : false;\n const reachable = e ? await embedReachable(e) : null;\n const vectors = e ? vectorState(db) : null;\n const presets = presetCoverage(db, cfg);\n const staticLanguages = e && e.provider === 'static' ? readLanguages(e.model) : undefined;\n const declaredLanguages = staticLanguages && staticLanguages.length > 0 ? staticLanguages : null;\n const detectedLanguages = languageDistribution(db) ?? null;\n const heartbeat = getMeta(db, 'watch_heartbeat');\n // Derived at open() (F): 3x the largest reconcile this cache has ever held its write\n // transaction for, floored at 30s -- read back from the connection rather than\n // recomputed here, so this always reports what open() actually set.\n const busyTimeoutMs = (db.prepare('PRAGMA busy_timeout').get() as { timeout: number }).timeout;\n // The env var holds the token; only its name and whether it is set are ever reported.\n const keyEnv = e?.key ? { name: e.key, set: (process.env[e.key] ?? '') !== '' } : null;\n const result = {\n config: cfg.configPath,\n configVersion: SUPPORTED_CONFIG_VERSION,\n migratedFrom: cfg.migratedFrom ?? null,\n unknownConfigKeys: cfg.unknownKeys ?? null,\n tree: cfg.baseDir,\n cache: dirname(dbPath),\n db: dbPath,\n cacheSchema: SCHEMA_VERSION,\n docs: docCount(db),\n unparseableFrontmatter: parseErrorCount(db),\n features: features.on,\n featuresOff: features.off,\n embed: e ? { provider: e.provider, model: e.model, dir: e.provider === 'static' ? modelDir(e.model) : null, url: e.url ? redactUrl(e.url) : null, reachable, keyEnv, downloaded: hasModel, languages: declaredLanguages, detectedLanguages, ...(vectors ?? {}) } : null,\n presets,\n queries: Object.keys(cfg.queries ?? {}).length,\n watcherPid: getMeta(db, 'watch_pid'),\n watcherHeartbeatSecondsAgo: heartbeat ? Math.round((Date.now() - Date.parse(heartbeat)) / 1000) : null,\n busyTimeoutMs,\n };\n\n if (format === 'json') {\n console.log(JSON.stringify(result, null, 2));\n } else {\n const migrated = cfg.migratedFrom !== undefined ? `, migrated from v${cfg.migratedFrom} on this run` : '';\n console.log(`config: ${result.config} (v${SUPPORTED_CONFIG_VERSION}${migrated})`);\n // A load-time stderr warning has scrolled away by the time anyone looks here.\n if (cfg.unknownKeys) console.log(` unknown keys, ignored: ${cfg.unknownKeys.join(', ')}`);\n console.log(`tree: ${result.tree}`);\n console.log(`cache: ${result.cache} (schema ${SCHEMA_VERSION}; delete this directory to reset)`);\n console.log(`docs: ${result.docs}${result.unparseableFrontmatter > 0 ? ` (${result.unparseableFrontmatter} with unparseable frontmatter: WHERE _parse_error IS NOT NULL)` : ''}`);\n console.log('');\n console.log(featuresLine(features));\n // One authoritative block, off states included: an absent model reads as a degraded search,\n // not as an empty corpus.\n if (!e) console.log('embed: off (no preset asks for vectors)');\n else {\n console.log(`embed: ${e.provider} ${e.model}`);\n if (e.provider === 'static') console.log(` model: ${modelDir(e.model)} (${hasModel ? 'present' : `missing, ${isDownloadable(e.model) ? `fetches automatically at first search, or run \\`${ctx.name} download\\` to prefetch` : `no ${MODEL_FILENAMES} in that directory`}`})`);\n if (declaredLanguages) console.log(` languages: ${declaredLanguages.join(', ')} (declared by the model card)`);\n // Cohere defaults to api.cohere.com with no url in the config; show it anyway so the\n // reachability probe below has something to name.\n if (e.url || reachable !== null) {\n const shown = e.url ? redactUrl(e.url) : 'https://api.cohere.com (default)';\n console.log(` url: ${shown}${reachable === null ? '' : reachable ? ' (reachable)' : ' (unreachable)'}`);\n }\n if (keyEnv) console.log(` key: env ${keyEnv.name} (${keyEnv.set ? 'set' : 'NOT SET'})`);\n if (!hasModel) console.log(' vectors: off until the model is present; search runs on words and links');\n else if (vectors) console.log(` vectors: ${vectors.embedded} embedded, ${vectors.pending} pending (embedded on the first semantic search)`);\n if (detectedLanguages) {\n const total = Object.values(detectedLanguages).reduce((a, b) => a + b, 0);\n const breakdown = Object.entries(detectedLanguages)\n .sort((a, b) => b[1] - a[1])\n .map(([code, n]) => `${code} ${Math.round((n / total) * 100)}%`)\n .join(', ');\n console.log(` detected: ${breakdown} (${total} chunks classified)`);\n }\n }\n console.log('');\n for (const line of presetsLines(presets)) console.log(line);\n console.log('');\n console.log(`queries: ${result.queries} saved (${ctx.name} --list)`);\n const pid = result.watcherPid ? `pid ${result.watcherPid}, ` : '';\n console.log(result.watcherHeartbeatSecondsAgo === null ? 'watcher: none' : `watcher: ${pid}last heartbeat ${result.watcherHeartbeatSecondsAgo}s ago`);\n console.log(`sqlite: busy_timeout ${result.busyTimeoutMs}ms (derived: 3x the largest reconcile this cache has recorded, floored at 30000ms)`);\n }\n db.close();\n};\nexport default status;\n"],"names":["dirname","presetCoverage","embedConfig","featureStates","SUPPORTED_CONFIG_VERSION","docCount","getMeta","open","SCHEMA_VERSION","probeReachable","languageDistribution","isDownloadable","MODEL_FILENAMES","modelDir","modelPresent","readLanguages","featuresLine","presetsLines","USAGE","CONFIG","FORMAT","formatOf","parse","printWarnings","vectorState","db","row","prepare","get","parseErrorCount","n","redactUrl","url","parsed","URL","username","password","toString","replace","embedReachable","e","provider","base","status","ctx","process","cfg","values","argv","name","format","resolveConfig","config","dbPath","warnings","features","hasModel","reachable","vectors","presets","staticLanguages","model","undefined","declaredLanguages","length","detectedLanguages","heartbeat","busyTimeoutMs","timeout","keyEnv","key","set","env","result","configPath","configVersion","migratedFrom","unknownConfigKeys","unknownKeys","tree","baseDir","cache","cacheSchema","docs","unparseableFrontmatter","on","featuresOff","off","embed","dir","downloaded","languages","queries","Object","keys","watcherPid","watcherHeartbeatSecondsAgo","Math","round","Date","now","console","log","JSON","stringify","migrated","join","shown","embedded","pending","total","reduce","a","b","breakdown","entries","sort","map","code","line","pid","close"],"mappings":"AAAA,SAASA,OAAO,QAAQ,YAAY;AACpC,SAASC,cAAc,QAAQ,uBAAuB;AACtD,SAASC,WAAW,EAAEC,aAAa,EAAEC,wBAAwB,QAAQ,qBAAqB;AAC1F,SAASC,QAAQ,EAAEC,OAAO,EAAEC,IAAI,EAAEC,cAAc,QAAQ,iBAAiB;AACzE,SAASC,cAAc,QAAQ,mBAAmB;AAClD,SAASC,oBAAoB,QAAQ,sBAAsB;AAC3D,SAASC,cAAc,EAAEC,eAAe,EAAEC,QAAQ,EAAEC,YAAY,EAAEC,aAAa,QAAQ,oBAAoB;AAC3G,SAASC,YAAY,EAAEC,YAAY,QAAQ,sBAAsB;AACjE,SAASC,KAAK,QAAQ,aAAa;AACnC,SAASC,MAAM,EAAEC,MAAM,EAAEC,QAAQ,EAAEC,KAAK,EAAEC,aAAa,QAAQ,cAAc;AAG7E,0FAA0F;AAC1F,+FAA+F;AAC/F,4FAA4F;AAC5F,kGAAkG;AAElG,2FAA2F;AAC3F,4FAA4F;AAC5F,yDAAyD;AACzD,SAASC,YAAYC,EAAiC;IACpD,IAAI;QACF,MAAMC,MAAMD,GAAGE,OAAO,CAAC,yFAAyFC,GAAG;QACnH,OAAOF;IACT,EAAE,OAAM;QACN,OAAO;IACT;AACF;AAEA,SAASG,gBAAgBJ,EAAiC;IACxD,MAAMC,MAAMD,GAAGE,OAAO,CAAC,CAAC,sEAAsE,CAAC,EAAEC,GAAG;IACpG,OAAOF,IAAII,CAAC;AACd;AAEA,8FAA8F;AAC9F,qEAAqE;AACrE,OAAO,SAASC,UAAUC,GAAW;IACnC,IAAI;QACF,MAAMC,SAAS,IAAIC,IAAIF;QACvB,IAAI,CAACC,OAAOE,QAAQ,IAAI,CAACF,OAAOG,QAAQ,EAAE,OAAOJ;QACjDC,OAAOE,QAAQ,GAAG;QAClBF,OAAOG,QAAQ,GAAG;QAClB,OAAO,GAAGH,OAAOI,QAAQ,GAAGC,OAAO,CAAC,OAAO,IAAI,uBAAuB,CAAC;IACzE,EAAE,OAAM;QACN,OAAON;IACT;AACF;AAEA,gGAAgG;AAChG,+FAA+F;AAC/F,eAAeO,eAAeC,CAAqC;QAEnDA;IADd,IAAIA,EAAEC,QAAQ,KAAK,YAAaD,EAAEC,QAAQ,KAAK,YAAY,CAACD,EAAER,GAAG,EAAG,OAAO;IAC3E,MAAMU,OAAO,EAACF,SAAAA,EAAER,GAAG,cAALQ,oBAAAA,SAAS,0BAA0BF,OAAO,CAAC,QAAQ;IACjE,OAAO7B,eAAe+B,EAAEC,QAAQ,KAAK,WAAW,GAAGC,KAAK,OAAO,CAAC,GAAGA;AACrE;AAEA,MAAMC,SAAkB,OAAOC;QAeHlC,uBAOmBmC,oBAI7BC,mBACKA,kBAWEA;IArCvB,MAAM,EAAEC,MAAM,EAAE,GAAGzB,MAAMsB,IAAII,IAAI,EAAE,CAAC,OAAO,EAAEJ,IAAIK,IAAI,CAAC,CAAC,EAAE/B,MAAMyB,MAAM,EAAE,EAAE;QAAE,GAAGvB,MAAM;QAAE,GAAGD,MAAM;IAAC;IAChG,MAAM+B,SAAS7B,SAAS0B;IACxB,MAAMD,MAAMF,IAAIO,aAAa,CAACJ,OAAOK,MAAM;IAC3C,MAAM,EAAE3B,EAAE,EAAE4B,MAAM,EAAEC,QAAQ,EAAE,GAAG/C,KAAKuC;IACtCvB,cAAc+B;IAEd,MAAMC,WAAWpD,cAAc2C;IAC/B,MAAMN,IAAItC,YAAY4C;IACtB,MAAMU,WAAWhB,IAAI1B,aAAagC,OAAO;IACzC,MAAMW,YAAYjB,IAAI,MAAMD,eAAeC,KAAK;IAChD,MAAMkB,UAAUlB,IAAIhB,YAAYC,MAAM;IACtC,MAAMkC,UAAU1D,eAAewB,IAAIqB;IACnC,MAAMc,kBAAkBpB,KAAKA,EAAEC,QAAQ,KAAK,WAAW1B,cAAcyB,EAAEqB,KAAK,IAAIC;IAChF,MAAMC,oBAAoBH,mBAAmBA,gBAAgBI,MAAM,GAAG,IAAIJ,kBAAkB;IAC5F,MAAMK,qBAAoBvD,wBAAAA,qBAAqBe,iBAArBf,mCAAAA,wBAA4B;IACtD,MAAMwD,YAAY5D,QAAQmB,IAAI;IAC9B,qFAAqF;IACrF,+EAA+E;IAC/E,oEAAoE;IACpE,MAAM0C,gBAAgB,AAAC1C,GAAGE,OAAO,CAAC,uBAAuBC,GAAG,GAA2BwC,OAAO;IAC9F,sFAAsF;IACtF,MAAMC,SAAS7B,CAAAA,cAAAA,wBAAAA,EAAG8B,GAAG,IAAG;QAAErB,MAAMT,EAAE8B,GAAG;QAAEC,KAAK,EAAC1B,qBAAAA,QAAQ2B,GAAG,CAAChC,EAAE8B,GAAG,CAAC,cAAlBzB,gCAAAA,qBAAsB,QAAQ;IAAG,IAAI;IAClF,MAAM4B,SAAS;QACbrB,QAAQN,IAAI4B,UAAU;QACtBC,eAAevE;QACfwE,YAAY,GAAE9B,oBAAAA,IAAI8B,YAAY,cAAhB9B,+BAAAA,oBAAoB;QAClC+B,iBAAiB,GAAE/B,mBAAAA,IAAIgC,WAAW,cAAfhC,8BAAAA,mBAAmB;QACtCiC,MAAMjC,IAAIkC,OAAO;QACjBC,OAAOjF,QAAQqD;QACf5B,IAAI4B;QACJ6B,aAAa1E;QACb2E,MAAM9E,SAASoB;QACf2D,wBAAwBvD,gBAAgBJ;QACxC8B,UAAUA,SAAS8B,EAAE;QACrBC,aAAa/B,SAASgC,GAAG;QACzBC,OAAOhD,IAAI;YAAEC,UAAUD,EAAEC,QAAQ;YAAEoB,OAAOrB,EAAEqB,KAAK;YAAE4B,KAAKjD,EAAEC,QAAQ,KAAK,WAAW5B,SAAS2B,EAAEqB,KAAK,IAAI;YAAM7B,KAAKQ,EAAER,GAAG,GAAGD,UAAUS,EAAER,GAAG,IAAI;YAAMyB;YAAWY;YAAQqB,YAAYlC;YAAUmC,WAAW5B;YAAmBE;eAAuBP,oBAAAA,qBAAAA,UAAW,CAAC,CAAhB;QAAmB,IAAI;QACnQC;QACAiC,SAASC,OAAOC,IAAI,EAAChD,eAAAA,IAAI8C,OAAO,cAAX9C,0BAAAA,eAAe,CAAC,GAAGkB,MAAM;QAC9C+B,YAAYzF,QAAQmB,IAAI;QACxBuE,4BAA4B9B,YAAY+B,KAAKC,KAAK,CAAC,AAACC,CAAAA,KAAKC,GAAG,KAAKD,KAAK7E,KAAK,CAAC4C,UAAS,IAAK,QAAQ;QAClGC;IACF;IAEA,IAAIjB,WAAW,QAAQ;QACrBmD,QAAQC,GAAG,CAACC,KAAKC,SAAS,CAAC/B,QAAQ,MAAM;IAC3C,OAAO;QACL,MAAMgC,WAAW3D,IAAI8B,YAAY,KAAKd,YAAY,CAAC,iBAAiB,EAAEhB,IAAI8B,YAAY,CAAC,YAAY,CAAC,GAAG;QACvGyB,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOrB,MAAM,CAAC,GAAG,EAAEhD,2BAA2BqG,SAAS,CAAC,CAAC;QAClF,8EAA8E;QAC9E,IAAI3D,IAAIgC,WAAW,EAAEuB,QAAQC,GAAG,CAAC,CAAC,iCAAiC,EAAExD,IAAIgC,WAAW,CAAC4B,IAAI,CAAC,OAAO;QACjGL,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOM,IAAI,EAAE;QACtCsB,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOQ,KAAK,CAAC,SAAS,EAAEzE,eAAe,iCAAiC,CAAC;QAClG6F,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOU,IAAI,GAAGV,OAAOW,sBAAsB,GAAG,IAAI,CAAC,GAAG,EAAEX,OAAOW,sBAAsB,CAAC,8DAA8D,CAAC,GAAG,IAAI;QACrLiB,QAAQC,GAAG,CAAC;QACZD,QAAQC,GAAG,CAACtF,aAAauC;QACzB,4FAA4F;QAC5F,0BAA0B;QAC1B,IAAI,CAACf,GAAG6D,QAAQC,GAAG,CAAC;aACf;YACHD,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE9D,EAAEC,QAAQ,CAAC,CAAC,EAAED,EAAEqB,KAAK,EAAE;YAChD,IAAIrB,EAAEC,QAAQ,KAAK,UAAU4D,QAAQC,GAAG,CAAC,CAAC,mBAAmB,EAAEzF,SAAS2B,EAAEqB,KAAK,EAAE,EAAE,EAAEL,WAAW,YAAY,CAAC,SAAS,EAAE7C,eAAe6B,EAAEqB,KAAK,IAAI,CAAC,gDAAgD,EAAEjB,IAAIK,IAAI,CAAC,uBAAuB,CAAC,GAAG,CAAC,GAAG,EAAErC,gBAAgB,kBAAkB,CAAC,EAAE,CAAC,CAAC,CAAC;YACvR,IAAImD,mBAAmBsC,QAAQC,GAAG,CAAC,CAAC,qBAAqB,EAAEvC,kBAAkB2C,IAAI,CAAC,MAAM,6BAA6B,CAAC;YACtH,qFAAqF;YACrF,kDAAkD;YAClD,IAAIlE,EAAER,GAAG,IAAIyB,cAAc,MAAM;gBAC/B,MAAMkD,QAAQnE,EAAER,GAAG,GAAGD,UAAUS,EAAER,GAAG,IAAI;gBACzCqE,QAAQC,GAAG,CAAC,CAAC,mBAAmB,EAAEK,QAAQlD,cAAc,OAAO,KAAKA,YAAY,iBAAiB,kBAAkB;YACrH;YACA,IAAIY,QAAQgC,QAAQC,GAAG,CAAC,CAAC,uBAAuB,EAAEjC,OAAOpB,IAAI,CAAC,EAAE,EAAEoB,OAAOE,GAAG,GAAG,QAAQ,UAAU,CAAC,CAAC;YACnG,IAAI,CAACf,UAAU6C,QAAQC,GAAG,CAAC;iBACtB,IAAI5C,SAAS2C,QAAQC,GAAG,CAAC,CAAC,mBAAmB,EAAE5C,QAAQkD,QAAQ,CAAC,WAAW,EAAElD,QAAQmD,OAAO,CAAC,gDAAgD,CAAC;YACnJ,IAAI5C,mBAAmB;gBACrB,MAAM6C,QAAQjB,OAAO9C,MAAM,CAACkB,mBAAmB8C,MAAM,CAAC,CAACC,GAAGC,IAAMD,IAAIC,GAAG;gBACvE,MAAMC,YAAYrB,OAAOsB,OAAO,CAAClD,mBAC9BmD,IAAI,CAAC,CAACJ,GAAGC,IAAMA,CAAC,CAAC,EAAE,GAAGD,CAAC,CAAC,EAAE,EAC1BK,GAAG,CAAC,CAAC,CAACC,MAAMxF,EAAE,GAAK,GAAGwF,KAAK,CAAC,EAAErB,KAAKC,KAAK,CAAC,AAACpE,IAAIgF,QAAS,KAAK,CAAC,CAAC,EAC9DJ,IAAI,CAAC;gBACRL,QAAQC,GAAG,CAAC,CAAC,oBAAoB,EAAEY,UAAU,EAAE,EAAEJ,MAAM,mBAAmB,CAAC;YAC7E;QACF;QACAT,QAAQC,GAAG,CAAC;QACZ,KAAK,MAAMiB,QAAQtG,aAAa0C,SAAU0C,QAAQC,GAAG,CAACiB;QACtDlB,QAAQC,GAAG,CAAC;QACZD,QAAQC,GAAG,CAAC,CAAC,UAAU,EAAE7B,OAAOmB,OAAO,CAAC,QAAQ,EAAEhD,IAAIK,IAAI,CAAC,QAAQ,CAAC;QACpE,MAAMuE,MAAM/C,OAAOsB,UAAU,GAAG,CAAC,IAAI,EAAEtB,OAAOsB,UAAU,CAAC,EAAE,CAAC,GAAG;QAC/DM,QAAQC,GAAG,CAAC7B,OAAOuB,0BAA0B,KAAK,OAAO,mBAAmB,CAAC,UAAU,EAAEwB,IAAI,eAAe,EAAE/C,OAAOuB,0BAA0B,CAAC,KAAK,CAAC;QACtJK,QAAQC,GAAG,CAAC,CAAC,uBAAuB,EAAE7B,OAAON,aAAa,CAAC,kFAAkF,CAAC;IAChJ;IACA1C,GAAGgG,KAAK;AACV;AACA,eAAe9E,OAAO"}
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
import type { DatabaseSync } from 'node:sqlite';
|
|
2
2
|
import type { FeatureName, ResolvedConfig, SearchOverrides } from '../config/index.js';
|
|
3
|
-
import type { Row } from '../output.js';
|
|
3
|
+
import type { Row } from '../output/output.js';
|
|
4
4
|
import type { PresetCoverage } from './status.js';
|
|
5
5
|
export interface TreeMap {
|
|
6
6
|
docs: {
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/map.ts"],"sourcesContent":["import type { DatabaseSync } from 'node:sqlite';\nimport type { FeatureName, ResolvedConfig, SearchOverrides } from '../config/index.ts';\nimport { featureEnabled, featureStates } from '../config/index.ts';\nimport type { Row } from '../output.ts';\nimport { INTERNAL_COLUMNS, scopedPaths, setupMapScope } from './scope.ts';\nimport type { PresetCoverage } from './status.ts';\nimport { presetCoverage } from './status.ts';\n\nexport interface TreeMap {\n docs: { count: number; bytes: number };\n fields: Row[]; // top 20 by coverage; fieldsTotal carries the real count\n fieldsTotal: number;\n features: { on: FeatureName[]; off: FeatureName[] };\n presets: PresetCoverage[];\n hubs: Row[];\n recent: Row[];\n recentCaveat: string | null;\n}\n\n// A result row is capped at SQLITE_MAX_COLUMN (2000, default); two aggregate expressions\n// per field keeps a chunk's row safely under that regardless of how many fields the tree has.\nconst MAP_COLUMN_CHUNK = 300;\n\nfunction chunk<T>(items: T[], size: number): T[][] {\n const out: T[][] = [];\n for (let i = 0; i < items.length; i += size) out.push(items.slice(i, i + size));\n return out;\n}\n\n// What is this scope: fixed-size output regardless of tree size. Coverage and features stay\n// global -- they describe the tree, not the current question.\nexport function mapTree(db: DatabaseSync, cfg: ResolvedConfig, overrides: SearchOverrides = {}): TreeMap {\n setupMapScope(db, scopedPaths(db, cfg, overrides));\n const scopeWhere = 'WHERE \"path\" IN (SELECT \"path\" FROM _map_scope)';\n const scopeAnd = 'AND f.\"path\" IN (SELECT \"path\" FROM _map_scope)';\n\n const docs = db.prepare(`SELECT COUNT(*) AS count, COALESCE(SUM(\"_size\"), 0) AS bytes FROM frontmatter ${scopeWhere}`).get() as { count: number; bytes: number };\n\n const columns = (db.prepare('PRAGMA table_info(frontmatter)').all() as Array<{ name: string }>).map((r) => r.name).filter((name) => !INTERNAL_COLUMNS.has(name));\n // Observed types, not declared: SQLite types per value, so a field can be text in most notes\n // and numeric in a few. One aggregate scan per chunk of columns; FILTER matches COUNT's nulls.\n const allFields: Row[] = [];\n for (const group of chunk(columns, MAP_COLUMN_CHUNK)) {\n const exprs = group.map((name, i) => {\n const quoted = `\"${name.split('\"').join('\"\"')}\"`;\n return `COUNT(${quoted}) AS n${i}, GROUP_CONCAT(DISTINCT typeof(${quoted})) FILTER (WHERE ${quoted} IS NOT NULL) AS t${i}`;\n });\n const result = db.prepare(`SELECT ${exprs.join(', ')} FROM frontmatter ${scopeWhere}`).get() as Record<string, number | string | null>;\n group.forEach((name, i) => allFields.push({ field: name, coverage: result[`n${i}`] as number, type: (result[`t${i}`] as string) ?? '' }));\n }\n allFields.sort((a, b) => (b.coverage as number) - (a.coverage as number));\n const fields = allFields.slice(0, 20);\n\n const hubs = featureEnabled(cfg, 'rank') ? (db.prepare(`SELECT f.\"path\" AS path, round(f.\"_rank\" * 100, 2) AS rank, content.title FROM frontmatter f JOIN content ON content.path = f.\"path\" WHERE f.\"_rank\" IS NOT NULL ${scopeAnd} ORDER BY f.\"_rank\" DESC LIMIT 8`).all() as Row[]) : [];\n\n const recent = db.prepare(`SELECT \"path\", datetime(\"_mtime\" / 1000, 'unixepoch') AS modified FROM frontmatter ${scopeWhere} ORDER BY \"_mtime\" DESC LIMIT 5`).all() as Row[];\n\n // A fresh clone/copy stamps files with checkout time, not edit history; second granularity\n // matches the `recent` table above and is coarse enough to catch that without an exact-ms match.\n const topSecond = db.prepare(`SELECT COUNT(*) AS n FROM frontmatter ${scopeWhere} GROUP BY CAST(\"_mtime\" / 1000 AS INTEGER) ORDER BY n DESC LIMIT 1`).get() as { n: number } | undefined;\n const recentCaveat = topSecond && docs.count > 1 && topSecond.n > docs.count / 2 ? `${topSecond.n} of ${docs.count} files share one modified second, so recency likely reflects a checkout or copy, not edit history` : null;\n\n return { docs, fields, fieldsTotal: allFields.length, features: featureStates(cfg), presets: presetCoverage(db, cfg), hubs, recent, recentCaveat };\n}\n"],"names":["featureEnabled","featureStates","INTERNAL_COLUMNS","scopedPaths","setupMapScope","presetCoverage","MAP_COLUMN_CHUNK","chunk","items","size","out","i","length","push","slice","mapTree","db","cfg","overrides","scopeWhere","scopeAnd","docs","prepare","get","columns","all","map","r","name","filter","has","allFields","group","exprs","quoted","split","join","result","forEach","field","coverage","type","sort","a","b","fields","hubs","recent","topSecond","recentCaveat","count","n","fieldsTotal","features","presets"],"mappings":"AAEA,SAASA,cAAc,EAAEC,aAAa,QAAQ,qBAAqB;AAEnE,SAASC,gBAAgB,EAAEC,WAAW,EAAEC,aAAa,QAAQ,aAAa;AAE1E,SAASC,cAAc,QAAQ,cAAc;AAa7C,yFAAyF;AACzF,8FAA8F;AAC9F,MAAMC,mBAAmB;AAEzB,SAASC,MAASC,KAAU,EAAEC,IAAY;IACxC,MAAMC,MAAa,EAAE;IACrB,IAAK,IAAIC,IAAI,GAAGA,IAAIH,MAAMI,MAAM,EAAED,KAAKF,KAAMC,IAAIG,IAAI,CAACL,MAAMM,KAAK,CAACH,GAAGA,IAAIF;IACzE,OAAOC;AACT;AAEA,4FAA4F;AAC5F,8DAA8D;AAC9D,OAAO,SAASK,QAAQC,EAAgB,EAAEC,GAAmB,EAAEC,YAA6B,CAAC,CAAC;IAC5Fd,cAAcY,IAAIb,YAAYa,IAAIC,KAAKC;IACvC,MAAMC,aAAa;IACnB,MAAMC,WAAW;IAEjB,MAAMC,OAAOL,GAAGM,OAAO,CAAC,CAAC,8EAA8E,EAAEH,YAAY,EAAEI,GAAG;IAE1H,MAAMC,UAAU,AAACR,GAAGM,OAAO,CAAC,kCAAkCG,GAAG,GAA+BC,GAAG,CAAC,CAACC,IAAMA,EAAEC,IAAI,EAAEC,MAAM,CAAC,CAACD,OAAS,CAAC1B,iBAAiB4B,GAAG,CAACF;IAC1J,6FAA6F;IAC7F,+FAA+F;IAC/F,MAAMG,YAAmB,EAAE;IAC3B,KAAK,MAAMC,SAASzB,MAAMiB,SAASlB,kBAAmB;QACpD,MAAM2B,QAAQD,MAAMN,GAAG,CAAC,CAACE,MAAMjB;YAC7B,MAAMuB,SAAS,CAAC,CAAC,EAAEN,KAAKO,KAAK,CAAC,KAAKC,IAAI,CAAC,MAAM,CAAC,CAAC;YAChD,OAAO,CAAC,MAAM,EAAEF,OAAO,MAAM,EAAEvB,EAAE,+BAA+B,EAAEuB,OAAO,iBAAiB,EAAEA,OAAO,kBAAkB,EAAEvB,GAAG;QAC5H;QACA,MAAM0B,SAASrB,GAAGM,OAAO,CAAC,CAAC,OAAO,EAAEW,MAAMG,IAAI,CAAC,MAAM,kBAAkB,EAAEjB,YAAY,EAAEI,GAAG;QAC1FS,MAAMM,OAAO,CAAC,CAACV,MAAMjB;gBAAgF0B;mBAA1EN,UAAUlB,IAAI,CAAC;gBAAE0B,OAAOX;gBAAMY,UAAUH,MAAM,CAAC,CAAC,CAAC,EAAE1B,GAAG,CAAC;gBAAY8B,IAAI,GAAGJ,WAAAA,MAAM,CAAC,CAAC,CAAC,EAAE1B,GAAG,CAAC,cAAf0B,sBAAAA,WAA8B;YAAG;;IACxI;IACAN,UAAUW,IAAI,CAAC,CAACC,GAAGC,IAAM,AAACA,EAAEJ,QAAQ,GAAeG,EAAEH,QAAQ;IAC7D,MAAMK,SAASd,UAAUjB,KAAK,CAAC,GAAG;IAElC,MAAMgC,OAAO9C,eAAeiB,KAAK,UAAWD,GAAGM,OAAO,CAAC,CAAC,iKAAiK,EAAEF,SAAS,gCAAgC,CAAC,EAAEK,GAAG,KAAe,EAAE;IAE3R,MAAMsB,SAAS/B,GAAGM,OAAO,CAAC,CAAC,mFAAmF,EAAEH,WAAW,+BAA+B,CAAC,EAAEM,GAAG;IAEhK,2FAA2F;IAC3F,iGAAiG;IACjG,MAAMuB,YAAYhC,GAAGM,OAAO,CAAC,CAAC,sCAAsC,EAAEH,WAAW,kEAAkE,CAAC,EAAEI,GAAG;IACzJ,MAAM0B,eAAeD,aAAa3B,KAAK6B,KAAK,GAAG,KAAKF,UAAUG,CAAC,GAAG9B,KAAK6B,KAAK,GAAG,IAAI,GAAGF,UAAUG,CAAC,CAAC,IAAI,EAAE9B,KAAK6B,KAAK,CAAC,iGAAiG,CAAC,GAAG;IAExN,OAAO;QAAE7B;QAAMwB;QAAQO,aAAarB,UAAUnB,MAAM;QAAEyC,UAAUpD,cAAcgB;QAAMqC,SAASjD,eAAeW,IAAIC;QAAM6B;QAAMC;QAAQE;IAAa;AACnJ"}
|
|
1
|
+
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/map.ts"],"sourcesContent":["import type { DatabaseSync } from 'node:sqlite';\nimport type { FeatureName, ResolvedConfig, SearchOverrides } from '../config/index.ts';\nimport { featureEnabled, featureStates } from '../config/index.ts';\nimport type { Row } from '../output/output.ts';\nimport { INTERNAL_COLUMNS, scopedPaths, setupMapScope } from './scope.ts';\nimport type { PresetCoverage } from './status.ts';\nimport { presetCoverage } from './status.ts';\n\nexport interface TreeMap {\n docs: { count: number; bytes: number };\n fields: Row[]; // top 20 by coverage; fieldsTotal carries the real count\n fieldsTotal: number;\n features: { on: FeatureName[]; off: FeatureName[] };\n presets: PresetCoverage[];\n hubs: Row[];\n recent: Row[];\n recentCaveat: string | null;\n}\n\n// A result row is capped at SQLITE_MAX_COLUMN (2000, default); two aggregate expressions\n// per field keeps a chunk's row safely under that regardless of how many fields the tree has.\nconst MAP_COLUMN_CHUNK = 300;\n\nfunction chunk<T>(items: T[], size: number): T[][] {\n const out: T[][] = [];\n for (let i = 0; i < items.length; i += size) out.push(items.slice(i, i + size));\n return out;\n}\n\n// What is this scope: fixed-size output regardless of tree size. Coverage and features stay\n// global -- they describe the tree, not the current question.\nexport function mapTree(db: DatabaseSync, cfg: ResolvedConfig, overrides: SearchOverrides = {}): TreeMap {\n setupMapScope(db, scopedPaths(db, cfg, overrides));\n const scopeWhere = 'WHERE \"path\" IN (SELECT \"path\" FROM _map_scope)';\n const scopeAnd = 'AND f.\"path\" IN (SELECT \"path\" FROM _map_scope)';\n\n const docs = db.prepare(`SELECT COUNT(*) AS count, COALESCE(SUM(\"_size\"), 0) AS bytes FROM frontmatter ${scopeWhere}`).get() as { count: number; bytes: number };\n\n const columns = (db.prepare('PRAGMA table_info(frontmatter)').all() as Array<{ name: string }>).map((r) => r.name).filter((name) => !INTERNAL_COLUMNS.has(name));\n // Observed types, not declared: SQLite types per value, so a field can be text in most notes\n // and numeric in a few. One aggregate scan per chunk of columns; FILTER matches COUNT's nulls.\n const allFields: Row[] = [];\n for (const group of chunk(columns, MAP_COLUMN_CHUNK)) {\n const exprs = group.map((name, i) => {\n const quoted = `\"${name.split('\"').join('\"\"')}\"`;\n return `COUNT(${quoted}) AS n${i}, GROUP_CONCAT(DISTINCT typeof(${quoted})) FILTER (WHERE ${quoted} IS NOT NULL) AS t${i}`;\n });\n const result = db.prepare(`SELECT ${exprs.join(', ')} FROM frontmatter ${scopeWhere}`).get() as Record<string, number | string | null>;\n group.forEach((name, i) => allFields.push({ field: name, coverage: result[`n${i}`] as number, type: (result[`t${i}`] as string) ?? '' }));\n }\n allFields.sort((a, b) => (b.coverage as number) - (a.coverage as number));\n const fields = allFields.slice(0, 20);\n\n const hubs = featureEnabled(cfg, 'rank') ? (db.prepare(`SELECT f.\"path\" AS path, round(f.\"_rank\" * 100, 2) AS rank, content.title FROM frontmatter f JOIN content ON content.path = f.\"path\" WHERE f.\"_rank\" IS NOT NULL ${scopeAnd} ORDER BY f.\"_rank\" DESC LIMIT 8`).all() as Row[]) : [];\n\n const recent = db.prepare(`SELECT \"path\", datetime(\"_mtime\" / 1000, 'unixepoch') AS modified FROM frontmatter ${scopeWhere} ORDER BY \"_mtime\" DESC LIMIT 5`).all() as Row[];\n\n // A fresh clone/copy stamps files with checkout time, not edit history; second granularity\n // matches the `recent` table above and is coarse enough to catch that without an exact-ms match.\n const topSecond = db.prepare(`SELECT COUNT(*) AS n FROM frontmatter ${scopeWhere} GROUP BY CAST(\"_mtime\" / 1000 AS INTEGER) ORDER BY n DESC LIMIT 1`).get() as { n: number } | undefined;\n const recentCaveat = topSecond && docs.count > 1 && topSecond.n > docs.count / 2 ? `${topSecond.n} of ${docs.count} files share one modified second, so recency likely reflects a checkout or copy, not edit history` : null;\n\n return { docs, fields, fieldsTotal: allFields.length, features: featureStates(cfg), presets: presetCoverage(db, cfg), hubs, recent, recentCaveat };\n}\n"],"names":["featureEnabled","featureStates","INTERNAL_COLUMNS","scopedPaths","setupMapScope","presetCoverage","MAP_COLUMN_CHUNK","chunk","items","size","out","i","length","push","slice","mapTree","db","cfg","overrides","scopeWhere","scopeAnd","docs","prepare","get","columns","all","map","r","name","filter","has","allFields","group","exprs","quoted","split","join","result","forEach","field","coverage","type","sort","a","b","fields","hubs","recent","topSecond","recentCaveat","count","n","fieldsTotal","features","presets"],"mappings":"AAEA,SAASA,cAAc,EAAEC,aAAa,QAAQ,qBAAqB;AAEnE,SAASC,gBAAgB,EAAEC,WAAW,EAAEC,aAAa,QAAQ,aAAa;AAE1E,SAASC,cAAc,QAAQ,cAAc;AAa7C,yFAAyF;AACzF,8FAA8F;AAC9F,MAAMC,mBAAmB;AAEzB,SAASC,MAASC,KAAU,EAAEC,IAAY;IACxC,MAAMC,MAAa,EAAE;IACrB,IAAK,IAAIC,IAAI,GAAGA,IAAIH,MAAMI,MAAM,EAAED,KAAKF,KAAMC,IAAIG,IAAI,CAACL,MAAMM,KAAK,CAACH,GAAGA,IAAIF;IACzE,OAAOC;AACT;AAEA,4FAA4F;AAC5F,8DAA8D;AAC9D,OAAO,SAASK,QAAQC,EAAgB,EAAEC,GAAmB,EAAEC,YAA6B,CAAC,CAAC;IAC5Fd,cAAcY,IAAIb,YAAYa,IAAIC,KAAKC;IACvC,MAAMC,aAAa;IACnB,MAAMC,WAAW;IAEjB,MAAMC,OAAOL,GAAGM,OAAO,CAAC,CAAC,8EAA8E,EAAEH,YAAY,EAAEI,GAAG;IAE1H,MAAMC,UAAU,AAACR,GAAGM,OAAO,CAAC,kCAAkCG,GAAG,GAA+BC,GAAG,CAAC,CAACC,IAAMA,EAAEC,IAAI,EAAEC,MAAM,CAAC,CAACD,OAAS,CAAC1B,iBAAiB4B,GAAG,CAACF;IAC1J,6FAA6F;IAC7F,+FAA+F;IAC/F,MAAMG,YAAmB,EAAE;IAC3B,KAAK,MAAMC,SAASzB,MAAMiB,SAASlB,kBAAmB;QACpD,MAAM2B,QAAQD,MAAMN,GAAG,CAAC,CAACE,MAAMjB;YAC7B,MAAMuB,SAAS,CAAC,CAAC,EAAEN,KAAKO,KAAK,CAAC,KAAKC,IAAI,CAAC,MAAM,CAAC,CAAC;YAChD,OAAO,CAAC,MAAM,EAAEF,OAAO,MAAM,EAAEvB,EAAE,+BAA+B,EAAEuB,OAAO,iBAAiB,EAAEA,OAAO,kBAAkB,EAAEvB,GAAG;QAC5H;QACA,MAAM0B,SAASrB,GAAGM,OAAO,CAAC,CAAC,OAAO,EAAEW,MAAMG,IAAI,CAAC,MAAM,kBAAkB,EAAEjB,YAAY,EAAEI,GAAG;QAC1FS,MAAMM,OAAO,CAAC,CAACV,MAAMjB;gBAAgF0B;mBAA1EN,UAAUlB,IAAI,CAAC;gBAAE0B,OAAOX;gBAAMY,UAAUH,MAAM,CAAC,CAAC,CAAC,EAAE1B,GAAG,CAAC;gBAAY8B,IAAI,GAAGJ,WAAAA,MAAM,CAAC,CAAC,CAAC,EAAE1B,GAAG,CAAC,cAAf0B,sBAAAA,WAA8B;YAAG;;IACxI;IACAN,UAAUW,IAAI,CAAC,CAACC,GAAGC,IAAM,AAACA,EAAEJ,QAAQ,GAAeG,EAAEH,QAAQ;IAC7D,MAAMK,SAASd,UAAUjB,KAAK,CAAC,GAAG;IAElC,MAAMgC,OAAO9C,eAAeiB,KAAK,UAAWD,GAAGM,OAAO,CAAC,CAAC,iKAAiK,EAAEF,SAAS,gCAAgC,CAAC,EAAEK,GAAG,KAAe,EAAE;IAE3R,MAAMsB,SAAS/B,GAAGM,OAAO,CAAC,CAAC,mFAAmF,EAAEH,WAAW,+BAA+B,CAAC,EAAEM,GAAG;IAEhK,2FAA2F;IAC3F,iGAAiG;IACjG,MAAMuB,YAAYhC,GAAGM,OAAO,CAAC,CAAC,sCAAsC,EAAEH,WAAW,kEAAkE,CAAC,EAAEI,GAAG;IACzJ,MAAM0B,eAAeD,aAAa3B,KAAK6B,KAAK,GAAG,KAAKF,UAAUG,CAAC,GAAG9B,KAAK6B,KAAK,GAAG,IAAI,GAAGF,UAAUG,CAAC,CAAC,IAAI,EAAE9B,KAAK6B,KAAK,CAAC,iGAAiG,CAAC,GAAG;IAExN,OAAO;QAAE7B;QAAMwB;QAAQO,aAAarB,UAAUnB,MAAM;QAAEyC,UAAUpD,cAAcgB;QAAMqC,SAASjD,eAAeW,IAAIC;QAAM6B;QAAMC;QAAQE;IAAa;AACnJ"}
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
import type { DatabaseSync } from 'node:sqlite';
|
|
2
2
|
import type { FeatureName, ResolvedConfig, SearchOverrides } from '../config/index.js';
|
|
3
|
-
import type { Row } from '../output.js';
|
|
3
|
+
import type { Row } from '../output/output.js';
|
|
4
4
|
export declare function resolveNote(paths: string[], arg: string): string;
|
|
5
5
|
export interface Peek {
|
|
6
6
|
path: string;
|
|
@@ -1,4 +1,5 @@
|
|
|
1
1
|
import posix from 'node:path/posix';
|
|
2
|
+
import { estimateTokens } from '../chunk/index.js';
|
|
2
3
|
import { featureEnabled } from '../config/index.js';
|
|
3
4
|
import { SenseError } from '../errors.js';
|
|
4
5
|
import { INTERNAL_COLUMNS, scopedPaths } from './scope.js';
|
|
@@ -18,6 +19,7 @@ const PEEK_LIST_LIMIT = 20;
|
|
|
18
19
|
// ranges + token estimates (so the follow-up Read is a range, not the file), links both ways.
|
|
19
20
|
export function peek(db, cfg, pathArg, overrides = {}) {
|
|
20
21
|
var _row__parse_error, _row__size;
|
|
22
|
+
var _db_prepare_get;
|
|
21
23
|
const paths = db.prepare('SELECT "path" FROM frontmatter').all().map((r)=>r.path);
|
|
22
24
|
const path = resolveNote(paths, pathArg);
|
|
23
25
|
const row = db.prepare('SELECT * FROM frontmatter WHERE "path" = ?').get(path);
|
|
@@ -42,9 +44,14 @@ export function peek(db, cfg, pathArg, overrides = {}) {
|
|
|
42
44
|
backlinksTotal = db.prepare('SELECT COUNT(DISTINCT src) AS n FROM links WHERE dst = ? AND src != dst').get(path).n;
|
|
43
45
|
backlinks = db.prepare('SELECT DISTINCT src FROM links WHERE dst = ? AND src != dst ORDER BY src LIMIT ?').all(path, PEEK_LIST_LIMIT).map((r)=>r.src);
|
|
44
46
|
}
|
|
47
|
+
// content.text is the stripped body already computed at reconcile time (no extra file read),
|
|
48
|
+
// floored by the byte/4 estimate so stripped-away syntax and frontmatter can't vanish from the price.
|
|
49
|
+
const body = (_db_prepare_get = db.prepare('SELECT text FROM content WHERE "path" = ?').get(path)) === null || _db_prepare_get === void 0 ? void 0 : _db_prepare_get.text;
|
|
50
|
+
const byteTokens = Math.ceil(((_row__size = row._size) !== null && _row__size !== void 0 ? _row__size : 0) / 4);
|
|
51
|
+
const tokens = body !== undefined ? Math.max(Math.ceil(estimateTokens(body)), byteTokens) : byteTokens;
|
|
45
52
|
return {
|
|
46
53
|
path,
|
|
47
|
-
tokens
|
|
54
|
+
tokens,
|
|
48
55
|
frontmatter,
|
|
49
56
|
parseError,
|
|
50
57
|
sections,
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/peek.ts"],"sourcesContent":["import posix from 'node:path/posix';\nimport type { DatabaseSync } from 'node:sqlite';\nimport type { FeatureName, ResolvedConfig, SearchOverrides } from '../config/index.ts';\nimport { featureEnabled } from '../config/index.ts';\nimport { SenseError } from '../errors.ts';\nimport type { Row } from '../output.ts';\nimport { INTERNAL_COLUMNS, scopedPaths } from './scope.ts';\n\n// Note resolution shared by peek and path: an exact path, or a unique basename (case\n// insensitive, .md stripped).\nexport function resolveNote(paths: string[], arg: string): string {\n const exact = paths.find((p) => p === arg);\n if (exact) return exact;\n const base = posix.basename(arg).replace(/\\.md$/i, '').toLowerCase();\n const matches = paths.filter((p) => posix.basename(p).replace(/\\.md$/i, '').toLowerCase() === base);\n if (matches.length === 1) return matches[0];\n if (matches.length > 1) throw new SenseError('NOTE_AMBIGUOUS', `\"${arg}\" is ambiguous: ${matches.join(', ')}`);\n throw new SenseError('NOTE_NOT_FOUND', `no note matches \"${arg}\"`);\n}\n\nexport interface Peek {\n path: string;\n tokens: number;\n frontmatter: Row;\n // Set when the note's frontmatter was refused, so an empty frontmatter block reads as \"did\n // not parse\" rather than \"has none\". Same reason `_parse_error` sits in the row.\n parseError: string | null;\n sections: Row[];\n outbound: string[];\n backlinks: string[];\n unresolved: string[];\n // Totals before truncation: a hub can have thousands of backlinks (or a note thousands of\n // headings), and peek's whole point is bounded output. Query the sections/links tables\n // directly for the full list.\n sectionsTotal: number;\n outboundTotal: number;\n backlinksTotal: number;\n unresolvedTotal: number;\n // Bounded k-hop expansion beyond the immediate ring already shown by outbound/backlinks\n // (depth starts at 2).\n off: FeatureName[]; // disabled features whose blocks are omitted (not empty)\n}\n\nconst PEEK_LIST_LIMIT = 20;\n\n// peek: everything about one note except its prose -- frontmatter, outline with line\n// ranges + token estimates (so the follow-up Read is a range, not the file), links both ways.\nexport function peek(db: DatabaseSync, cfg: ResolvedConfig, pathArg: string, overrides: SearchOverrides = {}): Peek {\n const paths = (db.prepare('SELECT \"path\" FROM frontmatter').all() as Array<{ path: string }>).map((r) => r.path);\n const path = resolveNote(paths, pathArg);\n\n const row = db.prepare('SELECT * FROM frontmatter WHERE \"path\" = ?').get(path) as Row;\n const parseError = (row._parse_error as string | null) ?? null;\n const frontmatter: Row = {};\n for (const [key, value] of Object.entries(row)) {\n if (!INTERNAL_COLUMNS.has(key) && value !== null) frontmatter[key] = value;\n }\n\n const sectionsTotal = featureEnabled(cfg, 'sections') ? (db.prepare('SELECT COUNT(*) AS n FROM sections WHERE \"path\" = ?').get(path) as { n: number }).n : 0;\n const sections = featureEnabled(cfg, 'sections') ? (db.prepare('SELECT level, heading, start_line, end_line, tokens FROM sections WHERE \"path\" = ? ORDER BY idx LIMIT ?').all(path, PEEK_LIST_LIMIT) as Row[]) : [];\n\n let outbound: string[] = [];\n let backlinks: string[] = [];\n let unresolved: string[] = [];\n let backlinksTotal = 0;\n const _allowed = scopedPaths(db, cfg, overrides);\n if (featureEnabled(cfg, 'links')) {\n const out = db.prepare('SELECT target, dst FROM links WHERE src = ? AND (dst IS NULL OR dst != src) ORDER BY target').all(path) as Array<{ target: string; dst: string | null }>;\n outbound = [...new Set(out.filter((l) => l.dst !== null).map((l) => l.dst as string))];\n unresolved = out.filter((l) => l.dst === null).map((l) => l.target);\n backlinksTotal = (db.prepare('SELECT COUNT(DISTINCT src) AS n FROM links WHERE dst = ? AND src != dst').get(path) as { n: number }).n;\n backlinks = (db.prepare('SELECT DISTINCT src FROM links WHERE dst = ? AND src != dst ORDER BY src LIMIT ?').all(path, PEEK_LIST_LIMIT) as Array<{ src: string }>).map((r) => r.src);\n }\n\n
|
|
1
|
+
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/peek.ts"],"sourcesContent":["import posix from 'node:path/posix';\nimport type { DatabaseSync } from 'node:sqlite';\nimport { estimateTokens } from '../chunk/index.ts';\nimport type { FeatureName, ResolvedConfig, SearchOverrides } from '../config/index.ts';\nimport { featureEnabled } from '../config/index.ts';\nimport { SenseError } from '../errors.ts';\nimport type { Row } from '../output/output.ts';\nimport { INTERNAL_COLUMNS, scopedPaths } from './scope.ts';\n\n// Note resolution shared by peek and path: an exact path, or a unique basename (case\n// insensitive, .md stripped).\nexport function resolveNote(paths: string[], arg: string): string {\n const exact = paths.find((p) => p === arg);\n if (exact) return exact;\n const base = posix.basename(arg).replace(/\\.md$/i, '').toLowerCase();\n const matches = paths.filter((p) => posix.basename(p).replace(/\\.md$/i, '').toLowerCase() === base);\n if (matches.length === 1) return matches[0];\n if (matches.length > 1) throw new SenseError('NOTE_AMBIGUOUS', `\"${arg}\" is ambiguous: ${matches.join(', ')}`);\n throw new SenseError('NOTE_NOT_FOUND', `no note matches \"${arg}\"`);\n}\n\nexport interface Peek {\n path: string;\n tokens: number;\n frontmatter: Row;\n // Set when the note's frontmatter was refused, so an empty frontmatter block reads as \"did\n // not parse\" rather than \"has none\". Same reason `_parse_error` sits in the row.\n parseError: string | null;\n sections: Row[];\n outbound: string[];\n backlinks: string[];\n unresolved: string[];\n // Totals before truncation: a hub can have thousands of backlinks (or a note thousands of\n // headings), and peek's whole point is bounded output. Query the sections/links tables\n // directly for the full list.\n sectionsTotal: number;\n outboundTotal: number;\n backlinksTotal: number;\n unresolvedTotal: number;\n // Bounded k-hop expansion beyond the immediate ring already shown by outbound/backlinks\n // (depth starts at 2).\n off: FeatureName[]; // disabled features whose blocks are omitted (not empty)\n}\n\nconst PEEK_LIST_LIMIT = 20;\n\n// peek: everything about one note except its prose -- frontmatter, outline with line\n// ranges + token estimates (so the follow-up Read is a range, not the file), links both ways.\nexport function peek(db: DatabaseSync, cfg: ResolvedConfig, pathArg: string, overrides: SearchOverrides = {}): Peek {\n const paths = (db.prepare('SELECT \"path\" FROM frontmatter').all() as Array<{ path: string }>).map((r) => r.path);\n const path = resolveNote(paths, pathArg);\n\n const row = db.prepare('SELECT * FROM frontmatter WHERE \"path\" = ?').get(path) as Row;\n const parseError = (row._parse_error as string | null) ?? null;\n const frontmatter: Row = {};\n for (const [key, value] of Object.entries(row)) {\n if (!INTERNAL_COLUMNS.has(key) && value !== null) frontmatter[key] = value;\n }\n\n const sectionsTotal = featureEnabled(cfg, 'sections') ? (db.prepare('SELECT COUNT(*) AS n FROM sections WHERE \"path\" = ?').get(path) as { n: number }).n : 0;\n const sections = featureEnabled(cfg, 'sections') ? (db.prepare('SELECT level, heading, start_line, end_line, tokens FROM sections WHERE \"path\" = ? ORDER BY idx LIMIT ?').all(path, PEEK_LIST_LIMIT) as Row[]) : [];\n\n let outbound: string[] = [];\n let backlinks: string[] = [];\n let unresolved: string[] = [];\n let backlinksTotal = 0;\n const _allowed = scopedPaths(db, cfg, overrides);\n if (featureEnabled(cfg, 'links')) {\n const out = db.prepare('SELECT target, dst FROM links WHERE src = ? AND (dst IS NULL OR dst != src) ORDER BY target').all(path) as Array<{ target: string; dst: string | null }>;\n outbound = [...new Set(out.filter((l) => l.dst !== null).map((l) => l.dst as string))];\n unresolved = out.filter((l) => l.dst === null).map((l) => l.target);\n backlinksTotal = (db.prepare('SELECT COUNT(DISTINCT src) AS n FROM links WHERE dst = ? AND src != dst').get(path) as { n: number }).n;\n backlinks = (db.prepare('SELECT DISTINCT src FROM links WHERE dst = ? AND src != dst ORDER BY src LIMIT ?').all(path, PEEK_LIST_LIMIT) as Array<{ src: string }>).map((r) => r.src);\n }\n\n // content.text is the stripped body already computed at reconcile time (no extra file read),\n // floored by the byte/4 estimate so stripped-away syntax and frontmatter can't vanish from the price.\n const body = (db.prepare('SELECT text FROM content WHERE \"path\" = ?').get(path) as { text: string } | undefined)?.text;\n const byteTokens = Math.ceil(((row._size as number) ?? 0) / 4);\n const tokens = body !== undefined ? Math.max(Math.ceil(estimateTokens(body)), byteTokens) : byteTokens;\n\n return {\n path,\n tokens,\n frontmatter,\n parseError,\n sections,\n outbound: outbound.slice(0, PEEK_LIST_LIMIT),\n backlinks,\n unresolved: unresolved.slice(0, PEEK_LIST_LIMIT),\n sectionsTotal,\n outboundTotal: outbound.length,\n backlinksTotal,\n unresolvedTotal: unresolved.length,\n off: (['sections', 'links'] as FeatureName[]).filter((name) => !featureEnabled(cfg, name)),\n };\n}\n"],"names":["posix","estimateTokens","featureEnabled","SenseError","INTERNAL_COLUMNS","scopedPaths","resolveNote","paths","arg","exact","find","p","base","basename","replace","toLowerCase","matches","filter","length","join","PEEK_LIST_LIMIT","peek","db","cfg","pathArg","overrides","row","prepare","all","map","r","path","get","parseError","_parse_error","frontmatter","key","value","Object","entries","has","sectionsTotal","n","sections","outbound","backlinks","unresolved","backlinksTotal","_allowed","out","Set","l","dst","target","src","body","text","byteTokens","Math","ceil","_size","tokens","undefined","max","slice","outboundTotal","unresolvedTotal","off","name"],"mappings":"AAAA,OAAOA,WAAW,kBAAkB;AAEpC,SAASC,cAAc,QAAQ,oBAAoB;AAEnD,SAASC,cAAc,QAAQ,qBAAqB;AACpD,SAASC,UAAU,QAAQ,eAAe;AAE1C,SAASC,gBAAgB,EAAEC,WAAW,QAAQ,aAAa;AAE3D,qFAAqF;AACrF,8BAA8B;AAC9B,OAAO,SAASC,YAAYC,KAAe,EAAEC,GAAW;IACtD,MAAMC,QAAQF,MAAMG,IAAI,CAAC,CAACC,IAAMA,MAAMH;IACtC,IAAIC,OAAO,OAAOA;IAClB,MAAMG,OAAOZ,MAAMa,QAAQ,CAACL,KAAKM,OAAO,CAAC,UAAU,IAAIC,WAAW;IAClE,MAAMC,UAAUT,MAAMU,MAAM,CAAC,CAACN,IAAMX,MAAMa,QAAQ,CAACF,GAAGG,OAAO,CAAC,UAAU,IAAIC,WAAW,OAAOH;IAC9F,IAAII,QAAQE,MAAM,KAAK,GAAG,OAAOF,OAAO,CAAC,EAAE;IAC3C,IAAIA,QAAQE,MAAM,GAAG,GAAG,MAAM,IAAIf,WAAW,kBAAkB,CAAC,CAAC,EAAEK,IAAI,gBAAgB,EAAEQ,QAAQG,IAAI,CAAC,OAAO;IAC7G,MAAM,IAAIhB,WAAW,kBAAkB,CAAC,iBAAiB,EAAEK,IAAI,CAAC,CAAC;AACnE;AAyBA,MAAMY,kBAAkB;AAExB,qFAAqF;AACrF,8FAA8F;AAC9F,OAAO,SAASC,KAAKC,EAAgB,EAAEC,GAAmB,EAAEC,OAAe,EAAEC,YAA6B,CAAC,CAAC;QAKtFC,mBAyBWA;QADjBJ;IA5Bd,MAAMf,QAAQ,AAACe,GAAGK,OAAO,CAAC,kCAAkCC,GAAG,GAA+BC,GAAG,CAAC,CAACC,IAAMA,EAAEC,IAAI;IAC/G,MAAMA,OAAOzB,YAAYC,OAAOiB;IAEhC,MAAME,MAAMJ,GAAGK,OAAO,CAAC,8CAA8CK,GAAG,CAACD;IACzE,MAAME,cAAcP,oBAAAA,IAAIQ,YAAY,cAAhBR,+BAAAA,oBAAsC;IAC1D,MAAMS,cAAmB,CAAC;IAC1B,KAAK,MAAM,CAACC,KAAKC,MAAM,IAAIC,OAAOC,OAAO,CAACb,KAAM;QAC9C,IAAI,CAACtB,iBAAiBoC,GAAG,CAACJ,QAAQC,UAAU,MAAMF,WAAW,CAACC,IAAI,GAAGC;IACvE;IAEA,MAAMI,gBAAgBvC,eAAeqB,KAAK,cAAc,AAACD,GAAGK,OAAO,CAAC,uDAAuDK,GAAG,CAACD,MAAwBW,CAAC,GAAG;IAC3J,MAAMC,WAAWzC,eAAeqB,KAAK,cAAeD,GAAGK,OAAO,CAAC,2GAA2GC,GAAG,CAACG,MAAMX,mBAA6B,EAAE;IAEnN,IAAIwB,WAAqB,EAAE;IAC3B,IAAIC,YAAsB,EAAE;IAC5B,IAAIC,aAAuB,EAAE;IAC7B,IAAIC,iBAAiB;IACrB,MAAMC,WAAW3C,YAAYiB,IAAIC,KAAKE;IACtC,IAAIvB,eAAeqB,KAAK,UAAU;QAChC,MAAM0B,MAAM3B,GAAGK,OAAO,CAAC,+FAA+FC,GAAG,CAACG;QAC1Ha,WAAW;eAAI,IAAIM,IAAID,IAAIhC,MAAM,CAAC,CAACkC,IAAMA,EAAEC,GAAG,KAAK,MAAMvB,GAAG,CAAC,CAACsB,IAAMA,EAAEC,GAAG;SAAa;QACtFN,aAAaG,IAAIhC,MAAM,CAAC,CAACkC,IAAMA,EAAEC,GAAG,KAAK,MAAMvB,GAAG,CAAC,CAACsB,IAAMA,EAAEE,MAAM;QAClEN,iBAAiB,AAACzB,GAAGK,OAAO,CAAC,2EAA2EK,GAAG,CAACD,MAAwBW,CAAC;QACrIG,YAAY,AAACvB,GAAGK,OAAO,CAAC,oFAAoFC,GAAG,CAACG,MAAMX,iBAA4CS,GAAG,CAAC,CAACC,IAAMA,EAAEwB,GAAG;IACpL;IAEA,6FAA6F;IAC7F,sGAAsG;IACtG,MAAMC,QAAQjC,kBAAAA,GAAGK,OAAO,CAAC,6CAA6CK,GAAG,CAACD,mBAA5DT,sCAAD,AAACA,gBAAoGkC,IAAI;IACtH,MAAMC,aAAaC,KAAKC,IAAI,CAAC,EAAEjC,aAAAA,IAAIkC,KAAK,cAATlC,wBAAAA,aAAwB,KAAK;IAC5D,MAAMmC,SAASN,SAASO,YAAYJ,KAAKK,GAAG,CAACL,KAAKC,IAAI,CAAC1D,eAAesD,QAAQE,cAAcA;IAE5F,OAAO;QACL1B;QACA8B;QACA1B;QACAF;QACAU;QACAC,UAAUA,SAASoB,KAAK,CAAC,GAAG5C;QAC5ByB;QACAC,YAAYA,WAAWkB,KAAK,CAAC,GAAG5C;QAChCqB;QACAwB,eAAerB,SAAS1B,MAAM;QAC9B6B;QACAmB,iBAAiBpB,WAAW5B,MAAM;QAClCiD,KAAK,AAAC;YAAC;YAAY;SAAQ,CAAmBlD,MAAM,CAAC,CAACmD,OAAS,CAAClE,eAAeqB,KAAK6C;IACtF;AACF"}
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
import { matchesGlob } from 'node:path';
|
|
2
|
-
import { columnHint } from '../column-hint.js';
|
|
3
2
|
import { anyPresetEmbeds, resolveSearch } from '../config/index.js';
|
|
3
|
+
import { columnHint } from '../output/column-hint.js';
|
|
4
4
|
export const INTERNAL_COLUMNS = new Set([
|
|
5
5
|
'path',
|
|
6
6
|
'_mtime',
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/scope.ts"],"sourcesContent":["import { matchesGlob } from 'node:path';\nimport type { DatabaseSync } from 'node:sqlite';\nimport {
|
|
1
|
+
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/scope.ts"],"sourcesContent":["import { matchesGlob } from 'node:path';\nimport type { DatabaseSync } from 'node:sqlite';\nimport type { ResolvedConfig, SearchOverrides } from '../config/index.ts';\nimport { anyPresetEmbeds, resolveSearch } from '../config/index.ts';\nimport { columnHint } from '../output/column-hint.ts';\n\nexport const INTERNAL_COLUMNS = new Set(['path', '_mtime', '_ctime', '_size', '_rank', '_parse_error']);\n\n// node:path's matchesGlob is experimental (stable behind an unstable-API flag) as of the\n// engines floor (Node >=22.20); scope filtering only ever needs single-pattern matching, so\n// it's used here in JS rather than running a directory walk in the query path.\nexport function inScope(path: string, include: string[], exclude?: string[]): boolean {\n if (!include.some((g) => matchesGlob(path, g))) return false;\n if (exclude?.some((g) => matchesGlob(path, g))) return false;\n return true;\n}\n\nexport function scopeHasEmbeddings(db: DatabaseSync, cfg: ResolvedConfig, scopedPaths: Set<string>): boolean {\n if (!anyPresetEmbeds(cfg)) return false; // the embeddings table doesn't exist at all in this case\n const rows = db.prepare('SELECT DISTINCT \"path\" FROM embeddings').all() as Array<{ path: string }>;\n return rows.some((r) => scopedPaths.has(r.path));\n}\n\n// Scope only (no --where): preset_files for a named preset, JS glob matching for an ad hoc\n// include/exclude override. Shared by scopedPaths() and search(), which also needs the\n// pre-where set to size the candidate-pool filter.\nexport function rawScope(db: DatabaseSync, cfg: ResolvedConfig, overrides: SearchOverrides, allPaths?: string[]): Set<string> {\n const effective = resolveSearch(cfg, overrides);\n const { include, exclude } = effective;\n const adHocScope = overrides.include !== undefined || overrides.exclude !== undefined || overrides.noExclude === true;\n if (!adHocScope) return new Set((db.prepare('SELECT \"path\" FROM preset_files WHERE preset = ?').all(effective.presetName) as Array<{ path: string }>).map((r) => r.path));\n const paths = allPaths ?? (db.prepare('SELECT \"path\" FROM frontmatter').all() as Array<{ path: string }>).map((r) => r.path);\n return new Set(paths.filter((p) => inScope(p, include, exclude)));\n}\n\nexport function narrowByWhere(db: DatabaseSync, paths: Set<string>, where: string | undefined): Set<string> {\n if (!where) return paths;\n let whereRows: Array<{ path: string }>;\n try {\n whereRows = db.prepare(`SELECT \"path\" FROM frontmatter f WHERE (${where})`).all() as Array<{ path: string }>;\n } catch (err) {\n throw columnHint(db, err as Error);\n }\n const wherePaths = new Set(whereRows.map((r) => r.path));\n return new Set([...paths].filter((p) => wherePaths.has(p)));\n}\n\n// The scope resolver for non-search commands (path, peek, map): same coverage rule search()\n// applies, then narrowed by the resolved `where`.\nexport function scopedPaths(db: DatabaseSync, cfg: ResolvedConfig, overrides: SearchOverrides): Set<string> {\n const effective = resolveSearch(cfg, overrides);\n return narrowByWhere(db, rawScope(db, cfg, overrides), effective.where);\n}\n\n// Materializes a path set into a named temp table (same shape as traverse.ts's allowed_nodes)\n// so a query can join/filter against it cheaply instead of binding a parameter per path.\nexport function materializeScope(db: DatabaseSync, table: string, paths: Set<string>): void {\n db.exec(`DROP TABLE IF EXISTS ${table}`);\n db.exec(`CREATE TEMP TABLE ${table} (\"path\" TEXT PRIMARY KEY)`);\n db.prepare(`INSERT INTO ${table} SELECT DISTINCT value FROM json_each(?1)`).run(JSON.stringify([...paths]));\n}\n\nexport function setupMapScope(db: DatabaseSync, paths: Set<string>): void {\n materializeScope(db, '_map_scope', paths);\n}\n"],"names":["matchesGlob","anyPresetEmbeds","resolveSearch","columnHint","INTERNAL_COLUMNS","Set","inScope","path","include","exclude","some","g","scopeHasEmbeddings","db","cfg","scopedPaths","rows","prepare","all","r","has","rawScope","overrides","allPaths","effective","adHocScope","undefined","noExclude","presetName","map","paths","filter","p","narrowByWhere","where","whereRows","err","wherePaths","materializeScope","table","exec","run","JSON","stringify","setupMapScope"],"mappings":"AAAA,SAASA,WAAW,QAAQ,YAAY;AAGxC,SAASC,eAAe,EAAEC,aAAa,QAAQ,qBAAqB;AACpE,SAASC,UAAU,QAAQ,2BAA2B;AAEtD,OAAO,MAAMC,mBAAmB,IAAIC,IAAI;IAAC;IAAQ;IAAU;IAAU;IAAS;IAAS;CAAe,EAAE;AAExG,yFAAyF;AACzF,4FAA4F;AAC5F,+EAA+E;AAC/E,OAAO,SAASC,QAAQC,IAAY,EAAEC,OAAiB,EAAEC,OAAkB;IACzE,IAAI,CAACD,QAAQE,IAAI,CAAC,CAACC,IAAMX,YAAYO,MAAMI,KAAK,OAAO;IACvD,IAAIF,oBAAAA,8BAAAA,QAASC,IAAI,CAAC,CAACC,IAAMX,YAAYO,MAAMI,KAAK,OAAO;IACvD,OAAO;AACT;AAEA,OAAO,SAASC,mBAAmBC,EAAgB,EAAEC,GAAmB,EAAEC,WAAwB;IAChG,IAAI,CAACd,gBAAgBa,MAAM,OAAO,OAAO,yDAAyD;IAClG,MAAME,OAAOH,GAAGI,OAAO,CAAC,0CAA0CC,GAAG;IACrE,OAAOF,KAAKN,IAAI,CAAC,CAACS,IAAMJ,YAAYK,GAAG,CAACD,EAAEZ,IAAI;AAChD;AAEA,2FAA2F;AAC3F,uFAAuF;AACvF,mDAAmD;AACnD,OAAO,SAASc,SAASR,EAAgB,EAAEC,GAAmB,EAAEQ,SAA0B,EAAEC,QAAmB;IAC7G,MAAMC,YAAYtB,cAAcY,KAAKQ;IACrC,MAAM,EAAEd,OAAO,EAAEC,OAAO,EAAE,GAAGe;IAC7B,MAAMC,aAAaH,UAAUd,OAAO,KAAKkB,aAAaJ,UAAUb,OAAO,KAAKiB,aAAaJ,UAAUK,SAAS,KAAK;IACjH,IAAI,CAACF,YAAY,OAAO,IAAIpB,IAAI,AAACQ,GAAGI,OAAO,CAAC,oDAAoDC,GAAG,CAACM,UAAUI,UAAU,EAA8BC,GAAG,CAAC,CAACV,IAAMA,EAAEZ,IAAI;IACvK,MAAMuB,QAAQP,qBAAAA,sBAAAA,WAAY,AAACV,GAAGI,OAAO,CAAC,kCAAkCC,GAAG,GAA+BW,GAAG,CAAC,CAACV,IAAMA,EAAEZ,IAAI;IAC3H,OAAO,IAAIF,IAAIyB,MAAMC,MAAM,CAAC,CAACC,IAAM1B,QAAQ0B,GAAGxB,SAASC;AACzD;AAEA,OAAO,SAASwB,cAAcpB,EAAgB,EAAEiB,KAAkB,EAAEI,KAAyB;IAC3F,IAAI,CAACA,OAAO,OAAOJ;IACnB,IAAIK;IACJ,IAAI;QACFA,YAAYtB,GAAGI,OAAO,CAAC,CAAC,wCAAwC,EAAEiB,MAAM,CAAC,CAAC,EAAEhB,GAAG;IACjF,EAAE,OAAOkB,KAAK;QACZ,MAAMjC,WAAWU,IAAIuB;IACvB;IACA,MAAMC,aAAa,IAAIhC,IAAI8B,UAAUN,GAAG,CAAC,CAACV,IAAMA,EAAEZ,IAAI;IACtD,OAAO,IAAIF,IAAI;WAAIyB;KAAM,CAACC,MAAM,CAAC,CAACC,IAAMK,WAAWjB,GAAG,CAACY;AACzD;AAEA,4FAA4F;AAC5F,kDAAkD;AAClD,OAAO,SAASjB,YAAYF,EAAgB,EAAEC,GAAmB,EAAEQ,SAA0B;IAC3F,MAAME,YAAYtB,cAAcY,KAAKQ;IACrC,OAAOW,cAAcpB,IAAIQ,SAASR,IAAIC,KAAKQ,YAAYE,UAAUU,KAAK;AACxE;AAEA,8FAA8F;AAC9F,yFAAyF;AACzF,OAAO,SAASI,iBAAiBzB,EAAgB,EAAE0B,KAAa,EAAET,KAAkB;IAClFjB,GAAG2B,IAAI,CAAC,CAAC,qBAAqB,EAAED,OAAO;IACvC1B,GAAG2B,IAAI,CAAC,CAAC,kBAAkB,EAAED,MAAM,0BAA0B,CAAC;IAC9D1B,GAAGI,OAAO,CAAC,CAAC,YAAY,EAAEsB,MAAM,yCAAyC,CAAC,EAAEE,GAAG,CAACC,KAAKC,SAAS,CAAC;WAAIb;KAAM;AAC3G;AAEA,OAAO,SAASc,cAAc/B,EAAgB,EAAEiB,KAAkB;IAChEQ,iBAAiBzB,IAAI,cAAciB;AACrC"}
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
import type { DatabaseSync } from 'node:sqlite';
|
|
2
2
|
import type { ResolvedConfig } from '../config/index.js';
|
|
3
|
-
import type { Row } from '../output.js';
|
|
3
|
+
import type { Row } from '../output/output.js';
|
|
4
4
|
export interface SearchOptions {
|
|
5
5
|
k?: number;
|
|
6
6
|
where?: string;
|
|
@@ -3,12 +3,12 @@ import { join } from 'node:path';
|
|
|
3
3
|
import { contentTokenize, embedConfig, featureEnabled, resolveSearch } from '../config/index.js';
|
|
4
4
|
import { localModelMissing, MODEL_FILENAMES } from '../embed/store.js';
|
|
5
5
|
import { SenseError } from '../errors.js';
|
|
6
|
-
import { searchError } from '../search-error.js';
|
|
7
|
-
import { segmentMatch } from '../segment.js';
|
|
6
|
+
import { searchError } from '../output/search-error.js';
|
|
7
|
+
import { segmentMatch } from '../text/segment.js';
|
|
8
8
|
import { materializeScope, narrowByWhere, rawScope, scopeHasEmbeddings } from './scope.js';
|
|
9
9
|
import { linksCandidates, vectorsCandidates, wordsCandidates } from './signals.js';
|
|
10
10
|
// snippet() re-tokenizes each candidate doc, superlinearly: ~10s for one 1MB doc
|
|
11
|
-
// (
|
|
11
|
+
// (benchmark/reports/2026-08-23-hub-release-battery.md). Past this bound, rows get the JS excerpt.
|
|
12
12
|
const EXCERPT_WINDOW = 160;
|
|
13
13
|
// Bare terms from an FTS5 query string: strips operators/quoting so the oversized-doc excerpt
|
|
14
14
|
// scan matches the same words the query matched on, not FTS5 syntax.
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/search.ts"],"sourcesContent":["import { readFileSync } from 'node:fs';\nimport { join } from 'node:path';\nimport type { DatabaseSync } from 'node:sqlite';\nimport type { ResolvedConfig } from '../config/index.ts';\nimport { contentTokenize, embedConfig, featureEnabled, resolveSearch } from '../config/index.ts';\nimport { localModelMissing, MODEL_FILENAMES } from '../embed/store.ts';\nimport { SenseError } from '../errors.ts';\nimport type { Row } from '../output.ts';\nimport { searchError } from '../search-error.ts';\nimport { segmentMatch } from '../segment.ts';\nimport { materializeScope, narrowByWhere, rawScope, scopeHasEmbeddings } from './scope.ts';\nimport { linksCandidates, vectorsCandidates, wordsCandidates } from './signals.ts';\n\n// snippet() re-tokenizes each candidate doc, superlinearly: ~10s for one 1MB doc\n// (BENCHMARKING.md). Past this bound, rows get the linear JS excerpt instead.\nconst EXCERPT_WINDOW = 160;\n\n// Bare terms from an FTS5 query string: strips operators/quoting so the oversized-doc excerpt\n// scan matches the same words the query matched on, not FTS5 syntax.\nfunction extractBareTerms(query: string): string[] {\n const cleaned = query\n .replace(/\"/g, ' ')\n .replace(/[()*]/g, ' ')\n .replace(/\\b(AND|OR|NOT|NEAR)\\b(\\/\\d+)?/gi, ' ');\n return cleaned\n .split(/\\s+/)\n .map((tok) => tok.replace(/^[A-Za-z_]\\w*:/, '')) // column filter, e.g. title:term\n .map((tok) => tok.toLowerCase().trim())\n .filter((tok) => tok.length > 0);\n}\n\nfunction findOccurrences(haystackLower: string, terms: string[]): Array<{ start: number; end: number; term: string }> {\n const occ: Array<{ start: number; end: number; term: string }> = [];\n for (const term of terms) {\n let idx = 0;\n for (;;) {\n const found = haystackLower.indexOf(term, idx);\n if (found === -1) break;\n occ.push({ start: found, end: found + term.length, term });\n idx = found + term.length;\n }\n }\n // Longest span first at equal start, so \"tests\" beats its substring \"test\" and the\n // whole word gets highlighted; the emit loop then absorbs the shorter overlap.\n return occ.sort((a, b) => a.start - b.start || b.end - a.end);\n}\n\n// Slides a window over the sorted occurrences, same semantics as FTS5's own best-window\n// pick: most distinct terms, ties broken by most total hits.\nfunction bestWindowStart(occ: Array<{ start: number; end: number; term: string }>, windowSize: number): number {\n let best = occ[0].start;\n let bestDistinct = 0;\n let bestCount = 0;\n for (let i = 0; i < occ.length; i++) {\n const winEnd = occ[i].start + windowSize;\n const seen = new Set<string>();\n let count = 0;\n for (let j = i; j < occ.length && occ[j].start < winEnd; j++) {\n seen.add(occ[j].term);\n count++;\n }\n if (seen.size > bestDistinct || (seen.size === bestDistinct && count > bestCount)) {\n best = occ[i].start;\n bestDistinct = seen.size;\n bestCount = count;\n }\n }\n return best;\n}\n\n// snippet()-shaped excerpt: matched terms wrapped in «», … at cut edges. Linear in doc\n// length, only run for rows actually returned (<= k), unlike snippet()'s per-row cost.\nfunction computeExcerpt(text: string, terms: string[]): { excerpt: string; offset: number } {\n const occ = findOccurrences(text.toLowerCase(), terms);\n if (occ.length === 0) {\n // This is a raw substring scan, not porter-stemmed: a doc matched only through a\n // stemmed variant (query \"negotiate\" vs doc \"negotiating\") finds no occurrence here.\n // Fall back to the doc's start, unmarked, rather than claim a match that isn't there.\n const end = Math.min(text.length, EXCERPT_WINDOW);\n return { excerpt: `${text.slice(0, end).replace(/\\s+/g, ' ').trim()}${end < text.length ? '…' : ''}`, offset: 0 };\n }\n const start = bestWindowStart(occ, EXCERPT_WINDOW);\n const end = Math.min(text.length, start + EXCERPT_WINDOW);\n let out = '';\n let cursor = start;\n for (const o of occ) {\n if (o.start < start || o.end > end) continue;\n // Occurrences of duplicate or substring-overlapping terms (\"test tests\") can overlap;\n // emitting each would duplicate document text. Keep the first, absorb the rest.\n if (o.start < cursor) continue;\n out += `${text.slice(cursor, o.start)}«${text.slice(o.start, o.end)}»`;\n cursor = o.end;\n }\n out += text.slice(cursor, end);\n const prefix = start > 0 ? '…' : '';\n const suffix = end < text.length ? '…' : '';\n return { excerpt: `${prefix}${out.replace(/\\s+/g, ' ')}${suffix}`, offset: start };\n}\n\nfunction lineNumberAt(text: string, offset: number): number {\n let line = 1;\n for (let i = 0; i < offset; i++) if (text.charCodeAt(i) === 10) line++;\n return line;\n}\n\nfunction lineRangeFor(db: DatabaseSync, path: string, line: number): string | null {\n const row = db.prepare('SELECT start_line, end_line FROM sections WHERE \"path\" = ? AND start_line <= ? AND end_line >= ? ORDER BY start_line DESC LIMIT 1').get(path, line, line) as { start_line: number; end_line: number } | undefined;\n return row ? `L${row.start_line}-${row.end_line}` : null;\n}\n\nexport interface SearchOptions {\n k?: number;\n where?: string; // SQL fragment against frontmatter alias `f`, e.g. \"f.status = 'active'\"\n preset?: string; // named preset; unknown name throws listing declared presets, undefined -> \"default\"\n include?: string[]; // ad hoc scope override (repeatable --include); independent of exclude\n exclude?: string[]; // ad hoc scope override (repeatable --exclude); independent of include\n noExclude?: boolean; // --no-exclude: drop the preset's exclude for this command\n}\n\n// The declared or defaulted signals compose via RRF; `via` names which ones produced each row.\n// `opts` arrives already resolved (config.ts:resolveSearch).\nexport async function search(db: DatabaseSync, cfg: ResolvedConfig, terms: string, opts: SearchOptions = {}): Promise<Row[]> {\n const effective = resolveSearch(cfg, opts);\n const { k, signals } = effective;\n\n const allPaths = (db.prepare('SELECT \"path\" FROM frontmatter').all() as Array<{ path: string }>).map((r) => r.path);\n const scopePaths = rawScope(db, cfg, opts, allPaths);\n const scopeActive = scopePaths.size < allPaths.length;\n // The set every candidate pool must be filtered to before truncation: scope narrowed by\n // --where, the same composition scopedPaths() gives the other commands. Runs the same where\n // fragment matchSql does, so a bad column gets the same attributed error either way.\n let allowedPaths: Set<string>;\n try {\n allowedPaths = narrowByWhere(db, scopePaths, effective.where);\n } catch (err) {\n throw searchError(err as Error, terms, effective.where);\n }\n const fetch = Math.max(k * 3, 30);\n\n // A downloadable HF id proceeds -- getProvider fetches it lazily on consent. Only a\n // local path with missing files errors here, since nothing will ever fetch it for itself.\n const wantsVectors = signals.vectors !== undefined;\n if (wantsVectors) {\n const e = embedConfig(cfg); // validate.ts guarantees this is set whenever \"vectors\" is declared\n if (localModelMissing(e)) {\n throw new SenseError('EMBED_MODEL_MISSING', `preset \"${effective.presetName}\" searches with vectors, but the local model path \"${e.model}\" is missing ${MODEL_FILENAMES}; point embed.model at a directory containing them, or drop \"vectors\" from that preset's signals to search without them`);\n }\n }\n const semanticEnabled = wantsVectors && scopeHasEmbeddings(db, cfg, allowedPaths);\n\n // Terms pass to MATCH as written, with one transform: a run of text whose language marks no\n // word boundaries becomes a quoted grapheme phrase (and a title:/summary:/text: qualifier\n // ahead of it retargets its _seg column), matching how it was indexed. Text that already\n // carries boundaries comes through untouched, so this is a no-op for the languages that\n // never needed it. Gated on the same predicate that decided whether the sidecars were\n // populated (contentTokenize(cfg) === undefined), so index and query can't disagree. Invalid\n // syntax still errors rather than being rewritten.\n // --where\n // applies inside the candidate query (a post-filter would drop matches ranked past the pool)\n // and again on the final select, for link-derived rows.\n const scope = effective.where;\n const whereJoin = scope ? `JOIN frontmatter f ON f.\"path\" = content.path` : '';\n const whereCond = scope ? `AND (${scope})` : '';\n // A scope narrower than the whole index must filter the candidate pool before LIMIT, not\n // after -- otherwise scoped notes ranking below the global top-`fetch` never reach the\n // filter. A join against a temp table, not a bound parameter list: real scopes run to\n // thousands of paths, past SQLITE_MAX_VARIABLE_NUMBER on older builds.\n if (scopeActive) materializeScope(db, '_search_scope', scopePaths);\n const scopeCond = scopeActive ? `AND content.path IN (SELECT \"path\" FROM _search_scope)` : '';\n const query = contentTokenize(cfg) === undefined ? segmentMatch(terms) : terms;\n\n const candidates = new Map<string, { score: number; via: string }>();\n let matchRows: Array<{ path: string; hit: string | null }> = [];\n // A query that rewrites to nothing (e.g. only unspaced-script punctuation) has no\n // searchable words: zero lexical rows, and the other signals still compose.\n if (signals.words !== undefined && query.trim() !== '') {\n try {\n matchRows = wordsCandidates(db, candidates, query, whereJoin, whereCond, scopeCond, fetch, signals.words);\n } catch (err) {\n throw searchError(err as Error, terms, scope);\n }\n }\n const hits = new Map(matchRows.map((r) => [r.path, r.hit]));\n // hit === null here means the bound suppressed snippet(), not \"no match\" -- distinguish\n // from via='link' rows (never in matchRows, so absent from this set) below.\n const oversized = new Set(matchRows.filter((r) => r.hit === null).map((r) => r.path));\n\n if (signals.links !== undefined) linksCandidates(db, candidates, matchRows, allPaths, allowedPaths, fetch, signals.links);\n\n let chunkLines = new Map<string, string>();\n let chunkSimilarity = new Map<string, number>();\n if (semanticEnabled) {\n ({ chunkLines, chunkSimilarity } = await vectorsCandidates(db, cfg, candidates, terms, fetch, allowedPaths, signals.vectors as number));\n }\n\n db.exec('DROP TABLE IF EXISTS _search');\n db.exec('CREATE TEMP TABLE _search (\"path\" TEXT PRIMARY KEY, score REAL, via TEXT, hit TEXT, lines TEXT, similarity REAL)');\n const insert = db.prepare('INSERT INTO _search (\"path\", score, via, hit, lines, similarity) VALUES (?, ?, ?, ?, ?, ?)');\n for (const [path, c] of candidates) insert.run(path, c.score, c.via, hits.get(path) ?? null, chunkLines.get(path) ?? null, chunkSimilarity.get(path) ?? null);\n\n // All three candidate paths (match, link, vector) already filtered to scope+where before\n // reaching _search; this reapplies --where anyway since the join to frontmatter is already\n // needed for the path column.\n const where = scope ? `WHERE (${scope})` : '';\n // lines is always present now: semantic rows carry their chunk's range, oversized-doc\n // lexical rows gain one below, everything else stays null. similarity stays semantic-only.\n const similarityCol = semanticEnabled ? ', _search.similarity' : '';\n const rows = db\n .prepare(\n `SELECT f.\"path\" AS path, content.title, content.summary, _search.hit, _search.via, round(_search.score, 4) AS score, _search.lines${similarityCol}\n FROM _search JOIN frontmatter f ON f.\"path\" = _search.\"path\" JOIN content ON content.path = _search.\"path\"\n ${where} ORDER BY _search.score DESC LIMIT ?`\n )\n .all(k) as Row[];\n\n if (oversized.size > 0) {\n const bareTerms = extractBareTerms(terms);\n for (const row of rows) {\n if (row.hit !== null || !oversized.has(row.path as string)) continue;\n let text: string;\n try {\n text = readFileSync(join(cfg.baseDir, row.path as string), 'utf8');\n } catch {\n continue; // vanished since the match; leave hit/lines null rather than throw\n }\n const { excerpt, offset } = computeExcerpt(text, bareTerms);\n row.hit = excerpt;\n if (row.lines == null) row.lines = featureEnabled(cfg, 'sections') ? lineRangeFor(db, row.path as string, lineNumberAt(text, offset)) : null;\n }\n }\n\n return rows;\n}\n"],"names":["readFileSync","join","contentTokenize","embedConfig","featureEnabled","resolveSearch","localModelMissing","MODEL_FILENAMES","SenseError","searchError","segmentMatch","materializeScope","narrowByWhere","rawScope","scopeHasEmbeddings","linksCandidates","vectorsCandidates","wordsCandidates","EXCERPT_WINDOW","extractBareTerms","query","cleaned","replace","split","map","tok","toLowerCase","trim","filter","length","findOccurrences","haystackLower","terms","occ","term","idx","found","indexOf","push","start","end","sort","a","b","bestWindowStart","windowSize","best","bestDistinct","bestCount","i","winEnd","seen","Set","count","j","add","size","computeExcerpt","text","Math","min","excerpt","slice","offset","out","cursor","o","prefix","suffix","lineNumberAt","line","charCodeAt","lineRangeFor","db","path","row","prepare","get","start_line","end_line","search","cfg","opts","hits","chunkLines","chunkSimilarity","effective","k","signals","allPaths","all","r","scopePaths","scopeActive","allowedPaths","where","err","fetch","max","wantsVectors","vectors","undefined","e","presetName","model","semanticEnabled","scope","whereJoin","whereCond","scopeCond","candidates","Map","matchRows","words","hit","oversized","links","exec","insert","c","run","score","via","similarityCol","rows","bareTerms","has","baseDir","lines"],"mappings":"AAAA,SAASA,YAAY,QAAQ,UAAU;AACvC,SAASC,IAAI,QAAQ,YAAY;AAGjC,SAASC,eAAe,EAAEC,WAAW,EAAEC,cAAc,EAAEC,aAAa,QAAQ,qBAAqB;AACjG,SAASC,iBAAiB,EAAEC,eAAe,QAAQ,oBAAoB;AACvE,SAASC,UAAU,QAAQ,eAAe;AAE1C,SAASC,WAAW,QAAQ,qBAAqB;AACjD,SAASC,YAAY,QAAQ,gBAAgB;AAC7C,SAASC,gBAAgB,EAAEC,aAAa,EAAEC,QAAQ,EAAEC,kBAAkB,QAAQ,aAAa;AAC3F,SAASC,eAAe,EAAEC,iBAAiB,EAAEC,eAAe,QAAQ,eAAe;AAEnF,iFAAiF;AACjF,8EAA8E;AAC9E,MAAMC,iBAAiB;AAEvB,8FAA8F;AAC9F,qEAAqE;AACrE,SAASC,iBAAiBC,KAAa;IACrC,MAAMC,UAAUD,MACbE,OAAO,CAAC,MAAM,KACdA,OAAO,CAAC,UAAU,KAClBA,OAAO,CAAC,mCAAmC;IAC9C,OAAOD,QACJE,KAAK,CAAC,OACNC,GAAG,CAAC,CAACC,MAAQA,IAAIH,OAAO,CAAC,kBAAkB,KAAK,iCAAiC;KACjFE,GAAG,CAAC,CAACC,MAAQA,IAAIC,WAAW,GAAGC,IAAI,IACnCC,MAAM,CAAC,CAACH,MAAQA,IAAII,MAAM,GAAG;AAClC;AAEA,SAASC,gBAAgBC,aAAqB,EAAEC,KAAe;IAC7D,MAAMC,MAA2D,EAAE;IACnE,KAAK,MAAMC,QAAQF,MAAO;QACxB,IAAIG,MAAM;QACV,OAAS;YACP,MAAMC,QAAQL,cAAcM,OAAO,CAACH,MAAMC;YAC1C,IAAIC,UAAU,CAAC,GAAG;YAClBH,IAAIK,IAAI,CAAC;gBAAEC,OAAOH;gBAAOI,KAAKJ,QAAQF,KAAKL,MAAM;gBAAEK;YAAK;YACxDC,MAAMC,QAAQF,KAAKL,MAAM;QAC3B;IACF;IACA,mFAAmF;IACnF,+EAA+E;IAC/E,OAAOI,IAAIQ,IAAI,CAAC,CAACC,GAAGC,IAAMD,EAAEH,KAAK,GAAGI,EAAEJ,KAAK,IAAII,EAAEH,GAAG,GAAGE,EAAEF,GAAG;AAC9D;AAEA,wFAAwF;AACxF,6DAA6D;AAC7D,SAASI,gBAAgBX,GAAwD,EAAEY,UAAkB;IACnG,IAAIC,OAAOb,GAAG,CAAC,EAAE,CAACM,KAAK;IACvB,IAAIQ,eAAe;IACnB,IAAIC,YAAY;IAChB,IAAK,IAAIC,IAAI,GAAGA,IAAIhB,IAAIJ,MAAM,EAAEoB,IAAK;QACnC,MAAMC,SAASjB,GAAG,CAACgB,EAAE,CAACV,KAAK,GAAGM;QAC9B,MAAMM,OAAO,IAAIC;QACjB,IAAIC,QAAQ;QACZ,IAAK,IAAIC,IAAIL,GAAGK,IAAIrB,IAAIJ,MAAM,IAAII,GAAG,CAACqB,EAAE,CAACf,KAAK,GAAGW,QAAQI,IAAK;YAC5DH,KAAKI,GAAG,CAACtB,GAAG,CAACqB,EAAE,CAACpB,IAAI;YACpBmB;QACF;QACA,IAAIF,KAAKK,IAAI,GAAGT,gBAAiBI,KAAKK,IAAI,KAAKT,gBAAgBM,QAAQL,WAAY;YACjFF,OAAOb,GAAG,CAACgB,EAAE,CAACV,KAAK;YACnBQ,eAAeI,KAAKK,IAAI;YACxBR,YAAYK;QACd;IACF;IACA,OAAOP;AACT;AAEA,uFAAuF;AACvF,uFAAuF;AACvF,SAASW,eAAeC,IAAY,EAAE1B,KAAe;IACnD,MAAMC,MAAMH,gBAAgB4B,KAAKhC,WAAW,IAAIM;IAChD,IAAIC,IAAIJ,MAAM,KAAK,GAAG;QACpB,iFAAiF;QACjF,qFAAqF;QACrF,sFAAsF;QACtF,MAAMW,MAAMmB,KAAKC,GAAG,CAACF,KAAK7B,MAAM,EAAEX;QAClC,OAAO;YAAE2C,SAAS,GAAGH,KAAKI,KAAK,CAAC,GAAGtB,KAAKlB,OAAO,CAAC,QAAQ,KAAKK,IAAI,KAAKa,MAAMkB,KAAK7B,MAAM,GAAG,MAAM,IAAI;YAAEkC,QAAQ;QAAE;IAClH;IACA,MAAMxB,QAAQK,gBAAgBX,KAAKf;IACnC,MAAMsB,MAAMmB,KAAKC,GAAG,CAACF,KAAK7B,MAAM,EAAEU,QAAQrB;IAC1C,IAAI8C,MAAM;IACV,IAAIC,SAAS1B;IACb,KAAK,MAAM2B,KAAKjC,IAAK;QACnB,IAAIiC,EAAE3B,KAAK,GAAGA,SAAS2B,EAAE1B,GAAG,GAAGA,KAAK;QACpC,sFAAsF;QACtF,gFAAgF;QAChF,IAAI0B,EAAE3B,KAAK,GAAG0B,QAAQ;QACtBD,OAAO,GAAGN,KAAKI,KAAK,CAACG,QAAQC,EAAE3B,KAAK,EAAE,CAAC,EAAEmB,KAAKI,KAAK,CAACI,EAAE3B,KAAK,EAAE2B,EAAE1B,GAAG,EAAE,CAAC,CAAC;QACtEyB,SAASC,EAAE1B,GAAG;IAChB;IACAwB,OAAON,KAAKI,KAAK,CAACG,QAAQzB;IAC1B,MAAM2B,SAAS5B,QAAQ,IAAI,MAAM;IACjC,MAAM6B,SAAS5B,MAAMkB,KAAK7B,MAAM,GAAG,MAAM;IACzC,OAAO;QAAEgC,SAAS,GAAGM,SAASH,IAAI1C,OAAO,CAAC,QAAQ,OAAO8C,QAAQ;QAAEL,QAAQxB;IAAM;AACnF;AAEA,SAAS8B,aAAaX,IAAY,EAAEK,MAAc;IAChD,IAAIO,OAAO;IACX,IAAK,IAAIrB,IAAI,GAAGA,IAAIc,QAAQd,IAAK,IAAIS,KAAKa,UAAU,CAACtB,OAAO,IAAIqB;IAChE,OAAOA;AACT;AAEA,SAASE,aAAaC,EAAgB,EAAEC,IAAY,EAAEJ,IAAY;IAChE,MAAMK,MAAMF,GAAGG,OAAO,CAAC,qIAAqIC,GAAG,CAACH,MAAMJ,MAAMA;IAC5K,OAAOK,MAAM,CAAC,CAAC,EAAEA,IAAIG,UAAU,CAAC,CAAC,EAAEH,IAAII,QAAQ,EAAE,GAAG;AACtD;AAWA,+FAA+F;AAC/F,6DAA6D;AAC7D,OAAO,eAAeC,OAAOP,EAAgB,EAAEQ,GAAmB,EAAEjD,KAAa,EAAEkD,OAAsB,CAAC,CAAC;QA6EpCC,WAAwBC,iBAA8BC;IA5E3H,MAAMC,YAAYjF,cAAc4E,KAAKC;IACrC,MAAM,EAAEK,CAAC,EAAEC,OAAO,EAAE,GAAGF;IAEvB,MAAMG,WAAW,AAAChB,GAAGG,OAAO,CAAC,kCAAkCc,GAAG,GAA+BlE,GAAG,CAAC,CAACmE,IAAMA,EAAEjB,IAAI;IAClH,MAAMkB,aAAa/E,SAAS4D,IAAIQ,KAAKC,MAAMO;IAC3C,MAAMI,cAAcD,WAAWpC,IAAI,GAAGiC,SAAS5D,MAAM;IACrD,wFAAwF;IACxF,4FAA4F;IAC5F,qFAAqF;IACrF,IAAIiE;IACJ,IAAI;QACFA,eAAelF,cAAc6D,IAAImB,YAAYN,UAAUS,KAAK;IAC9D,EAAE,OAAOC,KAAK;QACZ,MAAMvF,YAAYuF,KAAchE,OAAOsD,UAAUS,KAAK;IACxD;IACA,MAAME,QAAQtC,KAAKuC,GAAG,CAACX,IAAI,GAAG;IAE9B,oFAAoF;IACpF,0FAA0F;IAC1F,MAAMY,eAAeX,QAAQY,OAAO,KAAKC;IACzC,IAAIF,cAAc;QAChB,MAAMG,IAAInG,YAAY8E,MAAM,oEAAoE;QAChG,IAAI3E,kBAAkBgG,IAAI;YACxB,MAAM,IAAI9F,WAAW,uBAAuB,CAAC,QAAQ,EAAE8E,UAAUiB,UAAU,CAAC,mDAAmD,EAAED,EAAEE,KAAK,CAAC,aAAa,EAAEjG,gBAAgB,uHAAuH,CAAC;QAClS;IACF;IACA,MAAMkG,kBAAkBN,gBAAgBrF,mBAAmB2D,IAAIQ,KAAKa;IAEpE,4FAA4F;IAC5F,0FAA0F;IAC1F,yFAAyF;IACzF,wFAAwF;IACxF,sFAAsF;IACtF,6FAA6F;IAC7F,mDAAmD;IACnD,UAAU;IACV,6FAA6F;IAC7F,wDAAwD;IACxD,MAAMY,QAAQpB,UAAUS,KAAK;IAC7B,MAAMY,YAAYD,QAAQ,CAAC,6CAA6C,CAAC,GAAG;IAC5E,MAAME,YAAYF,QAAQ,CAAC,KAAK,EAAEA,MAAM,CAAC,CAAC,GAAG;IAC7C,yFAAyF;IACzF,uFAAuF;IACvF,sFAAsF;IACtF,uEAAuE;IACvE,IAAIb,aAAalF,iBAAiB8D,IAAI,iBAAiBmB;IACvD,MAAMiB,YAAYhB,cAAc,CAAC,sDAAsD,CAAC,GAAG;IAC3F,MAAMzE,QAAQlB,gBAAgB+E,SAASoB,YAAY3F,aAAasB,SAASA;IAEzE,MAAM8E,aAAa,IAAIC;IACvB,IAAIC,YAAyD,EAAE;IAC/D,kFAAkF;IAClF,4EAA4E;IAC5E,IAAIxB,QAAQyB,KAAK,KAAKZ,aAAajF,MAAMO,IAAI,OAAO,IAAI;QACtD,IAAI;YACFqF,YAAY/F,gBAAgBwD,IAAIqC,YAAY1F,OAAOuF,WAAWC,WAAWC,WAAWZ,OAAOT,QAAQyB,KAAK;QAC1G,EAAE,OAAOjB,KAAK;YACZ,MAAMvF,YAAYuF,KAAchE,OAAO0E;QACzC;IACF;IACA,MAAMvB,OAAO,IAAI4B,IAAIC,UAAUxF,GAAG,CAAC,CAACmE,IAAM;YAACA,EAAEjB,IAAI;YAAEiB,EAAEuB,GAAG;SAAC;IACzD,wFAAwF;IACxF,4EAA4E;IAC5E,MAAMC,YAAY,IAAI/D,IAAI4D,UAAUpF,MAAM,CAAC,CAAC+D,IAAMA,EAAEuB,GAAG,KAAK,MAAM1F,GAAG,CAAC,CAACmE,IAAMA,EAAEjB,IAAI;IAEnF,IAAIc,QAAQ4B,KAAK,KAAKf,WAAWtF,gBAAgB0D,IAAIqC,YAAYE,WAAWvB,UAAUK,cAAcG,OAAOT,QAAQ4B,KAAK;IAExH,IAAIhC,aAAa,IAAI2B;IACrB,IAAI1B,kBAAkB,IAAI0B;IAC1B,IAAIN,iBAAiB;QAClB,CAAA,EAAErB,UAAU,EAAEC,eAAe,EAAE,GAAG,MAAMrE,kBAAkByD,IAAIQ,KAAK6B,YAAY9E,OAAOiE,OAAOH,cAAcN,QAAQY,OAAO,CAAU;IACvI;IAEA3B,GAAG4C,IAAI,CAAC;IACR5C,GAAG4C,IAAI,CAAC;IACR,MAAMC,SAAS7C,GAAGG,OAAO,CAAC;IAC1B,KAAK,MAAM,CAACF,MAAM6C,EAAE,IAAIT,WAAYQ,OAAOE,GAAG,CAAC9C,MAAM6C,EAAEE,KAAK,EAAEF,EAAEG,GAAG,GAAEvC,YAAAA,KAAKN,GAAG,CAACH,mBAATS,uBAAAA,YAAkB,OAAMC,kBAAAA,WAAWP,GAAG,CAACH,mBAAfU,6BAAAA,kBAAwB,OAAMC,uBAAAA,gBAAgBR,GAAG,CAACH,mBAApBW,kCAAAA,uBAA6B;IAExJ,yFAAyF;IACzF,2FAA2F;IAC3F,8BAA8B;IAC9B,MAAMU,QAAQW,QAAQ,CAAC,OAAO,EAAEA,MAAM,CAAC,CAAC,GAAG;IAC3C,sFAAsF;IACtF,2FAA2F;IAC3F,MAAMiB,gBAAgBlB,kBAAkB,yBAAyB;IACjE,MAAMmB,OAAOnD,GACVG,OAAO,CACN,CAAC,kIAAkI,EAAE+C,cAAc;;OAElJ,EAAE5B,MAAM,oCAAoC,CAAC,EAE/CL,GAAG,CAACH;IAEP,IAAI4B,UAAU3D,IAAI,GAAG,GAAG;QACtB,MAAMqE,YAAY1G,iBAAiBa;QACnC,KAAK,MAAM2C,OAAOiD,KAAM;YACtB,IAAIjD,IAAIuC,GAAG,KAAK,QAAQ,CAACC,UAAUW,GAAG,CAACnD,IAAID,IAAI,GAAa;YAC5D,IAAIhB;YACJ,IAAI;gBACFA,OAAO1D,aAAaC,KAAKgF,IAAI8C,OAAO,EAAEpD,IAAID,IAAI,GAAa;YAC7D,EAAE,OAAM;gBACN,UAAU,mEAAmE;YAC/E;YACA,MAAM,EAAEb,OAAO,EAAEE,MAAM,EAAE,GAAGN,eAAeC,MAAMmE;YACjDlD,IAAIuC,GAAG,GAAGrD;YACV,IAAIc,IAAIqD,KAAK,IAAI,MAAMrD,IAAIqD,KAAK,GAAG5H,eAAe6E,KAAK,cAAcT,aAAaC,IAAIE,IAAID,IAAI,EAAYL,aAAaX,MAAMK,WAAW;QAC1I;IACF;IAEA,OAAO6D;AACT"}
|
|
1
|
+
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/search.ts"],"sourcesContent":["import { readFileSync } from 'node:fs';\nimport { join } from 'node:path';\nimport type { DatabaseSync } from 'node:sqlite';\nimport type { ResolvedConfig } from '../config/index.ts';\nimport { contentTokenize, embedConfig, featureEnabled, resolveSearch } from '../config/index.ts';\nimport { localModelMissing, MODEL_FILENAMES } from '../embed/store.ts';\nimport { SenseError } from '../errors.ts';\nimport type { Row } from '../output/output.ts';\nimport { searchError } from '../output/search-error.ts';\nimport { segmentMatch } from '../text/segment.ts';\nimport { materializeScope, narrowByWhere, rawScope, scopeHasEmbeddings } from './scope.ts';\nimport { linksCandidates, vectorsCandidates, wordsCandidates } from './signals.ts';\n\n// snippet() re-tokenizes each candidate doc, superlinearly: ~10s for one 1MB doc\n// (benchmark/reports/2026-08-23-hub-release-battery.md). Past this bound, rows get the JS excerpt.\nconst EXCERPT_WINDOW = 160;\n\n// Bare terms from an FTS5 query string: strips operators/quoting so the oversized-doc excerpt\n// scan matches the same words the query matched on, not FTS5 syntax.\nfunction extractBareTerms(query: string): string[] {\n const cleaned = query\n .replace(/\"/g, ' ')\n .replace(/[()*]/g, ' ')\n .replace(/\\b(AND|OR|NOT|NEAR)\\b(\\/\\d+)?/gi, ' ');\n return cleaned\n .split(/\\s+/)\n .map((tok) => tok.replace(/^[A-Za-z_]\\w*:/, '')) // column filter, e.g. title:term\n .map((tok) => tok.toLowerCase().trim())\n .filter((tok) => tok.length > 0);\n}\n\nfunction findOccurrences(haystackLower: string, terms: string[]): Array<{ start: number; end: number; term: string }> {\n const occ: Array<{ start: number; end: number; term: string }> = [];\n for (const term of terms) {\n let idx = 0;\n for (;;) {\n const found = haystackLower.indexOf(term, idx);\n if (found === -1) break;\n occ.push({ start: found, end: found + term.length, term });\n idx = found + term.length;\n }\n }\n // Longest span first at equal start, so \"tests\" beats its substring \"test\" and the\n // whole word gets highlighted; the emit loop then absorbs the shorter overlap.\n return occ.sort((a, b) => a.start - b.start || b.end - a.end);\n}\n\n// Slides a window over the sorted occurrences, same semantics as FTS5's own best-window\n// pick: most distinct terms, ties broken by most total hits.\nfunction bestWindowStart(occ: Array<{ start: number; end: number; term: string }>, windowSize: number): number {\n let best = occ[0].start;\n let bestDistinct = 0;\n let bestCount = 0;\n for (let i = 0; i < occ.length; i++) {\n const winEnd = occ[i].start + windowSize;\n const seen = new Set<string>();\n let count = 0;\n for (let j = i; j < occ.length && occ[j].start < winEnd; j++) {\n seen.add(occ[j].term);\n count++;\n }\n if (seen.size > bestDistinct || (seen.size === bestDistinct && count > bestCount)) {\n best = occ[i].start;\n bestDistinct = seen.size;\n bestCount = count;\n }\n }\n return best;\n}\n\n// snippet()-shaped excerpt: matched terms wrapped in «», … at cut edges. Linear in doc\n// length, only run for rows actually returned (<= k), unlike snippet()'s per-row cost.\nfunction computeExcerpt(text: string, terms: string[]): { excerpt: string; offset: number } {\n const occ = findOccurrences(text.toLowerCase(), terms);\n if (occ.length === 0) {\n // This is a raw substring scan, not porter-stemmed: a doc matched only through a\n // stemmed variant (query \"negotiate\" vs doc \"negotiating\") finds no occurrence here.\n // Fall back to the doc's start, unmarked, rather than claim a match that isn't there.\n const end = Math.min(text.length, EXCERPT_WINDOW);\n return { excerpt: `${text.slice(0, end).replace(/\\s+/g, ' ').trim()}${end < text.length ? '…' : ''}`, offset: 0 };\n }\n const start = bestWindowStart(occ, EXCERPT_WINDOW);\n const end = Math.min(text.length, start + EXCERPT_WINDOW);\n let out = '';\n let cursor = start;\n for (const o of occ) {\n if (o.start < start || o.end > end) continue;\n // Occurrences of duplicate or substring-overlapping terms (\"test tests\") can overlap;\n // emitting each would duplicate document text. Keep the first, absorb the rest.\n if (o.start < cursor) continue;\n out += `${text.slice(cursor, o.start)}«${text.slice(o.start, o.end)}»`;\n cursor = o.end;\n }\n out += text.slice(cursor, end);\n const prefix = start > 0 ? '…' : '';\n const suffix = end < text.length ? '…' : '';\n return { excerpt: `${prefix}${out.replace(/\\s+/g, ' ')}${suffix}`, offset: start };\n}\n\nfunction lineNumberAt(text: string, offset: number): number {\n let line = 1;\n for (let i = 0; i < offset; i++) if (text.charCodeAt(i) === 10) line++;\n return line;\n}\n\nfunction lineRangeFor(db: DatabaseSync, path: string, line: number): string | null {\n const row = db.prepare('SELECT start_line, end_line FROM sections WHERE \"path\" = ? AND start_line <= ? AND end_line >= ? ORDER BY start_line DESC LIMIT 1').get(path, line, line) as { start_line: number; end_line: number } | undefined;\n return row ? `L${row.start_line}-${row.end_line}` : null;\n}\n\nexport interface SearchOptions {\n k?: number;\n where?: string; // SQL fragment against frontmatter alias `f`, e.g. \"f.status = 'active'\"\n preset?: string; // named preset; unknown name throws listing declared presets, undefined -> \"default\"\n include?: string[]; // ad hoc scope override (repeatable --include); independent of exclude\n exclude?: string[]; // ad hoc scope override (repeatable --exclude); independent of include\n noExclude?: boolean; // --no-exclude: drop the preset's exclude for this command\n}\n\n// The declared or defaulted signals compose via RRF; `via` names which ones produced each row.\n// `opts` arrives already resolved (config.ts:resolveSearch).\nexport async function search(db: DatabaseSync, cfg: ResolvedConfig, terms: string, opts: SearchOptions = {}): Promise<Row[]> {\n const effective = resolveSearch(cfg, opts);\n const { k, signals } = effective;\n\n const allPaths = (db.prepare('SELECT \"path\" FROM frontmatter').all() as Array<{ path: string }>).map((r) => r.path);\n const scopePaths = rawScope(db, cfg, opts, allPaths);\n const scopeActive = scopePaths.size < allPaths.length;\n // The set every candidate pool must be filtered to before truncation: scope narrowed by\n // --where, the same composition scopedPaths() gives the other commands. Runs the same where\n // fragment matchSql does, so a bad column gets the same attributed error either way.\n let allowedPaths: Set<string>;\n try {\n allowedPaths = narrowByWhere(db, scopePaths, effective.where);\n } catch (err) {\n throw searchError(err as Error, terms, effective.where);\n }\n const fetch = Math.max(k * 3, 30);\n\n // A downloadable HF id proceeds -- getProvider fetches it lazily on consent. Only a\n // local path with missing files errors here, since nothing will ever fetch it for itself.\n const wantsVectors = signals.vectors !== undefined;\n if (wantsVectors) {\n const e = embedConfig(cfg); // validate.ts guarantees this is set whenever \"vectors\" is declared\n if (localModelMissing(e)) {\n throw new SenseError('EMBED_MODEL_MISSING', `preset \"${effective.presetName}\" searches with vectors, but the local model path \"${e.model}\" is missing ${MODEL_FILENAMES}; point embed.model at a directory containing them, or drop \"vectors\" from that preset's signals to search without them`);\n }\n }\n const semanticEnabled = wantsVectors && scopeHasEmbeddings(db, cfg, allowedPaths);\n\n // Terms pass to MATCH as written, with one transform: a run of text whose language marks no\n // word boundaries becomes a quoted grapheme phrase (and a title:/summary:/text: qualifier\n // ahead of it retargets its _seg column), matching how it was indexed. Text that already\n // carries boundaries comes through untouched, so this is a no-op for the languages that\n // never needed it. Gated on the same predicate that decided whether the sidecars were\n // populated (contentTokenize(cfg) === undefined), so index and query can't disagree. Invalid\n // syntax still errors rather than being rewritten.\n // --where\n // applies inside the candidate query (a post-filter would drop matches ranked past the pool)\n // and again on the final select, for link-derived rows.\n const scope = effective.where;\n const whereJoin = scope ? `JOIN frontmatter f ON f.\"path\" = content.path` : '';\n const whereCond = scope ? `AND (${scope})` : '';\n // A scope narrower than the whole index must filter the candidate pool before LIMIT, not\n // after -- otherwise scoped notes ranking below the global top-`fetch` never reach the\n // filter. A join against a temp table, not a bound parameter list: real scopes run to\n // thousands of paths, past SQLITE_MAX_VARIABLE_NUMBER on older builds.\n if (scopeActive) materializeScope(db, '_search_scope', scopePaths);\n const scopeCond = scopeActive ? `AND content.path IN (SELECT \"path\" FROM _search_scope)` : '';\n const query = contentTokenize(cfg) === undefined ? segmentMatch(terms) : terms;\n\n const candidates = new Map<string, { score: number; via: string }>();\n let matchRows: Array<{ path: string; hit: string | null }> = [];\n // A query that rewrites to nothing (e.g. only unspaced-script punctuation) has no\n // searchable words: zero lexical rows, and the other signals still compose.\n if (signals.words !== undefined && query.trim() !== '') {\n try {\n matchRows = wordsCandidates(db, candidates, query, whereJoin, whereCond, scopeCond, fetch, signals.words);\n } catch (err) {\n throw searchError(err as Error, terms, scope);\n }\n }\n const hits = new Map(matchRows.map((r) => [r.path, r.hit]));\n // hit === null here means the bound suppressed snippet(), not \"no match\" -- distinguish\n // from via='link' rows (never in matchRows, so absent from this set) below.\n const oversized = new Set(matchRows.filter((r) => r.hit === null).map((r) => r.path));\n\n if (signals.links !== undefined) linksCandidates(db, candidates, matchRows, allPaths, allowedPaths, fetch, signals.links);\n\n let chunkLines = new Map<string, string>();\n let chunkSimilarity = new Map<string, number>();\n if (semanticEnabled) {\n ({ chunkLines, chunkSimilarity } = await vectorsCandidates(db, cfg, candidates, terms, fetch, allowedPaths, signals.vectors as number));\n }\n\n db.exec('DROP TABLE IF EXISTS _search');\n db.exec('CREATE TEMP TABLE _search (\"path\" TEXT PRIMARY KEY, score REAL, via TEXT, hit TEXT, lines TEXT, similarity REAL)');\n const insert = db.prepare('INSERT INTO _search (\"path\", score, via, hit, lines, similarity) VALUES (?, ?, ?, ?, ?, ?)');\n for (const [path, c] of candidates) insert.run(path, c.score, c.via, hits.get(path) ?? null, chunkLines.get(path) ?? null, chunkSimilarity.get(path) ?? null);\n\n // All three candidate paths (match, link, vector) already filtered to scope+where before\n // reaching _search; this reapplies --where anyway since the join to frontmatter is already\n // needed for the path column.\n const where = scope ? `WHERE (${scope})` : '';\n // lines is always present now: semantic rows carry their chunk's range, oversized-doc\n // lexical rows gain one below, everything else stays null. similarity stays semantic-only.\n const similarityCol = semanticEnabled ? ', _search.similarity' : '';\n const rows = db\n .prepare(\n `SELECT f.\"path\" AS path, content.title, content.summary, _search.hit, _search.via, round(_search.score, 4) AS score, _search.lines${similarityCol}\n FROM _search JOIN frontmatter f ON f.\"path\" = _search.\"path\" JOIN content ON content.path = _search.\"path\"\n ${where} ORDER BY _search.score DESC LIMIT ?`\n )\n .all(k) as Row[];\n\n if (oversized.size > 0) {\n const bareTerms = extractBareTerms(terms);\n for (const row of rows) {\n if (row.hit !== null || !oversized.has(row.path as string)) continue;\n let text: string;\n try {\n text = readFileSync(join(cfg.baseDir, row.path as string), 'utf8');\n } catch {\n continue; // vanished since the match; leave hit/lines null rather than throw\n }\n const { excerpt, offset } = computeExcerpt(text, bareTerms);\n row.hit = excerpt;\n if (row.lines == null) row.lines = featureEnabled(cfg, 'sections') ? lineRangeFor(db, row.path as string, lineNumberAt(text, offset)) : null;\n }\n }\n\n return rows;\n}\n"],"names":["readFileSync","join","contentTokenize","embedConfig","featureEnabled","resolveSearch","localModelMissing","MODEL_FILENAMES","SenseError","searchError","segmentMatch","materializeScope","narrowByWhere","rawScope","scopeHasEmbeddings","linksCandidates","vectorsCandidates","wordsCandidates","EXCERPT_WINDOW","extractBareTerms","query","cleaned","replace","split","map","tok","toLowerCase","trim","filter","length","findOccurrences","haystackLower","terms","occ","term","idx","found","indexOf","push","start","end","sort","a","b","bestWindowStart","windowSize","best","bestDistinct","bestCount","i","winEnd","seen","Set","count","j","add","size","computeExcerpt","text","Math","min","excerpt","slice","offset","out","cursor","o","prefix","suffix","lineNumberAt","line","charCodeAt","lineRangeFor","db","path","row","prepare","get","start_line","end_line","search","cfg","opts","hits","chunkLines","chunkSimilarity","effective","k","signals","allPaths","all","r","scopePaths","scopeActive","allowedPaths","where","err","fetch","max","wantsVectors","vectors","undefined","e","presetName","model","semanticEnabled","scope","whereJoin","whereCond","scopeCond","candidates","Map","matchRows","words","hit","oversized","links","exec","insert","c","run","score","via","similarityCol","rows","bareTerms","has","baseDir","lines"],"mappings":"AAAA,SAASA,YAAY,QAAQ,UAAU;AACvC,SAASC,IAAI,QAAQ,YAAY;AAGjC,SAASC,eAAe,EAAEC,WAAW,EAAEC,cAAc,EAAEC,aAAa,QAAQ,qBAAqB;AACjG,SAASC,iBAAiB,EAAEC,eAAe,QAAQ,oBAAoB;AACvE,SAASC,UAAU,QAAQ,eAAe;AAE1C,SAASC,WAAW,QAAQ,4BAA4B;AACxD,SAASC,YAAY,QAAQ,qBAAqB;AAClD,SAASC,gBAAgB,EAAEC,aAAa,EAAEC,QAAQ,EAAEC,kBAAkB,QAAQ,aAAa;AAC3F,SAASC,eAAe,EAAEC,iBAAiB,EAAEC,eAAe,QAAQ,eAAe;AAEnF,iFAAiF;AACjF,mGAAmG;AACnG,MAAMC,iBAAiB;AAEvB,8FAA8F;AAC9F,qEAAqE;AACrE,SAASC,iBAAiBC,KAAa;IACrC,MAAMC,UAAUD,MACbE,OAAO,CAAC,MAAM,KACdA,OAAO,CAAC,UAAU,KAClBA,OAAO,CAAC,mCAAmC;IAC9C,OAAOD,QACJE,KAAK,CAAC,OACNC,GAAG,CAAC,CAACC,MAAQA,IAAIH,OAAO,CAAC,kBAAkB,KAAK,iCAAiC;KACjFE,GAAG,CAAC,CAACC,MAAQA,IAAIC,WAAW,GAAGC,IAAI,IACnCC,MAAM,CAAC,CAACH,MAAQA,IAAII,MAAM,GAAG;AAClC;AAEA,SAASC,gBAAgBC,aAAqB,EAAEC,KAAe;IAC7D,MAAMC,MAA2D,EAAE;IACnE,KAAK,MAAMC,QAAQF,MAAO;QACxB,IAAIG,MAAM;QACV,OAAS;YACP,MAAMC,QAAQL,cAAcM,OAAO,CAACH,MAAMC;YAC1C,IAAIC,UAAU,CAAC,GAAG;YAClBH,IAAIK,IAAI,CAAC;gBAAEC,OAAOH;gBAAOI,KAAKJ,QAAQF,KAAKL,MAAM;gBAAEK;YAAK;YACxDC,MAAMC,QAAQF,KAAKL,MAAM;QAC3B;IACF;IACA,mFAAmF;IACnF,+EAA+E;IAC/E,OAAOI,IAAIQ,IAAI,CAAC,CAACC,GAAGC,IAAMD,EAAEH,KAAK,GAAGI,EAAEJ,KAAK,IAAII,EAAEH,GAAG,GAAGE,EAAEF,GAAG;AAC9D;AAEA,wFAAwF;AACxF,6DAA6D;AAC7D,SAASI,gBAAgBX,GAAwD,EAAEY,UAAkB;IACnG,IAAIC,OAAOb,GAAG,CAAC,EAAE,CAACM,KAAK;IACvB,IAAIQ,eAAe;IACnB,IAAIC,YAAY;IAChB,IAAK,IAAIC,IAAI,GAAGA,IAAIhB,IAAIJ,MAAM,EAAEoB,IAAK;QACnC,MAAMC,SAASjB,GAAG,CAACgB,EAAE,CAACV,KAAK,GAAGM;QAC9B,MAAMM,OAAO,IAAIC;QACjB,IAAIC,QAAQ;QACZ,IAAK,IAAIC,IAAIL,GAAGK,IAAIrB,IAAIJ,MAAM,IAAII,GAAG,CAACqB,EAAE,CAACf,KAAK,GAAGW,QAAQI,IAAK;YAC5DH,KAAKI,GAAG,CAACtB,GAAG,CAACqB,EAAE,CAACpB,IAAI;YACpBmB;QACF;QACA,IAAIF,KAAKK,IAAI,GAAGT,gBAAiBI,KAAKK,IAAI,KAAKT,gBAAgBM,QAAQL,WAAY;YACjFF,OAAOb,GAAG,CAACgB,EAAE,CAACV,KAAK;YACnBQ,eAAeI,KAAKK,IAAI;YACxBR,YAAYK;QACd;IACF;IACA,OAAOP;AACT;AAEA,uFAAuF;AACvF,uFAAuF;AACvF,SAASW,eAAeC,IAAY,EAAE1B,KAAe;IACnD,MAAMC,MAAMH,gBAAgB4B,KAAKhC,WAAW,IAAIM;IAChD,IAAIC,IAAIJ,MAAM,KAAK,GAAG;QACpB,iFAAiF;QACjF,qFAAqF;QACrF,sFAAsF;QACtF,MAAMW,MAAMmB,KAAKC,GAAG,CAACF,KAAK7B,MAAM,EAAEX;QAClC,OAAO;YAAE2C,SAAS,GAAGH,KAAKI,KAAK,CAAC,GAAGtB,KAAKlB,OAAO,CAAC,QAAQ,KAAKK,IAAI,KAAKa,MAAMkB,KAAK7B,MAAM,GAAG,MAAM,IAAI;YAAEkC,QAAQ;QAAE;IAClH;IACA,MAAMxB,QAAQK,gBAAgBX,KAAKf;IACnC,MAAMsB,MAAMmB,KAAKC,GAAG,CAACF,KAAK7B,MAAM,EAAEU,QAAQrB;IAC1C,IAAI8C,MAAM;IACV,IAAIC,SAAS1B;IACb,KAAK,MAAM2B,KAAKjC,IAAK;QACnB,IAAIiC,EAAE3B,KAAK,GAAGA,SAAS2B,EAAE1B,GAAG,GAAGA,KAAK;QACpC,sFAAsF;QACtF,gFAAgF;QAChF,IAAI0B,EAAE3B,KAAK,GAAG0B,QAAQ;QACtBD,OAAO,GAAGN,KAAKI,KAAK,CAACG,QAAQC,EAAE3B,KAAK,EAAE,CAAC,EAAEmB,KAAKI,KAAK,CAACI,EAAE3B,KAAK,EAAE2B,EAAE1B,GAAG,EAAE,CAAC,CAAC;QACtEyB,SAASC,EAAE1B,GAAG;IAChB;IACAwB,OAAON,KAAKI,KAAK,CAACG,QAAQzB;IAC1B,MAAM2B,SAAS5B,QAAQ,IAAI,MAAM;IACjC,MAAM6B,SAAS5B,MAAMkB,KAAK7B,MAAM,GAAG,MAAM;IACzC,OAAO;QAAEgC,SAAS,GAAGM,SAASH,IAAI1C,OAAO,CAAC,QAAQ,OAAO8C,QAAQ;QAAEL,QAAQxB;IAAM;AACnF;AAEA,SAAS8B,aAAaX,IAAY,EAAEK,MAAc;IAChD,IAAIO,OAAO;IACX,IAAK,IAAIrB,IAAI,GAAGA,IAAIc,QAAQd,IAAK,IAAIS,KAAKa,UAAU,CAACtB,OAAO,IAAIqB;IAChE,OAAOA;AACT;AAEA,SAASE,aAAaC,EAAgB,EAAEC,IAAY,EAAEJ,IAAY;IAChE,MAAMK,MAAMF,GAAGG,OAAO,CAAC,qIAAqIC,GAAG,CAACH,MAAMJ,MAAMA;IAC5K,OAAOK,MAAM,CAAC,CAAC,EAAEA,IAAIG,UAAU,CAAC,CAAC,EAAEH,IAAII,QAAQ,EAAE,GAAG;AACtD;AAWA,+FAA+F;AAC/F,6DAA6D;AAC7D,OAAO,eAAeC,OAAOP,EAAgB,EAAEQ,GAAmB,EAAEjD,KAAa,EAAEkD,OAAsB,CAAC,CAAC;QA6EpCC,WAAwBC,iBAA8BC;IA5E3H,MAAMC,YAAYjF,cAAc4E,KAAKC;IACrC,MAAM,EAAEK,CAAC,EAAEC,OAAO,EAAE,GAAGF;IAEvB,MAAMG,WAAW,AAAChB,GAAGG,OAAO,CAAC,kCAAkCc,GAAG,GAA+BlE,GAAG,CAAC,CAACmE,IAAMA,EAAEjB,IAAI;IAClH,MAAMkB,aAAa/E,SAAS4D,IAAIQ,KAAKC,MAAMO;IAC3C,MAAMI,cAAcD,WAAWpC,IAAI,GAAGiC,SAAS5D,MAAM;IACrD,wFAAwF;IACxF,4FAA4F;IAC5F,qFAAqF;IACrF,IAAIiE;IACJ,IAAI;QACFA,eAAelF,cAAc6D,IAAImB,YAAYN,UAAUS,KAAK;IAC9D,EAAE,OAAOC,KAAK;QACZ,MAAMvF,YAAYuF,KAAchE,OAAOsD,UAAUS,KAAK;IACxD;IACA,MAAME,QAAQtC,KAAKuC,GAAG,CAACX,IAAI,GAAG;IAE9B,oFAAoF;IACpF,0FAA0F;IAC1F,MAAMY,eAAeX,QAAQY,OAAO,KAAKC;IACzC,IAAIF,cAAc;QAChB,MAAMG,IAAInG,YAAY8E,MAAM,oEAAoE;QAChG,IAAI3E,kBAAkBgG,IAAI;YACxB,MAAM,IAAI9F,WAAW,uBAAuB,CAAC,QAAQ,EAAE8E,UAAUiB,UAAU,CAAC,mDAAmD,EAAED,EAAEE,KAAK,CAAC,aAAa,EAAEjG,gBAAgB,uHAAuH,CAAC;QAClS;IACF;IACA,MAAMkG,kBAAkBN,gBAAgBrF,mBAAmB2D,IAAIQ,KAAKa;IAEpE,4FAA4F;IAC5F,0FAA0F;IAC1F,yFAAyF;IACzF,wFAAwF;IACxF,sFAAsF;IACtF,6FAA6F;IAC7F,mDAAmD;IACnD,UAAU;IACV,6FAA6F;IAC7F,wDAAwD;IACxD,MAAMY,QAAQpB,UAAUS,KAAK;IAC7B,MAAMY,YAAYD,QAAQ,CAAC,6CAA6C,CAAC,GAAG;IAC5E,MAAME,YAAYF,QAAQ,CAAC,KAAK,EAAEA,MAAM,CAAC,CAAC,GAAG;IAC7C,yFAAyF;IACzF,uFAAuF;IACvF,sFAAsF;IACtF,uEAAuE;IACvE,IAAIb,aAAalF,iBAAiB8D,IAAI,iBAAiBmB;IACvD,MAAMiB,YAAYhB,cAAc,CAAC,sDAAsD,CAAC,GAAG;IAC3F,MAAMzE,QAAQlB,gBAAgB+E,SAASoB,YAAY3F,aAAasB,SAASA;IAEzE,MAAM8E,aAAa,IAAIC;IACvB,IAAIC,YAAyD,EAAE;IAC/D,kFAAkF;IAClF,4EAA4E;IAC5E,IAAIxB,QAAQyB,KAAK,KAAKZ,aAAajF,MAAMO,IAAI,OAAO,IAAI;QACtD,IAAI;YACFqF,YAAY/F,gBAAgBwD,IAAIqC,YAAY1F,OAAOuF,WAAWC,WAAWC,WAAWZ,OAAOT,QAAQyB,KAAK;QAC1G,EAAE,OAAOjB,KAAK;YACZ,MAAMvF,YAAYuF,KAAchE,OAAO0E;QACzC;IACF;IACA,MAAMvB,OAAO,IAAI4B,IAAIC,UAAUxF,GAAG,CAAC,CAACmE,IAAM;YAACA,EAAEjB,IAAI;YAAEiB,EAAEuB,GAAG;SAAC;IACzD,wFAAwF;IACxF,4EAA4E;IAC5E,MAAMC,YAAY,IAAI/D,IAAI4D,UAAUpF,MAAM,CAAC,CAAC+D,IAAMA,EAAEuB,GAAG,KAAK,MAAM1F,GAAG,CAAC,CAACmE,IAAMA,EAAEjB,IAAI;IAEnF,IAAIc,QAAQ4B,KAAK,KAAKf,WAAWtF,gBAAgB0D,IAAIqC,YAAYE,WAAWvB,UAAUK,cAAcG,OAAOT,QAAQ4B,KAAK;IAExH,IAAIhC,aAAa,IAAI2B;IACrB,IAAI1B,kBAAkB,IAAI0B;IAC1B,IAAIN,iBAAiB;QAClB,CAAA,EAAErB,UAAU,EAAEC,eAAe,EAAE,GAAG,MAAMrE,kBAAkByD,IAAIQ,KAAK6B,YAAY9E,OAAOiE,OAAOH,cAAcN,QAAQY,OAAO,CAAU;IACvI;IAEA3B,GAAG4C,IAAI,CAAC;IACR5C,GAAG4C,IAAI,CAAC;IACR,MAAMC,SAAS7C,GAAGG,OAAO,CAAC;IAC1B,KAAK,MAAM,CAACF,MAAM6C,EAAE,IAAIT,WAAYQ,OAAOE,GAAG,CAAC9C,MAAM6C,EAAEE,KAAK,EAAEF,EAAEG,GAAG,GAAEvC,YAAAA,KAAKN,GAAG,CAACH,mBAATS,uBAAAA,YAAkB,OAAMC,kBAAAA,WAAWP,GAAG,CAACH,mBAAfU,6BAAAA,kBAAwB,OAAMC,uBAAAA,gBAAgBR,GAAG,CAACH,mBAApBW,kCAAAA,uBAA6B;IAExJ,yFAAyF;IACzF,2FAA2F;IAC3F,8BAA8B;IAC9B,MAAMU,QAAQW,QAAQ,CAAC,OAAO,EAAEA,MAAM,CAAC,CAAC,GAAG;IAC3C,sFAAsF;IACtF,2FAA2F;IAC3F,MAAMiB,gBAAgBlB,kBAAkB,yBAAyB;IACjE,MAAMmB,OAAOnD,GACVG,OAAO,CACN,CAAC,kIAAkI,EAAE+C,cAAc;;OAElJ,EAAE5B,MAAM,oCAAoC,CAAC,EAE/CL,GAAG,CAACH;IAEP,IAAI4B,UAAU3D,IAAI,GAAG,GAAG;QACtB,MAAMqE,YAAY1G,iBAAiBa;QACnC,KAAK,MAAM2C,OAAOiD,KAAM;YACtB,IAAIjD,IAAIuC,GAAG,KAAK,QAAQ,CAACC,UAAUW,GAAG,CAACnD,IAAID,IAAI,GAAa;YAC5D,IAAIhB;YACJ,IAAI;gBACFA,OAAO1D,aAAaC,KAAKgF,IAAI8C,OAAO,EAAEpD,IAAID,IAAI,GAAa;YAC7D,EAAE,OAAM;gBACN,UAAU,mEAAmE;YAC/E;YACA,MAAM,EAAEb,OAAO,EAAEE,MAAM,EAAE,GAAGN,eAAeC,MAAMmE;YACjDlD,IAAIuC,GAAG,GAAGrD;YACV,IAAIc,IAAIqD,KAAK,IAAI,MAAMrD,IAAIqD,KAAK,GAAG5H,eAAe6E,KAAK,cAAcT,aAAaC,IAAIE,IAAID,IAAI,EAAYL,aAAaX,MAAMK,WAAW;QAC1I;IACF;IAEA,OAAO6D;AACT"}
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
import { semanticCandidates } from '../embed/query.js';
|
|
2
2
|
import { linkEdges } from '../features/index.js';
|
|
3
|
-
import { personalizedRank } from '../graph.js';
|
|
3
|
+
import { personalizedRank } from '../graph/graph.js';
|
|
4
4
|
// RRF composes whatever the preset declares; the constant lives here since it belongs to
|
|
5
5
|
// the fusion these three functions share, not to search() itself.
|
|
6
6
|
const RRF_K = 60;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/signals.ts"],"sourcesContent":["import type { DatabaseSync } from 'node:sqlite';\nimport type { ResolvedConfig } from '../config/index.ts';\nimport { semanticCandidates } from '../embed/query.ts';\nimport { linkEdges } from '../features/index.ts';\nimport { personalizedRank } from '../graph.ts';\n\n// RRF composes whatever the preset declares; the constant lives here since it belongs to\n// the fusion these three functions share, not to search() itself.\nconst RRF_K = 60;\n\nexport type Candidates = Map<string, { score: number; via: string }>;\n\n// BM25 match rows, folded into `candidates` by reciprocal rank scaled by the preset's declared\n// weight for this signal (default 1, bit-identical to the unweighted score). Every other signal\n// is seeded from its output, so search() runs this one first when \"words\" is declared.\nexport function wordsCandidates(db: DatabaseSync, candidates: Candidates, query: string, whereJoin: string, whereCond: string, scopeCond: string, fetch: number, weight: number): Array<{ path: string; hit: string | null }> {\n const WEIGHTED_BM25 = 'bm25(content, 10.0, 5.0, 1.0, 0, 10.0, 5.0, 1.0)';\n const SNIPPET_BOUND = 16_384;\n const matchSql = `SELECT content.path AS path, CASE WHEN length(content.text) <= ${SNIPPET_BOUND} THEN snippet(content, 2, '«', '»', '…', 10) ELSE NULL END AS hit FROM content ${whereJoin} WHERE content MATCH ? ${whereCond} ${scopeCond} ORDER BY ${WEIGHTED_BM25} LIMIT ${fetch}`;\n const matchRows = db.prepare(matchSql).all(query) as Array<{ path: string; hit: string | null }>;\n matchRows.forEach((r, i) => {\n candidates.set(r.path, { score: weight / (RRF_K + i), via: 'match' });\n });\n return matchRows;\n}\n\n// Personalized-PageRank expansion from the word-match seeds, folded into `candidates` at this\n// signal's weight. A no-op when there is nothing to seed from (matchRows empty), same as before\n// this was a signal.\nexport function linksCandidates(db: DatabaseSync, candidates: Candidates, matchRows: Array<{ path: string; hit: string | null }>, allPaths: string[], allowedPaths: Set<string>, fetch: number, weight: number): void {\n if (matchRows.length === 0) return;\n const edges = linkEdges(db);\n if (edges.length === 0) return;\n // Gates the label only, not the score: restart mass ranks every seed without an incident\n // edge, but dropping it from the score reweights fusion (FEVER hit@10 0.997 -> 0.907).\n const linked = new Set(edges.flat());\n const seeds = new Map(matchRows.map((r, i) => [r.path, 1 / (i + 1)]));\n const ranked = [...personalizedRank(allPaths, edges, seeds)]\n .filter(([path, score]) => score > 1e-9 && allowedPaths.has(path))\n .sort((a, b) => b[1] - a[1])\n .slice(0, fetch);\n ranked.forEach(([path], i) => {\n const existing = candidates.get(path);\n if (existing) {\n existing.score += weight / (RRF_K + i);\n if (linked.has(path)) existing.via = 'match+link';\n } else {\n candidates.set(path, { score: weight / (RRF_K + i), via: 'link' });\n }\n });\n}\n\n// Vector expansion at the swept flat-region constants (pool = fetch), folded into `candidates`\n// at this signal's weight. Each row also carries its best chunk's line range and similarity,\n// for the caller.\nexport async function vectorsCandidates(db: DatabaseSync, cfg: ResolvedConfig, candidates: Candidates, terms: string, fetch: number, allowedPaths: Set<string>, weight: number): Promise<{ chunkLines: Map<string, string>; chunkSimilarity: Map<string, number> }> {\n const chunkLines = new Map<string, string>();\n const chunkSimilarity = new Map<string, number>();\n const vec = await semanticCandidates(db, cfg, terms, fetch, allowedPaths);\n vec.forEach(({ path, lines, similarity }, i) => {\n chunkLines.set(path, lines);\n chunkSimilarity.set(path, similarity);\n const existing = candidates.get(path);\n if (existing) {\n existing.score += weight / (RRF_K + i);\n existing.via = `${existing.via}+vector`;\n } else {\n candidates.set(path, { score: weight / (RRF_K + i), via: 'vector' });\n }\n });\n return { chunkLines, chunkSimilarity };\n}\n"],"names":["semanticCandidates","linkEdges","personalizedRank","RRF_K","wordsCandidates","db","candidates","query","whereJoin","whereCond","scopeCond","fetch","weight","WEIGHTED_BM25","SNIPPET_BOUND","matchSql","matchRows","prepare","all","forEach","r","i","set","path","score","via","linksCandidates","allPaths","allowedPaths","length","edges","linked","Set","flat","seeds","Map","map","ranked","filter","has","sort","a","b","slice","existing","get","vectorsCandidates","cfg","terms","chunkLines","chunkSimilarity","vec","lines","similarity"],"mappings":"AAEA,SAASA,kBAAkB,QAAQ,oBAAoB;AACvD,SAASC,SAAS,QAAQ,uBAAuB;AACjD,SAASC,gBAAgB,QAAQ,
|
|
1
|
+
{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/commands/signals.ts"],"sourcesContent":["import type { DatabaseSync } from 'node:sqlite';\nimport type { ResolvedConfig } from '../config/index.ts';\nimport { semanticCandidates } from '../embed/query.ts';\nimport { linkEdges } from '../features/index.ts';\nimport { personalizedRank } from '../graph/graph.ts';\n\n// RRF composes whatever the preset declares; the constant lives here since it belongs to\n// the fusion these three functions share, not to search() itself.\nconst RRF_K = 60;\n\nexport type Candidates = Map<string, { score: number; via: string }>;\n\n// BM25 match rows, folded into `candidates` by reciprocal rank scaled by the preset's declared\n// weight for this signal (default 1, bit-identical to the unweighted score). Every other signal\n// is seeded from its output, so search() runs this one first when \"words\" is declared.\nexport function wordsCandidates(db: DatabaseSync, candidates: Candidates, query: string, whereJoin: string, whereCond: string, scopeCond: string, fetch: number, weight: number): Array<{ path: string; hit: string | null }> {\n const WEIGHTED_BM25 = 'bm25(content, 10.0, 5.0, 1.0, 0, 10.0, 5.0, 1.0)';\n const SNIPPET_BOUND = 16_384;\n const matchSql = `SELECT content.path AS path, CASE WHEN length(content.text) <= ${SNIPPET_BOUND} THEN snippet(content, 2, '«', '»', '…', 10) ELSE NULL END AS hit FROM content ${whereJoin} WHERE content MATCH ? ${whereCond} ${scopeCond} ORDER BY ${WEIGHTED_BM25} LIMIT ${fetch}`;\n const matchRows = db.prepare(matchSql).all(query) as Array<{ path: string; hit: string | null }>;\n matchRows.forEach((r, i) => {\n candidates.set(r.path, { score: weight / (RRF_K + i), via: 'match' });\n });\n return matchRows;\n}\n\n// Personalized-PageRank expansion from the word-match seeds, folded into `candidates` at this\n// signal's weight. A no-op when there is nothing to seed from (matchRows empty), same as before\n// this was a signal.\nexport function linksCandidates(db: DatabaseSync, candidates: Candidates, matchRows: Array<{ path: string; hit: string | null }>, allPaths: string[], allowedPaths: Set<string>, fetch: number, weight: number): void {\n if (matchRows.length === 0) return;\n const edges = linkEdges(db);\n if (edges.length === 0) return;\n // Gates the label only, not the score: restart mass ranks every seed without an incident\n // edge, but dropping it from the score reweights fusion (FEVER hit@10 0.997 -> 0.907).\n const linked = new Set(edges.flat());\n const seeds = new Map(matchRows.map((r, i) => [r.path, 1 / (i + 1)]));\n const ranked = [...personalizedRank(allPaths, edges, seeds)]\n .filter(([path, score]) => score > 1e-9 && allowedPaths.has(path))\n .sort((a, b) => b[1] - a[1])\n .slice(0, fetch);\n ranked.forEach(([path], i) => {\n const existing = candidates.get(path);\n if (existing) {\n existing.score += weight / (RRF_K + i);\n if (linked.has(path)) existing.via = 'match+link';\n } else {\n candidates.set(path, { score: weight / (RRF_K + i), via: 'link' });\n }\n });\n}\n\n// Vector expansion at the swept flat-region constants (pool = fetch), folded into `candidates`\n// at this signal's weight. Each row also carries its best chunk's line range and similarity,\n// for the caller.\nexport async function vectorsCandidates(db: DatabaseSync, cfg: ResolvedConfig, candidates: Candidates, terms: string, fetch: number, allowedPaths: Set<string>, weight: number): Promise<{ chunkLines: Map<string, string>; chunkSimilarity: Map<string, number> }> {\n const chunkLines = new Map<string, string>();\n const chunkSimilarity = new Map<string, number>();\n const vec = await semanticCandidates(db, cfg, terms, fetch, allowedPaths);\n vec.forEach(({ path, lines, similarity }, i) => {\n chunkLines.set(path, lines);\n chunkSimilarity.set(path, similarity);\n const existing = candidates.get(path);\n if (existing) {\n existing.score += weight / (RRF_K + i);\n existing.via = `${existing.via}+vector`;\n } else {\n candidates.set(path, { score: weight / (RRF_K + i), via: 'vector' });\n }\n });\n return { chunkLines, chunkSimilarity };\n}\n"],"names":["semanticCandidates","linkEdges","personalizedRank","RRF_K","wordsCandidates","db","candidates","query","whereJoin","whereCond","scopeCond","fetch","weight","WEIGHTED_BM25","SNIPPET_BOUND","matchSql","matchRows","prepare","all","forEach","r","i","set","path","score","via","linksCandidates","allPaths","allowedPaths","length","edges","linked","Set","flat","seeds","Map","map","ranked","filter","has","sort","a","b","slice","existing","get","vectorsCandidates","cfg","terms","chunkLines","chunkSimilarity","vec","lines","similarity"],"mappings":"AAEA,SAASA,kBAAkB,QAAQ,oBAAoB;AACvD,SAASC,SAAS,QAAQ,uBAAuB;AACjD,SAASC,gBAAgB,QAAQ,oBAAoB;AAErD,yFAAyF;AACzF,kEAAkE;AAClE,MAAMC,QAAQ;AAId,+FAA+F;AAC/F,gGAAgG;AAChG,uFAAuF;AACvF,OAAO,SAASC,gBAAgBC,EAAgB,EAAEC,UAAsB,EAAEC,KAAa,EAAEC,SAAiB,EAAEC,SAAiB,EAAEC,SAAiB,EAAEC,KAAa,EAAEC,MAAc;IAC7K,MAAMC,gBAAgB;IACtB,MAAMC,gBAAgB;IACtB,MAAMC,WAAW,CAAC,+DAA+D,EAAED,cAAc,+EAA+E,EAAEN,UAAU,uBAAuB,EAAEC,UAAU,CAAC,EAAEC,UAAU,UAAU,EAAEG,cAAc,OAAO,EAAEF,OAAO;IACtR,MAAMK,YAAYX,GAAGY,OAAO,CAACF,UAAUG,GAAG,CAACX;IAC3CS,UAAUG,OAAO,CAAC,CAACC,GAAGC;QACpBf,WAAWgB,GAAG,CAACF,EAAEG,IAAI,EAAE;YAAEC,OAAOZ,SAAUT,CAAAA,QAAQkB,CAAAA;YAAII,KAAK;QAAQ;IACrE;IACA,OAAOT;AACT;AAEA,8FAA8F;AAC9F,gGAAgG;AAChG,qBAAqB;AACrB,OAAO,SAASU,gBAAgBrB,EAAgB,EAAEC,UAAsB,EAAEU,SAAsD,EAAEW,QAAkB,EAAEC,YAAyB,EAAEjB,KAAa,EAAEC,MAAc;IAC5M,IAAII,UAAUa,MAAM,KAAK,GAAG;IAC5B,MAAMC,QAAQ7B,UAAUI;IACxB,IAAIyB,MAAMD,MAAM,KAAK,GAAG;IACxB,yFAAyF;IACzF,uFAAuF;IACvF,MAAME,SAAS,IAAIC,IAAIF,MAAMG,IAAI;IACjC,MAAMC,QAAQ,IAAIC,IAAInB,UAAUoB,GAAG,CAAC,CAAChB,GAAGC,IAAM;YAACD,EAAEG,IAAI;YAAE,IAAKF,CAAAA,IAAI,CAAA;SAAG;IACnE,MAAMgB,SAAS;WAAInC,iBAAiByB,UAAUG,OAAOI;KAAO,CACzDI,MAAM,CAAC,CAAC,CAACf,MAAMC,MAAM,GAAKA,QAAQ,QAAQI,aAAaW,GAAG,CAAChB,OAC3DiB,IAAI,CAAC,CAACC,GAAGC,IAAMA,CAAC,CAAC,EAAE,GAAGD,CAAC,CAAC,EAAE,EAC1BE,KAAK,CAAC,GAAGhC;IACZ0B,OAAOlB,OAAO,CAAC,CAAC,CAACI,KAAK,EAAEF;QACtB,MAAMuB,WAAWtC,WAAWuC,GAAG,CAACtB;QAChC,IAAIqB,UAAU;YACZA,SAASpB,KAAK,IAAIZ,SAAUT,CAAAA,QAAQkB,CAAAA;YACpC,IAAIU,OAAOQ,GAAG,CAAChB,OAAOqB,SAASnB,GAAG,GAAG;QACvC,OAAO;YACLnB,WAAWgB,GAAG,CAACC,MAAM;gBAAEC,OAAOZ,SAAUT,CAAAA,QAAQkB,CAAAA;gBAAII,KAAK;YAAO;QAClE;IACF;AACF;AAEA,+FAA+F;AAC/F,6FAA6F;AAC7F,kBAAkB;AAClB,OAAO,eAAeqB,kBAAkBzC,EAAgB,EAAE0C,GAAmB,EAAEzC,UAAsB,EAAE0C,KAAa,EAAErC,KAAa,EAAEiB,YAAyB,EAAEhB,MAAc;IAC5K,MAAMqC,aAAa,IAAId;IACvB,MAAMe,kBAAkB,IAAIf;IAC5B,MAAMgB,MAAM,MAAMnD,mBAAmBK,IAAI0C,KAAKC,OAAOrC,OAAOiB;IAC5DuB,IAAIhC,OAAO,CAAC,CAAC,EAAEI,IAAI,EAAE6B,KAAK,EAAEC,UAAU,EAAE,EAAEhC;QACxC4B,WAAW3B,GAAG,CAACC,MAAM6B;QACrBF,gBAAgB5B,GAAG,CAACC,MAAM8B;QAC1B,MAAMT,WAAWtC,WAAWuC,GAAG,CAACtB;QAChC,IAAIqB,UAAU;YACZA,SAASpB,KAAK,IAAIZ,SAAUT,CAAAA,QAAQkB,CAAAA;YACpCuB,SAASnB,GAAG,GAAG,GAAGmB,SAASnB,GAAG,CAAC,OAAO,CAAC;QACzC,OAAO;YACLnB,WAAWgB,GAAG,CAACC,MAAM;gBAAEC,OAAOZ,SAAUT,CAAAA,QAAQkB,CAAAA;gBAAII,KAAK;YAAS;QACpE;IACF;IACA,OAAO;QAAEwB;QAAYC;IAAgB;AACvC"}
|
|
@@ -1,3 +1,4 @@
|
|
|
1
|
+
import type { Feature } from '../features/types.js';
|
|
1
2
|
import type { SignalName, SignalWeights } from './signals.js';
|
|
2
3
|
import { type Config, type FeatureName } from './types.js';
|
|
3
4
|
export declare function presetNames(cfg: Config): string[];
|
|
@@ -17,6 +18,7 @@ export declare function embedConfig(cfg: Config): {
|
|
|
17
18
|
url?: string;
|
|
18
19
|
key?: string;
|
|
19
20
|
languages?: string[];
|
|
21
|
+
chunkTokens?: number;
|
|
20
22
|
} | null;
|
|
21
23
|
export declare function contentTokenize(cfg: Config): string | undefined;
|
|
22
|
-
export declare function featureSignature(cfg: Config): string;
|
|
24
|
+
export declare function featureSignature(cfg: Config, features: Feature[]): string;
|