sensemaking 0.15.4 → 0.17.0

package/dist/cjs/scan.js

@@ -1,433 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", {
-    value: true
-});
-function _export(target, all) {
-    for(var name in all)Object.defineProperty(target, name, {
-        enumerable: true,
-        get: Object.getOwnPropertyDescriptor(all, name).get
-    });
-}
-_export(exports, {
-    get RESERVED_COLUMNS () {
-        return RESERVED_COLUMNS;
-    },
-    get listFiles () {
-        return listFiles;
-    },
-    get looksLikeDatetime () {
-        return looksLikeDatetime;
-    },
-    get normalizeDate () {
-        return normalizeDate;
-    },
-    get parseFile () {
-        return parseFile;
-    },
-    get toPosixPath () {
-        return toPosixPath;
-    }
-});
-var _nodefs = require("node:fs");
-var _nodepath = require("node:path");
-var _removemarkdown = /*#__PURE__*/ _interop_require_default(require("remove-markdown"));
-var _yaml = require("yaml");
-var _indexts = require("./config/index.js");
-function _array_like_to_array(arr, len) {
-    if (len == null || len > arr.length) len = arr.length;
-    for(var i = 0, arr2 = new Array(len); i < len; i++)arr2[i] = arr[i];
-    return arr2;
-}
-function _array_with_holes(arr) {
-    if (Array.isArray(arr)) return arr;
-}
-function _array_without_holes(arr) {
-    if (Array.isArray(arr)) return _array_like_to_array(arr);
-}
-function _interop_require_default(obj) {
-    return obj && obj.__esModule ? obj : {
-        default: obj
-    };
-}
-function _iterable_to_array(iter) {
-    if (typeof Symbol !== "undefined" && iter[Symbol.iterator] != null || iter["@@iterator"] != null) {
-        return Array.from(iter);
-    }
-}
-function _iterable_to_array_limit(arr, i) {
-    var _i = arr == null ? null : typeof Symbol !== "undefined" && arr[Symbol.iterator] || arr["@@iterator"];
-    if (_i == null) return;
-    var _arr = [];
-    var _n = true;
-    var _d = false;
-    var _s, _e;
-    try {
-        for(_i = _i.call(arr); !(_n = (_s = _i.next()).done); _n = true){
-            _arr.push(_s.value);
-            if (i && _arr.length === i) break;
-        }
-    } catch (err) {
-        _d = true;
-        _e = err;
-    } finally{
-        try {
-            if (!_n && _i["return"] != null) _i["return"]();
-        } finally{
-            if (_d) throw _e;
-        }
-    }
-    return _arr;
-}
-function _non_iterable_rest() {
-    throw new TypeError("Invalid attempt to destructure non-iterable instance.\nIn order to be iterable, non-array objects must have a [Symbol.iterator]() method.");
-}
-function _non_iterable_spread() {
-    throw new TypeError("Invalid attempt to spread non-iterable instance.\nIn order to be iterable, non-array objects must have a [Symbol.iterator]() method.");
-}
-function _sliced_to_array(arr, i) {
-    return _array_with_holes(arr) || _iterable_to_array_limit(arr, i) || _unsupported_iterable_to_array(arr, i) || _non_iterable_rest();
-}
-function _to_consumable_array(arr) {
-    return _array_without_holes(arr) || _iterable_to_array(arr) || _unsupported_iterable_to_array(arr) || _non_iterable_spread();
-}
-function _type_of(obj) {
-    "@swc/helpers - typeof";
-    return obj && typeof Symbol !== "undefined" && obj.constructor === Symbol ? "symbol" : typeof obj;
-}
-function _unsupported_iterable_to_array(o, minLen) {
-    if (!o) return;
-    if (typeof o === "string") return _array_like_to_array(o, minLen);
-    var n = Object.prototype.toString.call(o).slice(8, -1);
-    if (n === "Object" && o.constructor) n = o.constructor.name;
-    if (n === "Map" || n === "Set") return Array.from(n);
-    if (n === "Arguments" || /^(?:Ui|I)nt(?:8|16|32)(?:Clamped)?Array$/.test(n)) return _array_like_to_array(o, minLen);
-}
-var RESERVED_COLUMNS = new Set([
-    'path',
-    '_mtime',
-    '_ctime',
-    '_size',
-    '_rank',
-    '_parse_error',
-    'content',
-    'links',
-    'sections'
-]);
-// YAML error codes whose recovery is unambiguous, so the parse is accepted rather than
-// quarantined. Only one qualifies: YAML 1.2 reserves `@` and `` ` `` at the start of a plain
-// scalar for future use, so they can never be valid and the text can only be what was typed
-// (`aliases: [@handle]` -> ["@handle"]). Every other code has a second reading -- an unquoted
-// `:` swallows the keys after it, an unquoted `[..](..)` drops the URL, a duplicate key picks
-// one value in silence -- so it writes values nobody wrote.
-var ACCEPTED_YAML_CODES = new Set([
-    'BAD_SCALAR_START'
-]);
-function normalizeText(value) {
-    if (value === null || value === undefined) return '';
-    return String(value).replace(/\s+/g, ' ').trim();
-}
-// Keeps URL query strings, asset filenames, and HTML attributes out of the index: rare terms
-// carry high IDF, so they outrank prose. remove-markdown misses wikilinks and tables.
-function stripText(value) {
-    var withoutWikilinks = value.replace(/\[\[([^\]|]+)\|([^\]]+)\]\]/g, '$2').replace(/\[\[([^\]]+)\]\]/g, '$1');
-    var withoutMarkdown = (0, _removemarkdown.default)(withoutWikilinks);
-    var withoutTables = withoutMarkdown.replace(/^\s*\|?[-\s|:]+\|\s*$/gm, '').replace(/\|/g, ' ');
-    return normalizeText(withoutTables);
-}
-function toPosixPath(relPath) {
-    var separator = arguments.length > 1 && arguments[1] !== void 0 ? arguments[1] : _nodepath.sep;
-    return separator === '\\' ? relPath.split(separator).join('/') : relPath;
-}
-// Every command pays listFiles before it answers (the freshness check stats each file), so
-// per-file work here is the hottest path in the package. Everything derivable from the config
-// alone is computed once, above the loop.
-var NO_THROW = {
-    throwIfNoEntry: false
-};
-function listFiles(cfg, baseDir) {
-    var coverage = new Map();
-    var posixNeeded = _nodepath.sep === '\\';
-    var _iteratorNormalCompletion = true, _didIteratorError = false, _iteratorError = undefined;
-    try {
-        for(var _iterator = (0, _indexts.presetNames)(cfg)[Symbol.iterator](), _step; !(_iteratorNormalCompletion = (_step = _iterator.next()).done); _iteratorNormalCompletion = true){
-            var name = _step.value;
-            var preset = cfg.presets[name];
-            var _iteratorNormalCompletion1 = true, _didIteratorError1 = false, _iteratorError1 = undefined;
-            try {
-                for(var _iterator1 = (0, _nodefs.globSync)(preset.include, {
-                    cwd: baseDir,
-                    exclude: preset.exclude
-                })[Symbol.iterator](), _step1; !(_iteratorNormalCompletion1 = (_step1 = _iterator1.next()).done); _iteratorNormalCompletion1 = true){
-                    var matched = _step1.value;
-                    var _coverage_get;
-                    var relPath = posixNeeded ? toPosixPath(matched) : matched;
-                    var set = (_coverage_get = coverage.get(relPath)) !== null && _coverage_get !== void 0 ? _coverage_get : new Set();
-                    set.add(name);
-                    coverage.set(relPath, set);
-                }
-            } catch (err) {
-                _didIteratorError1 = true;
-                _iteratorError1 = err;
-            } finally{
-                try {
-                    if (!_iteratorNormalCompletion1 && _iterator1.return != null) {
-                        _iterator1.return();
-                    }
-                } finally{
-                    if (_didIteratorError1) {
-                        throw _iteratorError1;
-                    }
-                }
-            }
-        }
-    } catch (err) {
-        _didIteratorError = true;
-        _iteratorError = err;
-    } finally{
-        try {
-            if (!_iteratorNormalCompletion && _iterator.return != null) {
-                _iterator.return();
-            }
-        } finally{
-            if (_didIteratorError) {
-                throw _iteratorError;
-            }
-        }
-    }
-    // Which presets want vectors is a property of the config, not of any file.
-    var embedding = (0, _indexts.embedEnabled)(cfg);
-    var semanticPresets = embedding ? new Set((0, _indexts.presetNames)(cfg).filter(function(name) {
-        return (0, _indexts.presetSemanticEnabled)(cfg, name);
-    })) : null;
-    var files = [];
-    var _iteratorNormalCompletion2 = true, _didIteratorError2 = false, _iteratorError2 = undefined;
-    try {
-        for(var _iterator2 = _to_consumable_array(coverage.keys()).sort()[Symbol.iterator](), _step2; !(_iteratorNormalCompletion2 = (_step2 = _iterator2.next()).done); _iteratorNormalCompletion2 = true){
-            var relPath1 = _step2.value;
-            var absPath = (0, _nodepath.join)(baseDir, relPath1); // join re-applies the platform separator for fs calls
-            // node:fs glob matches directories and dangling symlinks; fast-glob returned neither, so
-            // one stat filters both back out (throwIfNoEntry keeps a dangling link from throwing).
-            var st = (0, _nodefs.statSync)(absPath, NO_THROW);
-            if (!(st === null || st === void 0 ? void 0 : st.isFile())) continue;
-            var presets = _to_consumable_array(coverage.get(relPath1)).sort();
-            var embed = semanticPresets !== null && presets.some(function(name) {
-                return semanticPresets.has(name);
-            });
-            files.push({
-                relPath: relPath1,
-                absPath: absPath,
-                mtimeMs: st.mtimeMs,
-                ctimeMs: st.birthtimeMs,
-                size: st.size,
-                presets: presets,
-                embed: embed
-            });
-        }
-    } catch (err) {
-        _didIteratorError2 = true;
-        _iteratorError2 = err;
-    } finally{
-        try {
-            if (!_iteratorNormalCompletion2 && _iterator2.return != null) {
-                _iterator2.return();
-            }
-        } finally{
-            if (_didIteratorError2) {
-                throw _iteratorError2;
-            }
-        }
-    }
-    return files;
-}
-// SQLite's datetime() rejects a colonless offset (`-0800`) and a space separator, which ISO 8601
-// allows and producers emit. A rejected date is invisible, not excluded: every comparison is NULL.
-var ISO_DATETIME = /^(\d{4}-\d{2}-\d{2})[T ](\d{2}:\d{2}(?::\d{2})?(?:\.\d+)?)(Z|[+-]\d{2}(?::?\d{2})?)?$/;
-// A value opening `YYYY-MM-DDT` was meant to be a datetime; prose never is. Reported when it
-// cannot be normalized, so a typo surfaces on the next crawl instead of at some later audit.
-var MEANT_AS_DATETIME = /^\d{4}-\d{2}-\d{2}[T ]\d/;
-function looksLikeDatetime(value) {
-    return MEANT_AS_DATETIME.test(value);
-}
-function normalizeDate(value) {
-    var m = ISO_DATETIME.exec(value);
-    if (m === null) return value;
-    var _m = _sliced_to_array(m, 4), date = _m[1], time = _m[2], zone = _m[3];
-    var digits = zone === undefined || zone === 'Z' ? '' : zone.replace(':', '');
-    var offset = digits === '' ? zone !== null && zone !== void 0 ? zone : '' : digits.length === 3 ? "".concat(digits, ":00") : "".concat(digits.slice(0, 3), ":").concat(digits.slice(3));
-    var normalized = "".concat(date, "T").concat(time).concat(offset);
-    return Number.isNaN(Date.parse(normalized)) ? value : normalized;
-}
-// Storage class follows the YAML scalar. Booleans store as 1/0, so `WHERE flag = 1` matches
-// and `WHERE flag = 'true'` cannot; `map` prints observed types so the mismatch is visible.
-function mapValue(value) {
-    if (value === null || value === undefined) return null;
-    if (typeof value === 'boolean') return BigInt(value ? 1 : 0);
-    if (typeof value === 'number') return Number.isSafeInteger(value) ? BigInt(value) : value;
-    if (typeof value === 'string') return normalizeDate(value);
-    return JSON.stringify(value);
-}
-// The delimiter split is all this package used gray-matter for.
-function splitFrontmatter(raw) {
-    var open = raw.match(/^---\r?\n/);
-    if (!open) return {
-        fm: null,
-        body: raw
-    };
-    var rest = raw.slice(open[0].length);
-    var close = rest.match(/^---\r?(\n|$)/m);
-    if (!close || close.index === undefined) return {
-        fm: null,
-        body: raw
-    };
-    return {
-        fm: rest.slice(0, close.index),
-        body: rest.slice(close.index + close[0].length)
-    };
-}
-// A well-formed document can still hold a value nobody meant: `created: {{date}}` is valid
-// YAML for a flow map used as a mapping key, so it raises no error and stores
-// {"{ date }": null}. No error code can catch that, but yaml notices the stringified key, so
-// this reports it with the path instead (yaml's own warning has none, fires once per document,
-// and is what trains readers to discard stderr).
-function warnStringifiedKeys(relPath, doc, warnings) {
-    var found = false;
-    // Nested, not top level: `created: {{date}}` puts the collection key one level down, inside
-    // the flow map that `{{...}}` parses as.
-    (0, _yaml.visit)(doc, {
-        Pair: function Pair(_key, pair) {
-            if (!(0, _yaml.isCollection)(pair.key)) return undefined;
-            found = true;
-            return _yaml.visit.BREAK;
-        }
-    });
-    // One per file: a template repeats the same mistake on every field it stamps.
-    if (found) warnings.push("warning: ".concat(relPath, " frontmatter has a key that is itself a list or mapping, stored as text; this is usually an unrendered template placeholder like {{date}}"));
-}
-// Accept a clean parse, and one whose every error is unambiguous (ACCEPTED_YAML_CODES).
-// Anything else is quarantined: no frontmatter columns at all, and `_parse_error` carries the
-// reason. Recovering it would write values nobody wrote, which is worse than absence because
-// no query can see it. The file is still indexed -- content, links and sections never touch
-// frontmatter -- so a broken note stays searchable while it is being hunted for.
-// yaml's message continues onto a source excerpt, so the first line is the sentence -- minus
-// the colon that introduced the part being dropped.
-function firstLine(message) {
-    return message.split('\n')[0].replace(/:\s*$/, '');
-}
-function parseFrontmatter(relPath, fm, warnings) {
-    // logLevel silences yaml's own pathless warnings; warnStringifiedKeys re-reports the one
-    // that carries information, with the file it came from.
-    var doc = (0, _yaml.parseDocument)(fm, {
-        logLevel: 'silent'
-    });
-    var refused = doc.errors.filter(function(err) {
-        return !ACCEPTED_YAML_CODES.has(err.code);
-    });
-    if (refused.length > 0) {
-        var detail = refused.length > 1 ? " (and ".concat(refused.length - 1, " more)") : '';
-        var parseError = "".concat(firstLine(refused[0].message)).concat(detail);
-        warnings.push("warning: ".concat(relPath, " frontmatter did not parse, so none of it is indexed: ").concat(parseError));
-        return {
-            data: {},
-            parseError: parseError
-        };
-    }
-    var data;
-    try {
-        data = doc.toJS();
-    } catch (err) {
-        // Reaches here with doc.errors empty: `title: **Bold**` parses, then opens an alias on
-        // materialisation. An empty error list is not a successful parse.
-        var parseError1 = firstLine(err.message);
-        warnings.push("warning: ".concat(relPath, " frontmatter did not parse, so none of it is indexed: ").concat(parseError1));
-        return {
-            data: {},
-            parseError: parseError1
-        };
-    }
-    if (data === null || data === undefined) return {
-        data: {},
-        parseError: null
-    };
-    if ((typeof data === "undefined" ? "undefined" : _type_of(data)) !== 'object' || Array.isArray(data)) {
-        var parseError2 = 'frontmatter is not a key-value mapping';
-        warnings.push("warning: ".concat(relPath, " ").concat(parseError2, "; none of it is indexed"));
-        return {
-            data: {},
-            parseError: parseError2
-        };
-    }
-    warnStringifiedKeys(relPath, doc, warnings);
-    return {
-        data: data,
-        parseError: null
-    };
-}
-function parseFile(file) {
-    var extractors = arguments.length > 1 && arguments[1] !== void 0 ? arguments[1] : [];
-    var raw = (0, _nodefs.readFileSync)(file.absPath, 'utf8');
-    var warnings = [];
-    var _splitFrontmatter = splitFrontmatter(raw), fm = _splitFrontmatter.fm, content = _splitFrontmatter.body;
-    var _ref = fm === null ? {
-        data: {},
-        parseError: null
-    } : parseFrontmatter(file.relPath, fm, warnings), data = _ref.data, parseError = _ref.parseError;
-    var mapped = {};
-    var _iteratorNormalCompletion = true, _didIteratorError = false, _iteratorError = undefined;
-    try {
-        for(var _iterator = Object.keys(data)[Symbol.iterator](), _step; !(_iteratorNormalCompletion = (_step = _iterator.next()).done); _iteratorNormalCompletion = true){
-            var key = _step.value;
-            if (RESERVED_COLUMNS.has(key)) {
-                warnings.push("warning: ".concat(file.relPath, ' has a frontmatter key named "').concat(key, '", which is reserved; ignoring it'));
-                continue;
-            }
-            var value = mapValue(data[key]);
-            if (typeof value === 'string' && looksLikeDatetime(value) && Number.isNaN(Date.parse(value))) {
-                warnings.push("warning: ".concat(file.relPath, ": ").concat(key, " is not a valid date (").concat(value, "), so it is invisible to every date comparison"));
-            }
-            mapped[key] = value;
-        }
-    } catch (err) {
-        _didIteratorError = true;
-        _iteratorError = err;
-    } finally{
-        try {
-            if (!_iteratorNormalCompletion && _iterator.return != null) {
-                _iterator.return();
-            }
-        } finally{
-            if (_didIteratorError) {
-                throw _iteratorError;
-            }
-        }
-    }
-    // title/summary are plain YAML strings -- whitespace-collapse only;
-    // the prose gets the full markdown strip.
-    var search = {
-        title: normalizeText(data.title),
-        summary: normalizeText(data.summary),
-        text: stripText(content)
-    };
-    return {
-        doc: {
-            relPath: file.relPath,
-            mtimeMs: file.mtimeMs,
-            ctimeMs: file.ctimeMs,
-            size: file.size,
-            presets: file.presets,
-            data: mapped,
-            parseError: parseError,
-            search: search,
-            extracted: Object.fromEntries(extractors.filter(function(f) {
-                return f.extract;
-            }).map(function(f) {
-                var _f_extract;
-                return [
-                    f.name,
-                    (_f_extract = f.extract) === null || _f_extract === void 0 ? void 0 : _f_extract.call(f, raw, content, search, data)
-                ];
-            }))
-        },
-        warnings: warnings
-    };
-}
-/* CJS INTEROP */ if (exports.__esModule && exports.default) { try { Object.defineProperty(exports.default, '__esModule', { value: true }); for (var key in exports) { exports.default[key] = exports[key]; } } catch (_) {}; module.exports = exports.default; }

package/dist/cjs/scan.js.map

@@ -1 +0,0 @@
-{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/scan.ts"],"sourcesContent":["import { globSync, readFileSync, statSync } from 'node:fs';\nimport { join, sep } from 'node:path';\nimport removeMarkdown from 'remove-markdown';\nimport { isCollection, parseDocument, visit } from 'yaml';\nimport type { Config } from './config/index.ts';\nimport { embedEnabled, presetNames, presetSemanticEnabled } from './config/index.ts';\nimport type { Feature } from './features/types.ts';\n\n// Filesystem -> rows. Pure data in, data + warnings out; db.ts does the SQL.\n\n// Frontmatter keys that would collide with table columns. Exported so db.ts's upsert can tell\n// a feature-owned column (`_rank`) from a parsed one and leave it alone on reparse.\nexport const RESERVED_COLUMNS = new Set(['path', '_mtime', '_ctime', '_size', '_rank', '_parse_error', 'content', 'links', 'sections']);\n\n// YAML error codes whose recovery is unambiguous, so the parse is accepted rather than\n// quarantined. Only one qualifies: YAML 1.2 reserves `@` and `` ` `` at the start of a plain\n// scalar for future use, so they can never be valid and the text can only be what was typed\n// (`aliases: [@handle]` -> [\"@handle\"]). Every other code has a second reading -- an unquoted\n// `:` swallows the keys after it, an unquoted `[..](..)` drops the URL, a duplicate key picks\n// one value in silence -- so it writes values nobody wrote.\nconst ACCEPTED_YAML_CODES = new Set(['BAD_SCALAR_START']);\n\nfunction normalizeText(value: unknown): string {\n  if (value === null || value === undefined) return '';\n  return String(value).replace(/\\s+/g, ' ').trim();\n}\n\n// Keeps URL query strings, asset filenames, and HTML attributes out of the index: rare terms\n// carry high IDF, so they outrank prose. remove-markdown misses wikilinks and tables.\nfunction stripText(value: string): string {\n  const withoutWikilinks = value.replace(/\\[\\[([^\\]|]+)\\|([^\\]]+)\\]\\]/g, '$2').replace(/\\[\\[([^\\]]+)\\]\\]/g, '$1');\n  const withoutMarkdown = removeMarkdown(withoutWikilinks);\n  const withoutTables = withoutMarkdown.replace(/^\\s*\\|?[-\\s|:]+\\|\\s*$/gm, '').replace(/\\|/g, ' ');\n  return normalizeText(withoutTables);\n}\n\nexport interface FileStat {\n  relPath: string;\n  absPath: string;\n  mtimeMs: number;\n  ctimeMs: number; // filesystem birthtime; a clone or copy resets it, like _mtime\n  size: number;\n  presets: string[]; // every declared preset covering this file (>= 1; union, overlap allowed)\n  embed: boolean; // true iff a model is named and some covering preset has semantic on\n}\n\n// Presets are views, not partitions: they overlap freely, and a file's covering set (not one\n// owner) drives indexing. Globs resolve relative to baseDir; unmatched files are not indexed.\nexport function toPosixPath(relPath: string, separator: string = sep): string {\n  return separator === '\\\\' ? relPath.split(separator).join('/') : relPath;\n}\n\n// Every command pays listFiles before it answers (the freshness check stats each file), so\n// per-file work here is the hottest path in the package. Everything derivable from the config\n// alone is computed once, above the loop.\nconst NO_THROW = { throwIfNoEntry: false } as const;\n\nexport function listFiles(cfg: Config, baseDir: string): FileStat[] {\n  const coverage = new Map<string, Set<string>>();\n  const posixNeeded = sep === '\\\\';\n  for (const name of presetNames(cfg)) {\n    const preset = cfg.presets[name];\n    for (const matched of globSync(preset.include, { cwd: baseDir, exclude: preset.exclude })) {\n      const relPath = posixNeeded ? toPosixPath(matched) : matched;\n      const set = coverage.get(relPath) ?? new Set<string>();\n      set.add(name);\n      coverage.set(relPath, set);\n    }\n  }\n\n  // Which presets want vectors is a property of the config, not of any file.\n  const embedding = embedEnabled(cfg);\n  const semanticPresets = embedding ? new Set(presetNames(cfg).filter((name) => presetSemanticEnabled(cfg, name))) : null;\n\n  const files: FileStat[] = [];\n  for (const relPath of [...coverage.keys()].sort()) {\n    const absPath = join(baseDir, relPath); // join re-applies the platform separator for fs calls\n    // node:fs glob matches directories and dangling symlinks; fast-glob returned neither, so\n    // one stat filters both back out (throwIfNoEntry keeps a dangling link from throwing).\n    const st = statSync(absPath, NO_THROW);\n    if (!st?.isFile()) continue;\n    const presets = [...(coverage.get(relPath) as Set<string>)].sort();\n    const embed = semanticPresets !== null && presets.some((name) => semanticPresets.has(name));\n    files.push({ relPath, absPath, mtimeMs: st.mtimeMs, ctimeMs: st.birthtimeMs, size: st.size, presets, embed });\n  }\n  return files;\n}\n\nexport interface ParsedDoc {\n  relPath: string;\n  mtimeMs: number;\n  ctimeMs: number;\n  size: number;\n  presets: string[];\n  data: Record<string, string | number | bigint | null>;\n  // NULL when the frontmatter parsed, the first YAML message otherwise. In the row rather than\n  // a side table so `SELECT *` and any `IS NULL` investigation trip over it without being asked.\n  parseError: string | null;\n  // title/summary are duplicated from frontmatter so bm25() can weight them above the body text.\n  search: { title: string; summary: string; text: string };\n  // Per-feature extraction results, keyed by feature name; features store them at reconcile.\n  extracted: Record<string, unknown>;\n}\n\n// SQLite's datetime() rejects a colonless offset (`-0800`) and a space separator, which ISO 8601\n// allows and producers emit. A rejected date is invisible, not excluded: every comparison is NULL.\nconst ISO_DATETIME = /^(\\d{4}-\\d{2}-\\d{2})[T ](\\d{2}:\\d{2}(?::\\d{2})?(?:\\.\\d+)?)(Z|[+-]\\d{2}(?::?\\d{2})?)?$/;\n\n// A value opening `YYYY-MM-DDT` was meant to be a datetime; prose never is. Reported when it\n// cannot be normalized, so a typo surfaces on the next crawl instead of at some later audit.\nconst MEANT_AS_DATETIME = /^\\d{4}-\\d{2}-\\d{2}[T ]\\d/;\n\nexport function looksLikeDatetime(value: string): boolean {\n  return MEANT_AS_DATETIME.test(value);\n}\n\n// Punctuation only, never a timezone conversion: the offset survives, so substr(d,1,10) is still\n// the local date. A shape that is not a real instant is left as written, and stays auditable.\nexport function normalizeDate(value: string): string {\n  const m = ISO_DATETIME.exec(value);\n  if (m === null) return value;\n  const [, date, time, zone] = m;\n  const digits = zone === undefined || zone === 'Z' ? '' : zone.replace(':', '');\n  const offset = digits === '' ? (zone ?? '') : digits.length === 3 ? `${digits}:00` : `${digits.slice(0, 3)}:${digits.slice(3)}`;\n  const normalized = `${date}T${time}${offset}`;\n  return Number.isNaN(Date.parse(normalized)) ? value : normalized;\n}\n\n// Storage class follows the YAML scalar. Booleans store as 1/0, so `WHERE flag = 1` matches\n// and `WHERE flag = 'true'` cannot; `map` prints observed types so the mismatch is visible.\nfunction mapValue(value: unknown): string | number | bigint | null {\n  if (value === null || value === undefined) return null;\n  if (typeof value === 'boolean') return BigInt(value ? 1 : 0);\n  if (typeof value === 'number') return Number.isSafeInteger(value) ? BigInt(value) : value;\n  if (typeof value === 'string') return normalizeDate(value);\n  return JSON.stringify(value);\n}\n\n// The delimiter split is all this package used gray-matter for.\nfunction splitFrontmatter(raw: string): { fm: string | null; body: string } {\n  const open = raw.match(/^---\\r?\\n/);\n  if (!open) return { fm: null, body: raw };\n  const rest = raw.slice(open[0].length);\n  const close = rest.match(/^---\\r?(\\n|$)/m);\n  if (!close || close.index === undefined) return { fm: null, body: raw };\n  return { fm: rest.slice(0, close.index), body: rest.slice(close.index + close[0].length) };\n}\n\n// A well-formed document can still hold a value nobody meant: `created: {{date}}` is valid\n// YAML for a flow map used as a mapping key, so it raises no error and stores\n// {\"{ date }\": null}. No error code can catch that, but yaml notices the stringified key, so\n// this reports it with the path instead (yaml's own warning has none, fires once per document,\n// and is what trains readers to discard stderr).\nfunction warnStringifiedKeys(relPath: string, doc: ReturnType<typeof parseDocument>, warnings: string[]): void {\n  let found = false;\n  // Nested, not top level: `created: {{date}}` puts the collection key one level down, inside\n  // the flow map that `{{...}}` parses as.\n  visit(doc, {\n    Pair(_key, pair) {\n      if (!isCollection(pair.key)) return undefined;\n      found = true;\n      return visit.BREAK;\n    },\n  });\n  // One per file: a template repeats the same mistake on every field it stamps.\n  if (found) warnings.push(`warning: ${relPath} frontmatter has a key that is itself a list or mapping, stored as text; this is usually an unrendered template placeholder like {{date}}`);\n}\n\n// Accept a clean parse, and one whose every error is unambiguous (ACCEPTED_YAML_CODES).\n// Anything else is quarantined: no frontmatter columns at all, and `_parse_error` carries the\n// reason. Recovering it would write values nobody wrote, which is worse than absence because\n// no query can see it. The file is still indexed -- content, links and sections never touch\n// frontmatter -- so a broken note stays searchable while it is being hunted for.\n// yaml's message continues onto a source excerpt, so the first line is the sentence -- minus\n// the colon that introduced the part being dropped.\nfunction firstLine(message: string): string {\n  return message.split('\\n')[0].replace(/:\\s*$/, '');\n}\n\nfunction parseFrontmatter(relPath: string, fm: string, warnings: string[]): { data: Record<string, unknown>; parseError: string | null } {\n  // logLevel silences yaml's own pathless warnings; warnStringifiedKeys re-reports the one\n  // that carries information, with the file it came from.\n  const doc = parseDocument(fm, { logLevel: 'silent' });\n  const refused = doc.errors.filter((err) => !ACCEPTED_YAML_CODES.has(err.code));\n  if (refused.length > 0) {\n    const detail = refused.length > 1 ? ` (and ${refused.length - 1} more)` : '';\n    const parseError = `${firstLine(refused[0].message)}${detail}`;\n    warnings.push(`warning: ${relPath} frontmatter did not parse, so none of it is indexed: ${parseError}`);\n    return { data: {}, parseError };\n  }\n\n  let data: unknown;\n  try {\n    data = doc.toJS();\n  } catch (err) {\n    // Reaches here with doc.errors empty: `title: **Bold**` parses, then opens an alias on\n    // materialisation. An empty error list is not a successful parse.\n    const parseError = firstLine((err as Error).message);\n    warnings.push(`warning: ${relPath} frontmatter did not parse, so none of it is indexed: ${parseError}`);\n    return { data: {}, parseError };\n  }\n\n  if (data === null || data === undefined) return { data: {}, parseError: null };\n  if (typeof data !== 'object' || Array.isArray(data)) {\n    const parseError = 'frontmatter is not a key-value mapping';\n    warnings.push(`warning: ${relPath} ${parseError}; none of it is indexed`);\n    return { data: {}, parseError };\n  }\n  warnStringifiedKeys(relPath, doc, warnings);\n  return { data: data as Record<string, unknown>, parseError: null };\n}\n\nexport function parseFile(file: FileStat, extractors: Feature[] = []): { doc: ParsedDoc; warnings: string[] } {\n  const raw = readFileSync(file.absPath, 'utf8');\n  const warnings: string[] = [];\n\n  const { fm, body: content } = splitFrontmatter(raw);\n  const { data, parseError } = fm === null ? { data: {} as Record<string, unknown>, parseError: null } : parseFrontmatter(file.relPath, fm, warnings);\n  const mapped: Record<string, string | number | bigint | null> = {};\n\n  for (const key of Object.keys(data)) {\n    if (RESERVED_COLUMNS.has(key)) {\n      warnings.push(`warning: ${file.relPath} has a frontmatter key named \"${key}\", which is reserved; ignoring it`);\n      continue;\n    }\n    const value = mapValue(data[key]);\n    if (typeof value === 'string' && looksLikeDatetime(value) && Number.isNaN(Date.parse(value))) {\n      warnings.push(`warning: ${file.relPath}: ${key} is not a valid date (${value}), so it is invisible to every date comparison`);\n    }\n    mapped[key] = value;\n  }\n\n  // title/summary are plain YAML strings -- whitespace-collapse only;\n  // the prose gets the full markdown strip.\n  const search = { title: normalizeText(data.title), summary: normalizeText(data.summary), text: stripText(content) };\n\n  return {\n    doc: {\n      relPath: file.relPath,\n      mtimeMs: file.mtimeMs,\n      ctimeMs: file.ctimeMs,\n      size: file.size,\n      presets: file.presets,\n      data: mapped,\n      parseError,\n      search,\n      extracted: Object.fromEntries(extractors.filter((f) => f.extract).map((f) => [f.name, f.extract?.(raw, content, search, data)])),\n    },\n    warnings,\n  };\n}\n"],"names":["RESERVED_COLUMNS","listFiles","looksLikeDatetime","normalizeDate","parseFile","toPosixPath","Set","ACCEPTED_YAML_CODES","normalizeText","value","undefined","String","replace","trim","stripText","withoutWikilinks","withoutMarkdown","removeMarkdown","withoutTables","relPath","separator","sep","split","join","NO_THROW","throwIfNoEntry","cfg","baseDir","coverage","Map","posixNeeded","presetNames","name","preset","presets","globSync","include","cwd","exclude","matched","set","get","add","embedding","embedEnabled","semanticPresets","filter","presetSemanticEnabled","files","keys","sort","absPath","st","statSync","isFile","embed","some","has","push","mtimeMs","ctimeMs","birthtimeMs","size","ISO_DATETIME","MEANT_AS_DATETIME","test","m","exec","date","time","zone","digits","offset","length","slice","normalized","Number","isNaN","Date","parse","mapValue","BigInt","isSafeInteger","JSON","stringify","splitFrontmatter","raw","open","match","fm","body","rest","close","index","warnStringifiedKeys","doc","warnings","found","visit","Pair","_key","pair","isCollection","key","BREAK","firstLine","message","parseFrontmatter","parseDocument","logLevel","refused","errors","err","code","detail","parseError","data","toJS","Array","isArray","file","extractors","readFileSync","content","mapped","Object","search","title","summary","text","extracted","fromEntries","f","extract","map"],"mappings":";;;;;;;;;;;QAYaA;eAAAA;;QA6CGC;eAAAA;;QAuDAC;eAAAA;;QAMAC;eAAAA;;QA8FAC;eAAAA;;QApKAC;eAAAA;;;sBAhDiC;wBACvB;qEACC;oBACwB;uBAEc;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAO1D,IAAML,mBAAmB,IAAIM,IAAI;IAAC;IAAQ;IAAU;IAAU;IAAS;IAAS;IAAgB;IAAW;IAAS;CAAW;AAEtI,uFAAuF;AACvF,6FAA6F;AAC7F,4FAA4F;AAC5F,8FAA8F;AAC9F,8FAA8F;AAC9F,4DAA4D;AAC5D,IAAMC,sBAAsB,IAAID,IAAI;IAAC;CAAmB;AAExD,SAASE,cAAcC,KAAc;IACnC,IAAIA,UAAU,QAAQA,UAAUC,WAAW,OAAO;IAClD,OAAOC,OAAOF,OAAOG,OAAO,CAAC,QAAQ,KAAKC,IAAI;AAChD;AAEA,6FAA6F;AAC7F,sFAAsF;AACtF,SAASC,UAAUL,KAAa;IAC9B,IAAMM,mBAAmBN,MAAMG,OAAO,CAAC,gCAAgC,MAAMA,OAAO,CAAC,qBAAqB;IAC1G,IAAMI,kBAAkBC,IAAAA,uBAAc,EAACF;IACvC,IAAMG,gBAAgBF,gBAAgBJ,OAAO,CAAC,2BAA2B,IAAIA,OAAO,CAAC,OAAO;IAC5F,OAAOJ,cAAcU;AACvB;AAcO,SAASb,YAAYc,OAAe;QAAEC,YAAAA,iEAAoBC,aAAG;IAClE,OAAOD,cAAc,OAAOD,QAAQG,KAAK,CAACF,WAAWG,IAAI,CAAC,OAAOJ;AACnE;AAEA,2FAA2F;AAC3F,8FAA8F;AAC9F,0CAA0C;AAC1C,IAAMK,WAAW;IAAEC,gBAAgB;AAAM;AAElC,SAASxB,UAAUyB,GAAW,EAAEC,OAAe;IACpD,IAAMC,WAAW,IAAIC;IACrB,IAAMC,cAAcT,aAAG,KAAK;QACvB,kCAAA,2BAAA;;QAAL,QAAK,YAAcU,IAAAA,oBAAW,EAACL,yBAA1B,SAAA,6BAAA,QAAA,yBAAA,iCAAgC;YAAhC,IAAMM,OAAN;YACH,IAAMC,SAASP,IAAIQ,OAAO,CAACF,KAAK;gBAC3B,mCAAA,4BAAA;;gBAAL,QAAK,aAAiBG,IAAAA,gBAAQ,EAACF,OAAOG,OAAO,EAAE;oBAAEC,KAAKV;oBAASW,SAASL,OAAOK,OAAO;gBAAC,uBAAlF,UAAA,8BAAA,SAAA,0BAAA,kCAAsF;oBAAtF,IAAMC,UAAN;wBAESX;oBADZ,IAAMT,UAAUW,cAAczB,YAAYkC,WAAWA;oBACrD,IAAMC,OAAMZ,gBAAAA,SAASa,GAAG,CAACtB,sBAAbS,2BAAAA,gBAAyB,IAAItB;oBACzCkC,IAAIE,GAAG,CAACV;oBACRJ,SAASY,GAAG,CAACrB,SAASqB;gBACxB;;gBALK;gBAAA;;;yBAAA,8BAAA;wBAAA;;;wBAAA;8BAAA;;;;QAMP;;QARK;QAAA;;;iBAAA,6BAAA;gBAAA;;;gBAAA;sBAAA;;;;IAUL,2EAA2E;IAC3E,IAAMG,YAAYC,IAAAA,qBAAY,EAAClB;IAC/B,IAAMmB,kBAAkBF,YAAY,IAAIrC,IAAIyB,IAAAA,oBAAW,EAACL,KAAKoB,MAAM,CAAC,SAACd;eAASe,IAAAA,8BAAqB,EAACrB,KAAKM;UAAU;IAEnH,IAAMgB,QAAoB,EAAE;QACvB,mCAAA,4BAAA;;QAAL,QAAK,aAAiB,AAAC,qBAAGpB,SAASqB,IAAI,IAAIC,IAAI,uBAA1C,UAAA,8BAAA,SAAA,0BAAA,kCAA8C;YAA9C,IAAM/B,WAAN;YACH,IAAMgC,UAAU5B,IAAAA,cAAI,EAACI,SAASR,WAAU,sDAAsD;YAC9F,yFAAyF;YACzF,uFAAuF;YACvF,IAAMiC,KAAKC,IAAAA,gBAAQ,EAACF,SAAS3B;YAC7B,IAAI,EAAC4B,eAAAA,yBAAAA,GAAIE,MAAM,KAAI;YACnB,IAAMpB,UAAU,AAAC,qBAAIN,SAASa,GAAG,CAACtB,WAA0B+B,IAAI;YAChE,IAAMK,QAAQV,oBAAoB,QAAQX,QAAQsB,IAAI,CAAC,SAACxB;uBAASa,gBAAgBY,GAAG,CAACzB;;YACrFgB,MAAMU,IAAI,CAAC;gBAAEvC,SAAAA;gBAASgC,SAAAA;gBAASQ,SAASP,GAAGO,OAAO;gBAAEC,SAASR,GAAGS,WAAW;gBAAEC,MAAMV,GAAGU,IAAI;gBAAE5B,SAAAA;gBAASqB,OAAAA;YAAM;QAC7G;;QATK;QAAA;;;iBAAA,8BAAA;gBAAA;;;gBAAA;sBAAA;;;;IAUL,OAAOP;AACT;AAkBA,iGAAiG;AACjG,mGAAmG;AACnG,IAAMe,eAAe;AAErB,6FAA6F;AAC7F,6FAA6F;AAC7F,IAAMC,oBAAoB;AAEnB,SAAS9D,kBAAkBO,KAAa;IAC7C,OAAOuD,kBAAkBC,IAAI,CAACxD;AAChC;AAIO,SAASN,cAAcM,KAAa;IACzC,IAAMyD,IAAIH,aAAaI,IAAI,CAAC1D;IAC5B,IAAIyD,MAAM,MAAM,OAAOzD;IACvB,IAA6ByD,sBAAAA,OAApBE,OAAoBF,OAAdG,OAAcH,OAARI,OAAQJ;IAC7B,IAAMK,SAASD,SAAS5D,aAAa4D,SAAS,MAAM,KAAKA,KAAK1D,OAAO,CAAC,KAAK;IAC3E,IAAM4D,SAASD,WAAW,KAAMD,iBAAAA,kBAAAA,OAAQ,KAAMC,OAAOE,MAAM,KAAK,IAAI,AAAC,GAAS,OAAPF,QAAO,SAAO,AAAC,GAAwBA,OAAtBA,OAAOG,KAAK,CAAC,GAAG,IAAG,KAAmB,OAAhBH,OAAOG,KAAK,CAAC;IAC3H,IAAMC,aAAa,AAAC,GAAUN,OAARD,MAAK,KAAUI,OAAPH,MAAc,OAAPG;IACrC,OAAOI,OAAOC,KAAK,CAACC,KAAKC,KAAK,CAACJ,eAAelE,QAAQkE;AACxD;AAEA,4FAA4F;AAC5F,4FAA4F;AAC5F,SAASK,SAASvE,KAAc;IAC9B,IAAIA,UAAU,QAAQA,UAAUC,WAAW,OAAO;IAClD,IAAI,OAAOD,UAAU,WAAW,OAAOwE,OAAOxE,QAAQ,IAAI;IAC1D,IAAI,OAAOA,UAAU,UAAU,OAAOmE,OAAOM,aAAa,CAACzE,SAASwE,OAAOxE,SAASA;IACpF,IAAI,OAAOA,UAAU,UAAU,OAAON,cAAcM;IACpD,OAAO0E,KAAKC,SAAS,CAAC3E;AACxB;AAEA,gEAAgE;AAChE,SAAS4E,iBAAiBC,GAAW;IACnC,IAAMC,OAAOD,IAAIE,KAAK,CAAC;IACvB,IAAI,CAACD,MAAM,OAAO;QAAEE,IAAI;QAAMC,MAAMJ;IAAI;IACxC,IAAMK,OAAOL,IAAIZ,KAAK,CAACa,IAAI,CAAC,EAAE,CAACd,MAAM;IACrC,IAAMmB,QAAQD,KAAKH,KAAK,CAAC;IACzB,IAAI,CAACI,SAASA,MAAMC,KAAK,KAAKnF,WAAW,OAAO;QAAE+E,IAAI;QAAMC,MAAMJ;IAAI;IACtE,OAAO;QAAEG,IAAIE,KAAKjB,KAAK,CAAC,GAAGkB,MAAMC,KAAK;QAAGH,MAAMC,KAAKjB,KAAK,CAACkB,MAAMC,KAAK,GAAGD,KAAK,CAAC,EAAE,CAACnB,MAAM;IAAE;AAC3F;AAEA,2FAA2F;AAC3F,8EAA8E;AAC9E,6FAA6F;AAC7F,+FAA+F;AAC/F,iDAAiD;AACjD,SAASqB,oBAAoB3E,OAAe,EAAE4E,GAAqC,EAAEC,QAAkB;IACrG,IAAIC,QAAQ;IACZ,4FAA4F;IAC5F,yCAAyC;IACzCC,IAAAA,WAAK,EAACH,KAAK;QACTI,MAAAA,SAAAA,KAAKC,IAAI,EAAEC,IAAI;YACb,IAAI,CAACC,IAAAA,kBAAY,EAACD,KAAKE,GAAG,GAAG,OAAO7F;YACpCuF,QAAQ;YACR,OAAOC,WAAK,CAACM,KAAK;QACpB;IACF;IACA,8EAA8E;IAC9E,IAAIP,OAAOD,SAAStC,IAAI,CAAC,AAAC,YAAmB,OAARvC,SAAQ;AAC/C;AAEA,wFAAwF;AACxF,8FAA8F;AAC9F,6FAA6F;AAC7F,4FAA4F;AAC5F,iFAAiF;AACjF,6FAA6F;AAC7F,oDAAoD;AACpD,SAASsF,UAAUC,OAAe;IAChC,OAAOA,QAAQpF,KAAK,CAAC,KAAK,CAAC,EAAE,CAACV,OAAO,CAAC,SAAS;AACjD;AAEA,SAAS+F,iBAAiBxF,OAAe,EAAEsE,EAAU,EAAEO,QAAkB;IACvE,yFAAyF;IACzF,wDAAwD;IACxD,IAAMD,MAAMa,IAAAA,mBAAa,EAACnB,IAAI;QAAEoB,UAAU;IAAS;IACnD,IAAMC,UAAUf,IAAIgB,MAAM,CAACjE,MAAM,CAAC,SAACkE;eAAQ,CAACzG,oBAAoBkD,GAAG,CAACuD,IAAIC,IAAI;;IAC5E,IAAIH,QAAQrC,MAAM,GAAG,GAAG;QACtB,IAAMyC,SAASJ,QAAQrC,MAAM,GAAG,IAAI,AAAC,SAA2B,OAAnBqC,QAAQrC,MAAM,GAAG,GAAE,YAAU;QAC1E,IAAM0C,aAAa,AAAC,GAAkCD,OAAhCT,UAAUK,OAAO,CAAC,EAAE,CAACJ,OAAO,GAAW,OAAPQ;QACtDlB,SAAStC,IAAI,CAAC,AAAC,YAA2EyD,OAAhEhG,SAAQ,0DAAmE,OAAXgG;QAC1F,OAAO;YAAEC,MAAM,CAAC;YAAGD,YAAAA;QAAW;IAChC;IAEA,IAAIC;IACJ,IAAI;QACFA,OAAOrB,IAAIsB,IAAI;IACjB,EAAE,OAAOL,KAAK;QACZ,uFAAuF;QACvF,kEAAkE;QAClE,IAAMG,cAAaV,UAAU,AAACO,IAAcN,OAAO;QACnDV,SAAStC,IAAI,CAAC,AAAC,YAA2EyD,OAAhEhG,SAAQ,0DAAmE,OAAXgG;QAC1F,OAAO;YAAEC,MAAM,CAAC;YAAGD,YAAAA;QAAW;IAChC;IAEA,IAAIC,SAAS,QAAQA,SAAS1G,WAAW,OAAO;QAAE0G,MAAM,CAAC;QAAGD,YAAY;IAAK;IAC7E,IAAI,CAAA,OAAOC,qCAAP,SAAOA,KAAG,MAAM,YAAYE,MAAMC,OAAO,CAACH,OAAO;QACnD,IAAMD,cAAa;QACnBnB,SAAStC,IAAI,CAAC,AAAC,YAAsByD,OAAXhG,SAAQ,KAAc,OAAXgG,aAAW;QAChD,OAAO;YAAEC,MAAM,CAAC;YAAGD,YAAAA;QAAW;IAChC;IACArB,oBAAoB3E,SAAS4E,KAAKC;IAClC,OAAO;QAAEoB,MAAMA;QAAiCD,YAAY;IAAK;AACnE;AAEO,SAAS/G,UAAUoH,IAAc;QAAEC,aAAAA,iEAAwB,EAAE;IAClE,IAAMnC,MAAMoC,IAAAA,oBAAY,EAACF,KAAKrE,OAAO,EAAE;IACvC,IAAM6C,WAAqB,EAAE;IAE7B,IAA8BX,oBAAAA,iBAAiBC,MAAvCG,KAAsBJ,kBAAtBI,IAAIC,AAAMiC,UAAYtC,kBAAlBK;IACZ,IAA6BD,OAAAA,OAAO,OAAO;QAAE2B,MAAM,CAAC;QAA8BD,YAAY;IAAK,IAAIR,iBAAiBa,KAAKrG,OAAO,EAAEsE,IAAIO,WAAlIoB,OAAqB3B,KAArB2B,MAAMD,aAAe1B,KAAf0B;IACd,IAAMS,SAA0D,CAAC;QAE5D,kCAAA,2BAAA;;QAAL,QAAK,YAAaC,OAAO5E,IAAI,CAACmE,0BAAzB,SAAA,6BAAA,QAAA,yBAAA,iCAAgC;YAAhC,IAAMb,MAAN;YACH,IAAIvG,iBAAiByD,GAAG,CAAC8C,MAAM;gBAC7BP,SAAStC,IAAI,CAAC,AAAC,YAAwD6C,OAA7CiB,KAAKrG,OAAO,EAAC,kCAAoC,OAAJoF,KAAI;gBAC3E;YACF;YACA,IAAM9F,QAAQuE,SAASoC,IAAI,CAACb,IAAI;YAChC,IAAI,OAAO9F,UAAU,YAAYP,kBAAkBO,UAAUmE,OAAOC,KAAK,CAACC,KAAKC,KAAK,CAACtE,SAAS;gBAC5FuF,SAAStC,IAAI,CAAC,AAAC,YAA4B6C,OAAjBiB,KAAKrG,OAAO,EAAC,MAAgCV,OAA5B8F,KAAI,0BAA8B,OAAN9F,OAAM;YAC/E;YACAmH,MAAM,CAACrB,IAAI,GAAG9F;QAChB;;QAVK;QAAA;;;iBAAA,6BAAA;gBAAA;;;gBAAA;sBAAA;;;;IAYL,oEAAoE;IACpE,0CAA0C;IAC1C,IAAMqH,SAAS;QAAEC,OAAOvH,cAAc4G,KAAKW,KAAK;QAAGC,SAASxH,cAAc4G,KAAKY,OAAO;QAAGC,MAAMnH,UAAU6G;IAAS;IAElH,OAAO;QACL5B,KAAK;YACH5E,SAASqG,KAAKrG,OAAO;YACrBwC,SAAS6D,KAAK7D,OAAO;YACrBC,SAAS4D,KAAK5D,OAAO;YACrBE,MAAM0D,KAAK1D,IAAI;YACf5B,SAASsF,KAAKtF,OAAO;YACrBkF,MAAMQ;YACNT,YAAAA;YACAW,QAAAA;YACAI,WAAWL,OAAOM,WAAW,CAACV,WAAW3E,MAAM,CAAC,SAACsF;uBAAMA,EAAEC,OAAO;eAAEC,GAAG,CAAC,SAACF;oBAAeA;uBAAT;oBAACA,EAAEpG,IAAI;qBAAEoG,aAAAA,EAAEC,OAAO,cAATD,iCAAAA,gBAAAA,GAAY9C,KAAKqC,SAASG,QAAQV;iBAAM;;QAChI;QACApB,UAAAA;IACF;AACF"}

package/dist/cjs/search-error.js.map

@@ -1 +0,0 @@
-{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/search-error.ts"],"sourcesContent":["import { SenseError } from './errors.ts';\n\n// FTS5 reads punctuation as syntax, so `end-to-end` parses as a filter on column `to` and\n// errors `no such column: to` -- true about the parse, misleading about the input. Matched as\n// \"not a bareword\": an operator list is a list to keep current, and missing a character costs\n// the remedy, not the error.\nconst FTS5_PUNCTUATION = /[^\\p{L}\\p{N}_\\s]/u;\n\nexport function searchError(err: Error, terms: string, scope?: string): Error {\n  const message = err.message;\n  if (!/no such column|fts5: syntax error|malformed MATCH/.test(message)) return err;\n  const suspects = (terms.match(/\\S+/g) ?? []).filter((t) => !t.startsWith('\"') && FTS5_PUNCTUATION.test(t));\n  // Blame the terms only when the failing token actually came from one -- a typo'd column\n  // in --where (or the tree's default scope) raises \"no such column\" through this same\n  // statement, and naming a term for it would state a false fact about the input.\n  const col = /no such column: (\\S+)/.exec(message)?.[1];\n  const fromTerms = col === undefined ? suspects.length > 0 : suspects.some((t) => t.split(/[^\\p{L}\\p{N}]+/u).includes(col));\n  if (fromTerms && suspects.length > 0) {\n    return new SenseError('SEARCH_SYNTAX', `${message} -- the punctuation in ${suspects.map((t) => `\\`${t}\\``).join(', ')} is FTS5 syntax, not literal text; search for it literally by double-quoting: '\"${suspects[0]}\"'. Searchable columns are title, summary, text.`);\n  }\n  if (col !== undefined && scope !== undefined) {\n    return new SenseError('SEARCH_SYNTAX', `${message} -- the where condition (${scope}) references it; frontmatter columns are listed by sense sql \"SELECT name FROM pragma_table_info('frontmatter')\".`);\n  }\n  return new SenseError('SEARCH_SYNTAX', `${message} -- searchable columns are title, summary, text; frontmatter fields are queried with --where or sense sql (list them with pragma_table_info('frontmatter')).`);\n}\n"],"names":["searchError","FTS5_PUNCTUATION","err","terms","scope","message","test","suspects","match","filter","t","startsWith","col","exec","fromTerms","undefined","length","some","split","includes","SenseError","map","join"],"mappings":";;;;+BAQgBA;;;eAAAA;;;wBARW;AAE3B,0FAA0F;AAC1F,8FAA8F;AAC9F,8FAA8F;AAC9F,6BAA6B;AAC7B,IAAMC,mBAAmB;AAElB,SAASD,YAAYE,GAAU,EAAEC,KAAa,EAAEC,KAAc;QAGjDD;QAIN;IANZ,IAAME,UAAUH,IAAIG,OAAO;IAC3B,IAAI,CAAC,oDAAoDC,IAAI,CAACD,UAAU,OAAOH;IAC/E,IAAMK,WAAW,EAACJ,eAAAA,MAAMK,KAAK,CAAC,qBAAZL,0BAAAA,eAAuB,EAAE,EAAEM,MAAM,CAAC,SAACC;eAAM,CAACA,EAAEC,UAAU,CAAC,QAAQV,iBAAiBK,IAAI,CAACI;;IACvG,wFAAwF;IACxF,qFAAqF;IACrF,gFAAgF;IAChF,IAAME,OAAM,QAAA,wBAAwBC,IAAI,CAACR,sBAA7B,4BAAA,KAAuC,CAAC,EAAE;IACtD,IAAMS,YAAYF,QAAQG,YAAYR,SAASS,MAAM,GAAG,IAAIT,SAASU,IAAI,CAAC,SAACP;eAAMA,EAAEQ,KAAK,CAAC,8maAAmBC,QAAQ,CAACP;;IACrH,IAAIE,aAAaP,SAASS,MAAM,GAAG,GAAG;QACpC,OAAO,IAAII,oBAAU,CAAC,iBAAiB,AAAC,GAAmCb,OAAjCF,SAAQ,2BAAsJE,OAA7HA,SAASc,GAAG,CAAC,SAACX;mBAAM,AAAC,IAAM,OAAFA,GAAE;WAAKY,IAAI,CAAC,OAAM,qFAA8F,OAAZf,QAAQ,CAAC,EAAE,EAAC;IACtN;IACA,IAAIK,QAAQG,aAAaX,UAAUW,WAAW;QAC5C,OAAO,IAAIK,oBAAU,CAAC,iBAAiB,AAAC,GAAqChB,OAAnCC,SAAQ,6BAAiC,OAAND,OAAM;IACrF;IACA,OAAO,IAAIgB,oBAAU,CAAC,iBAAiB,AAAC,GAAU,OAARf,SAAQ;AACpD"}

package/dist/cjs/segment.d.cts

@@ -1,2 +0,0 @@
-export declare function segmentField(text: string): string;
-export declare function segmentMatch(terms: string): string;

package/dist/cjs/segment.d.ts

@@ -1,2 +0,0 @@
-export declare function segmentField(text: string): string;
-export declare function segmentMatch(terms: string): string;

package/dist/cjs/segment.js

@@ -1,98 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", {
-    value: true
-});
-function _export(target, all) {
-    for(var name in all)Object.defineProperty(target, name, {
-        enumerable: true,
-        get: Object.getOwnPropertyDescriptor(all, name).get
-    });
-}
-_export(exports, {
-    get segmentField () {
-        return segmentField;
-    },
-    get segmentMatch () {
-        return segmentMatch;
-    }
-});
-// Word boundaries FTS5's tokenizers cannot find on their own.
-//
-// Segmentation is context-free: seg(query) appears inside seg(document) whenever the query is
-// a substring of the document. That is the contract -- substring semantics, what `grep` and
-// `LIKE '%..%'` give -- and it is what makes index and query agree on every input, always.
-// Earlier attempts used ICU word segmentation (Intl.Segmenter), which is context-dependent:
-// `东京都政府` splits as `东 | 京都 | 政府` while the query `东京` splits as one word, so the
-// index and the query disagree about identical text. No ICU here. No dictionary.
-// Scripts written without word spaces: a closed set of writing systems.
-// Script_Extensions, not Script: the katakana long-vowel mark ー is Script=Common.
-var SCRIPTS = '\\p{scx=Han}\\p{scx=Hiragana}\\p{scx=Katakana}\\p{scx=Thai}\\p{scx=Khmer}\\p{scx=Lao}\\p{scx=Myanmar}';
-// A run is script BASE characters with their combining marks attached; a bare mark after a
-// Latin letter (decomposed é) never starts one.
-var RUN = new RegExp("((?:[".concat(SCRIPTS, "]\\p{M}*)+)"), 'gu');
-var HAS_RUN = new RegExp("[".concat(SCRIPTS, "]"), 'u');
-// Grapheme = base char plus its marks. Context-free: index and query can never disagree.
-var GRAPHEME = RegExp("([^\\u0300-\\u036F\\u0483-\\u0489\\u0591-\\u05BD\\u05BF\\u05C1-\\u05C2\\u05C4-\\u05C5\\u05C7\\u0610-\\u061A\\u064B-\\u065F\\u0670\\u06D6-\\u06DC\\u06DF-\\u06E4\\u06E7-\\u06E8\\u06EA-\\u06ED\\u0711\\u0730-\\u074A\\u07A6-\\u07B0\\u07EB-\\u07F3\\u07FD\\u0816-\\u0819\\u081B-\\u0823\\u0825-\\u0827\\u0829-\\u082D\\u0859-\\u085B\\u0897-\\u089F\\u08CA-\\u08E1\\u08E3-\\u0903\\u093A-\\u093C\\u093E-\\u094F\\u0951-\\u0957\\u0962-\\u0963\\u0981-\\u0983\\u09BC\\u09BE-\\u09C4\\u09C7-\\u09C8\\u09CB-\\u09CD\\u09D7\\u09E2-\\u09E3\\u09FE\\u0A01-\\u0A03\\u0A3C\\u0A3E-\\u0A42\\u0A47-\\u0A48\\u0A4B-\\u0A4D\\u0A51\\u0A70-\\u0A71\\u0A75\\u0A81-\\u0A83\\u0ABC\\u0ABE-\\u0AC5\\u0AC7-\\u0AC9\\u0ACB-\\u0ACD\\u0AE2-\\u0AE3\\u0AFA-\\u0AFF\\u0B01-\\u0B03\\u0B3C\\u0B3E-\\u0B44\\u0B47-\\u0B48\\u0B4B-\\u0B4D\\u0B55-\\u0B57\\u0B62-\\u0B63\\u0B82\\u0BBE-\\u0BC2\\u0BC6-\\u0BC8\\u0BCA-\\u0BCD\\u0BD7\\u0C00-\\u0C04\\u0C3C\\u0C3E-\\u0C44\\u0C46-\\u0C48\\u0C4A-\\u0C4D\\u0C55-\\u0C56\\u0C62-\\u0C63\\u0C81-\\u0C83\\u0CBC\\u0CBE-\\u0CC4\\u0CC6-\\u0CC8\\u0CCA-\\u0CCD\\u0CD5-\\u0CD6\\u0CE2-\\u0CE3\\u0CF3\\u0D00-\\u0D03\\u0D3B-\\u0D3C\\u0D3E-\\u0D44\\u0D46-\\u0D48\\u0D4A-\\u0D4D\\u0D57\\u0D62-\\u0D63\\u0D81-\\u0D83\\u0DCA\\u0DCF-\\u0DD4\\u0DD6\\u0DD8-\\u0DDF\\u0DF2-\\u0DF3\\u0E31\\u0E34-\\u0E3A\\u0E47-\\u0E4E\\u0EB1\\u0EB4-\\u0EBC\\u0EC8-\\u0ECE\\u0F18-\\u0F19\\u0F35\\u0F37\\u0F39\\u0F3E-\\u0F3F\\u0F71-\\u0F84\\u0F86-\\u0F87\\u0F8D-\\u0F97\\u0F99-\\u0FBC\\u0FC6\\u102B-\\u103E\\u1056-\\u1059\\u105E-\\u1060\\u1062-\\u1064\\u1067-\\u106D\\u1071-\\u1074\\u1082-\\u108D\\u108F\\u109A-\\u109D\\u135D-\\u135F\\u1712-\\u1715\\u1732-\\u1734\\u1752-\\u1753\\u1772-\\u1773\\u17B4-\\u17D3\\u17DD\\u180B-\\u180D\\u180F\\u1885-\\u1886\\u18A9\\u1920-\\u192B\\u1930-\\u193B\\u1A17-\\u1A1B\\u1A55-\\u1A5E\\u1A60-\\u1A7C\\u1A7F\\u1AB0-\\u1ADD\\u1AE0-\\u1AEB\\u1B00-\\u1B04\\u1B34-\\u1B44\\u1B6B-\\u1B73\\u1B80-\\u1B82\\u1BA1-\\u1BAD\\u1BE6-\\u1BF3\\u1C24-\\u1C37\\u1CD0-\\u1CD2\\u1CD4-\\u1CE8\\u1CED\\u1CF4\\u1CF7-\\u1CF9\\u1DC0-\\u1DFF\\u20D0-\\u20F0\\u2CEF-\\u2CF1\\u2D7F\\u2DE0-\\u2DFF\\u302A-\\u302F\\u3099-\\u309A\\uA66F-\\uA672\\uA674-\\uA67D\\uA69E-\\uA69F\\uA6F0-\\uA6F1\\uA802\\uA806\\uA80B\\uA823-\\uA827\\uA82C\\uA880-\\uA881\\uA8B4-\\uA8C5\\uA8E0-\\uA8F1\\uA8FF\\uA926-\\uA92D\\uA947-\\uA953\\uA980-\\uA983\\uA9B3-\\uA9C0\\uA9E5\\uAA29-\\uAA36\\uAA43\\uAA4C-\\uAA4D\\uAA7B-\\uAA7D\\uAAB0\\uAAB2-\\uAAB4\\uAAB7-\\uAAB8\\uAABE-\\uAABF\\uAAC1\\uAAEB-\\uAAEF\\uAAF5-\\uAAF6\\uABE3-\\uABEA\\uABEC-\\uABED\\uFB1E\\uFE00-\\uFE0F\\uFE20-\\uFE2F\\u{101FD}\\u{102E0}\\u{10376}-\\u{1037A}\\u{10A01}-\\u{10A03}\\u{10A05}-\\u{10A06}\\u{10A0C}-\\u{10A0F}\\u{10A38}-\\u{10A3A}\\u{10A3F}\\u{10AE5}-\\u{10AE6}\\u{10D24}-\\u{10D27}\\u{10D69}-\\u{10D6D}\\u{10EAB}-\\u{10EAC}\\u{10EFA}-\\u{10EFF}\\u{10F46}-\\u{10F50}\\u{10F82}-\\u{10F85}\\u{11000}-\\u{11002}\\u{11038}-\\u{11046}\\u{11070}\\u{11073}-\\u{11074}\\u{1107F}-\\u{11082}\\u{110B0}-\\u{110BA}\\u{110C2}\\u{11100}-\\u{11102}\\u{11127}-\\u{11134}\\u{11145}-\\u{11146}\\u{11173}\\u{11180}-\\u{11182}\\u{111B3}-\\u{111C0}\\u{111C9}-\\u{111CC}\\u{111CE}-\\u{111CF}\\u{1122C}-\\u{11237}\\u{1123E}\\u{11241}\\u{112DF}-\\u{112EA}\\u{11300}-\\u{11303}\\u{1133B}-\\u{1133C}\\u{1133E}-\\u{11344}\\u{11347}-\\u{11348}\\u{1134B}-\\u{1134D}\\u{11357}\\u{11362}-\\u{11363}\\u{11366}-\\u{1136C}\\u{11370}-\\u{11374}\\u{113B8}-\\u{113C0}\\u{113C2}\\u{113C5}\\u{113C7}-\\u{113CA}\\u{113CC}-\\u{113D0}\\u{113D2}\\u{113E1}-\\u{113E2}\\u{11435}-\\u{11446}\\u{1145E}\\u{114B0}-\\u{114C3}\\u{115AF}-\\u{115B5}\\u{115B8}-\\u{115C0}\\u{115DC}-\\u{115DD}\\u{11630}-\\u{11640}\\u{116AB}-\\u{116B7}\\u{1171D}-\\u{1172B}\\u{1182C}-\\u{1183A}\\u{11930}-\\u{11935}\\u{11937}-\\u{11938}\\u{1193B}-\\u{1193E}\\u{11940}\\u{11942}-\\u{11943}\\u{119D1}-\\u{119D7}\\u{119DA}-\\u{119E0}\\u{119E4}\\u{11A01}-\\u{11A0A}\\u{11A33}-\\u{11A39}\\u{11A3B}-\\u{11A3E}\\u{11A47}\\u{11A51}-\\u{11A5B}\\u{11A8A}-\\u{11A99}\\u{11B60}-\\u{11B67}\\u{11C2F}-\\u{11C36}\\u{11C38}-\\u{11C3F}\\u{11C92}-\\u{11CA7}\\u{11CA9}-\\u{11CB6}\\u{11D31}-\\u{11D36}\\u{11D3A}\\u{11D3C}-\\u{11D3D}\\u{11D3F}-\\u{11D45}\\u{11D47}\\u{11D8A}-\\u{11D8E}\\u{11D90}-\\u{11D91}\\u{11D93}-\\u{11D97}\\u{11EF3}-\\u{11EF6}\\u{11F00}-\\u{11F01}\\u{11F03}\\u{11F34}-\\u{11F3A}\\u{11F3E}-\\u{11F42}\\u{11F5A}\\u{13440}\\u{13447}-\\u{13455}\\u{1611E}-\\u{1612F}\\u{16AF0}-\\u{16AF4}\\u{16B30}-\\u{16B36}\\u{16F4F}\\u{16F51}-\\u{16F87}\\u{16F8F}-\\u{16F92}\\u{16FE4}\\u{16FF0}-\\u{16FF1}\\u{1BC9D}-\\u{1BC9E}\\u{1CF00}-\\u{1CF2D}\\u{1CF30}-\\u{1CF46}\\u{1D165}-\\u{1D169}\\u{1D16D}-\\u{1D172}\\u{1D17B}-\\u{1D182}\\u{1D185}-\\u{1D18B}\\u{1D1AA}-\\u{1D1AD}\\u{1D242}-\\u{1D244}\\u{1DA00}-\\u{1DA36}\\u{1DA3B}-\\u{1DA6C}\\u{1DA75}\\u{1DA84}\\u{1DA9B}-\\u{1DA9F}\\u{1DAA1}-\\u{1DAAF}\\u{1E000}-\\u{1E006}\\u{1E008}-\\u{1E018}\\u{1E01B}-\\u{1E021}\\u{1E023}-\\u{1E024}\\u{1E026}-\\u{1E02A}\\u{1E08F}\\u{1E130}-\\u{1E136}\\u{1E2AE}\\u{1E2EC}-\\u{1E2EF}\\u{1E4EC}-\\u{1E4EF}\\u{1E5EE}-\\u{1E5EF}\\u{1E6E3}\\u{1E6E6}\\u{1E6EE}-\\u{1E6EF}\\u{1E6F5}\\u{1E8D0}-\\u{1E8D6}\\u{1E944}-\\u{1E94A}\\u{E0100}-\\u{E01EF}][\\u0300-\\u036F\\u0483-\\u0489\\u0591-\\u05BD\\u05BF\\u05C1-\\u05C2\\u05C4-\\u05C5\\u05C7\\u0610-\\u061A\\u064B-\\u065F\\u0670\\u06D6-\\u06DC\\u06DF-\\u06E4\\u06E7-\\u06E8\\u06EA-\\u06ED\\u0711\\u0730-\\u074A\\u07A6-\\u07B0\\u07EB-\\u07F3\\u07FD\\u0816-\\u0819\\u081B-\\u0823\\u0825-\\u0827\\u0829-\\u082D\\u0859-\\u085B\\u0897-\\u089F\\u08CA-\\u08E1\\u08E3-\\u0903\\u093A-\\u093C\\u093E-\\u094F\\u0951-\\u0957\\u0962-\\u0963\\u0981-\\u0983\\u09BC\\u09BE-\\u09C4\\u09C7-\\u09C8\\u09CB-\\u09CD\\u09D7\\u09E2-\\u09E3\\u09FE\\u0A01-\\u0A03\\u0A3C\\u0A3E-\\u0A42\\u0A47-\\u0A48\\u0A4B-\\u0A4D\\u0A51\\u0A70-\\u0A71\\u0A75\\u0A81-\\u0A83\\u0ABC\\u0ABE-\\u0AC5\\u0AC7-\\u0AC9\\u0ACB-\\u0ACD\\u0AE2-\\u0AE3\\u0AFA-\\u0AFF\\u0B01-\\u0B03\\u0B3C\\u0B3E-\\u0B44\\u0B47-\\u0B48\\u0B4B-\\u0B4D\\u0B55-\\u0B57\\u0B62-\\u0B63\\u0B82\\u0BBE-\\u0BC2\\u0BC6-\\u0BC8\\u0BCA-\\u0BCD\\u0BD7\\u0C00-\\u0C04\\u0C3C\\u0C3E-\\u0C44\\u0C46-\\u0C48\\u0C4A-\\u0C4D\\u0C55-\\u0C56\\u0C62-\\u0C63\\u0C81-\\u0C83\\u0CBC\\u0CBE-\\u0CC4\\u0CC6-\\u0CC8\\u0CCA-\\u0CCD\\u0CD5-\\u0CD6\\u0CE2-\\u0CE3\\u0CF3\\u0D00-\\u0D03\\u0D3B-\\u0D3C\\u0D3E-\\u0D44\\u0D46-\\u0D48\\u0D4A-\\u0D4D\\u0D57\\u0D62-\\u0D63\\u0D81-\\u0D83\\u0DCA\\u0DCF-\\u0DD4\\u0DD6\\u0DD8-\\u0DDF\\u0DF2-\\u0DF3\\u0E31\\u0E34-\\u0E3A\\u0E47-\\u0E4E\\u0EB1\\u0EB4-\\u0EBC\\u0EC8-\\u0ECE\\u0F18-\\u0F19\\u0F35\\u0F37\\u0F39\\u0F3E-\\u0F3F\\u0F71-\\u0F84\\u0F86-\\u0F87\\u0F8D-\\u0F97\\u0F99-\\u0FBC\\u0FC6\\u102B-\\u103E\\u1056-\\u1059\\u105E-\\u1060\\u1062-\\u1064\\u1067-\\u106D\\u1071-\\u1074\\u1082-\\u108D\\u108F\\u109A-\\u109D\\u135D-\\u135F\\u1712-\\u1715\\u1732-\\u1734\\u1752-\\u1753\\u1772-\\u1773\\u17B4-\\u17D3\\u17DD\\u180B-\\u180D\\u180F\\u1885-\\u1886\\u18A9\\u1920-\\u192B\\u1930-\\u193B\\u1A17-\\u1A1B\\u1A55-\\u1A5E\\u1A60-\\u1A7C\\u1A7F\\u1AB0-\\u1ADD\\u1AE0-\\u1AEB\\u1B00-\\u1B04\\u1B34-\\u1B44\\u1B6B-\\u1B73\\u1B80-\\u1B82\\u1BA1-\\u1BAD\\u1BE6-\\u1BF3\\u1C24-\\u1C37\\u1CD0-\\u1CD2\\u1CD4-\\u1CE8\\u1CED\\u1CF4\\u1CF7-\\u1CF9\\u1DC0-\\u1DFF\\u20D0-\\u20F0\\u2CEF-\\u2CF1\\u2D7F\\u2DE0-\\u2DFF\\u302A-\\u302F\\u3099-\\u309A\\uA66F-\\uA672\\uA674-\\uA67D\\uA69E-\\uA69F\\uA6F0-\\uA6F1\\uA802\\uA806\\uA80B\\uA823-\\uA827\\uA82C\\uA880-\\uA881\\uA8B4-\\uA8C5\\uA8E0-\\uA8F1\\uA8FF\\uA926-\\uA92D\\uA947-\\uA953\\uA980-\\uA983\\uA9B3-\\uA9C0\\uA9E5\\uAA29-\\uAA36\\uAA43\\uAA4C-\\uAA4D\\uAA7B-\\uAA7D\\uAAB0\\uAAB2-\\uAAB4\\uAAB7-\\uAAB8\\uAABE-\\uAABF\\uAAC1\\uAAEB-\\uAAEF\\uAAF5-\\uAAF6\\uABE3-\\uABEA\\uABEC-\\uABED\\uFB1E\\uFE00-\\uFE0F\\uFE20-\\uFE2F\\u{101FD}\\u{102E0}\\u{10376}-\\u{1037A}\\u{10A01}-\\u{10A03}\\u{10A05}-\\u{10A06}\\u{10A0C}-\\u{10A0F}\\u{10A38}-\\u{10A3A}\\u{10A3F}\\u{10AE5}-\\u{10AE6}\\u{10D24}-\\u{10D27}\\u{10D69}-\\u{10D6D}\\u{10EAB}-\\u{10EAC}\\u{10EFA}-\\u{10EFF}\\u{10F46}-\\u{10F50}\\u{10F82}-\\u{10F85}\\u{11000}-\\u{11002}\\u{11038}-\\u{11046}\\u{11070}\\u{11073}-\\u{11074}\\u{1107F}-\\u{11082}\\u{110B0}-\\u{110BA}\\u{110C2}\\u{11100}-\\u{11102}\\u{11127}-\\u{11134}\\u{11145}-\\u{11146}\\u{11173}\\u{11180}-\\u{11182}\\u{111B3}-\\u{111C0}\\u{111C9}-\\u{111CC}\\u{111CE}-\\u{111CF}\\u{1122C}-\\u{11237}\\u{1123E}\\u{11241}\\u{112DF}-\\u{112EA}\\u{11300}-\\u{11303}\\u{1133B}-\\u{1133C}\\u{1133E}-\\u{11344}\\u{11347}-\\u{11348}\\u{1134B}-\\u{1134D}\\u{11357}\\u{11362}-\\u{11363}\\u{11366}-\\u{1136C}\\u{11370}-\\u{11374}\\u{113B8}-\\u{113C0}\\u{113C2}\\u{113C5}\\u{113C7}-\\u{113CA}\\u{113CC}-\\u{113D0}\\u{113D2}\\u{113E1}-\\u{113E2}\\u{11435}-\\u{11446}\\u{1145E}\\u{114B0}-\\u{114C3}\\u{115AF}-\\u{115B5}\\u{115B8}-\\u{115C0}\\u{115DC}-\\u{115DD}\\u{11630}-\\u{11640}\\u{116AB}-\\u{116B7}\\u{1171D}-\\u{1172B}\\u{1182C}-\\u{1183A}\\u{11930}-\\u{11935}\\u{11937}-\\u{11938}\\u{1193B}-\\u{1193E}\\u{11940}\\u{11942}-\\u{11943}\\u{119D1}-\\u{119D7}\\u{119DA}-\\u{119E0}\\u{119E4}\\u{11A01}-\\u{11A0A}\\u{11A33}-\\u{11A39}\\u{11A3B}-\\u{11A3E}\\u{11A47}\\u{11A51}-\\u{11A5B}\\u{11A8A}-\\u{11A99}\\u{11B60}-\\u{11B67}\\u{11C2F}-\\u{11C36}\\u{11C38}-\\u{11C3F}\\u{11C92}-\\u{11CA7}\\u{11CA9}-\\u{11CB6}\\u{11D31}-\\u{11D36}\\u{11D3A}\\u{11D3C}-\\u{11D3D}\\u{11D3F}-\\u{11D45}\\u{11D47}\\u{11D8A}-\\u{11D8E}\\u{11D90}-\\u{11D91}\\u{11D93}-\\u{11D97}\\u{11EF3}-\\u{11EF6}\\u{11F00}-\\u{11F01}\\u{11F03}\\u{11F34}-\\u{11F3A}\\u{11F3E}-\\u{11F42}\\u{11F5A}\\u{13440}\\u{13447}-\\u{13455}\\u{1611E}-\\u{1612F}\\u{16AF0}-\\u{16AF4}\\u{16B30}-\\u{16B36}\\u{16F4F}\\u{16F51}-\\u{16F87}\\u{16F8F}-\\u{16F92}\\u{16FE4}\\u{16FF0}-\\u{16FF1}\\u{1BC9D}-\\u{1BC9E}\\u{1CF00}-\\u{1CF2D}\\u{1CF30}-\\u{1CF46}\\u{1D165}-\\u{1D169}\\u{1D16D}-\\u{1D172}\\u{1D17B}-\\u{1D182}\\u{1D185}-\\u{1D18B}\\u{1D1AA}-\\u{1D1AD}\\u{1D242}-\\u{1D244}\\u{1DA00}-\\u{1DA36}\\u{1DA3B}-\\u{1DA6C}\\u{1DA75}\\u{1DA84}\\u{1DA9B}-\\u{1DA9F}\\u{1DAA1}-\\u{1DAAF}\\u{1E000}-\\u{1E006}\\u{1E008}-\\u{1E018}\\u{1E01B}-\\u{1E021}\\u{1E023}-\\u{1E024}\\u{1E026}-\\u{1E02A}\\u{1E08F}\\u{1E130}-\\u{1E136}\\u{1E2AE}\\u{1E2EC}-\\u{1E2EF}\\u{1E4EC}-\\u{1E4EF}\\u{1E5EE}-\\u{1E5EF}\\u{1E6E3}\\u{1E6E6}\\u{1E6EE}-\\u{1E6EF}\\u{1E6F5}\\u{1E8D0}-\\u{1E8D6}\\u{1E944}-\\u{1E94A}\\u{E0100}-\\u{E01EF}]*)", "gu");
-// Token barrier between separate runs so their graphemes are never phrase-adjacent.
-// U+A7F7: a letter (so FTS5 keeps it as a token) that no one types into a search.
-var BARRIER = 'ꟷ';
-function graphemes(run) {
-    var _run_match;
-    return (_run_match = run.match(GRAPHEME)) !== null && _run_match !== void 0 ? _run_match : [];
-}
-function segmentField(text) {
-    if (!HAS_RUN.test(text)) return '';
-    var out = text.replace(RUN, function(run) {
-        return " ".concat(BARRIER, " ").concat(graphemes(run).join(' '), " ").concat(BARRIER, " ");
-    });
-    return out.replace(/\s+/g, ' ').trim();
-}
-// A `title:`/`summary:`/`text:` qualifier directly before a run that is about to become a
-// quoted grapheme phrase, so the rewrite can retarget it at the matching `_seg` column.
-var QUALIFIER = /(^|[\s(])(-?)(title|summary|text)\s*:\s*$/;
-function segmentMatch(terms) {
-    if (!HAS_RUN.test(terms)) return terms;
-    var out = '';
-    var quoted = false;
-    var pieces = terms.split(RUN); // split keeps captured runs at odd indices
-    for(var i = 0; i < pieces.length; i++){
-        if (i % 2 === 0) {
-            var _iteratorNormalCompletion = true, _didIteratorError = false, _iteratorError = undefined;
-            try {
-                for(var _iterator = pieces[i][Symbol.iterator](), _step; !(_iteratorNormalCompletion = (_step = _iterator.next()).done); _iteratorNormalCompletion = true){
-                    var ch = _step.value;
-                    if (ch === '"') quoted = !quoted;
-                }
-            } catch (err) {
-                _didIteratorError = true;
-                _iteratorError = err;
-            } finally{
-                try {
-                    if (!_iteratorNormalCompletion && _iterator.return != null) {
-                        _iterator.return();
-                    }
-                } finally{
-                    if (_didIteratorError) {
-                        throw _iteratorError;
-                    }
-                }
-            }
-            out += pieces[i];
-            continue;
-        }
-        if (quoted) {
-            out += pieces[i]; // an author's phrase is matched as written
-            continue;
-        }
-        var g = graphemes(pieces[i]);
-        var phrase = g.length > 1 ? '"'.concat(g.join(' '), '"') : g[0];
-        var m = out.match(QUALIFIER);
-        if (m) {
-            out = "".concat(out.slice(0, m.index)).concat(m[1]).concat(m[2]).concat(m[3], "_seg:").concat(phrase);
-        } else {
-            out += phrase;
-        }
-    }
-    return out;
-}
-/* CJS INTEROP */ if (exports.__esModule && exports.default) { try { Object.defineProperty(exports.default, '__esModule', { value: true }); for (var key in exports) { exports.default[key] = exports[key]; } } catch (_) {}; module.exports = exports.default; }

package/dist/cjs/segment.js.map

@@ -1 +0,0 @@
-{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/segment.ts"],"sourcesContent":["// Word boundaries FTS5's tokenizers cannot find on their own.\n//\n// Segmentation is context-free: seg(query) appears inside seg(document) whenever the query is\n// a substring of the document. That is the contract -- substring semantics, what `grep` and\n// `LIKE '%..%'` give -- and it is what makes index and query agree on every input, always.\n// Earlier attempts used ICU word segmentation (Intl.Segmenter), which is context-dependent:\n// `东京都政府` splits as `东 | 京都 | 政府` while the query `东京` splits as one word, so the\n// index and the query disagree about identical text. No ICU here. No dictionary.\n\n// Scripts written without word spaces: a closed set of writing systems.\n// Script_Extensions, not Script: the katakana long-vowel mark ー is Script=Common.\nconst SCRIPTS = '\\\\p{scx=Han}\\\\p{scx=Hiragana}\\\\p{scx=Katakana}\\\\p{scx=Thai}\\\\p{scx=Khmer}\\\\p{scx=Lao}\\\\p{scx=Myanmar}';\n// A run is script BASE characters with their combining marks attached; a bare mark after a\n// Latin letter (decomposed é) never starts one.\nconst RUN = new RegExp(`((?:[${SCRIPTS}]\\\\p{M}*)+)`, 'gu');\nconst HAS_RUN = new RegExp(`[${SCRIPTS}]`, 'u');\n// Grapheme = base char plus its marks. Context-free: index and query can never disagree.\nconst GRAPHEME = /(\\P{M}\\p{M}*)/gu;\n// Token barrier between separate runs so their graphemes are never phrase-adjacent.\n// U+A7F7: a letter (so FTS5 keeps it as a token) that no one types into a search.\nconst BARRIER = 'ꟷ';\n\nfunction graphemes(run: string): string[] {\n  return run.match(GRAPHEME) ?? [];\n}\n\n// Index side: '' when the field has no unspaced-script run (the common case, paid for by\n// nothing). Otherwise each run explodes into its graphemes, barrier-delimited from its\n// neighbors so graphemes from separate runs are never phrase-adjacent, and whitespace collapses.\nexport function segmentField(text: string): string {\n  if (!HAS_RUN.test(text)) return '';\n  const out = text.replace(RUN, (run) => ` ${BARRIER} ${graphemes(run).join(' ')} ${BARRIER} `);\n  return out.replace(/\\s+/g, ' ').trim();\n}\n\n// A `title:`/`summary:`/`text:` qualifier directly before a run that is about to become a\n// quoted grapheme phrase, so the rewrite can retarget it at the matching `_seg` column.\nconst QUALIFIER = /(^|[\\s(])(-?)(title|summary|text)\\s*:\\s*$/;\n\n// Query side: each unspaced run becomes a quoted phrase of its graphemes, matching how\n// segmentField indexed it. A qualifier ahead of such a run maps to its `_seg` column, since\n// segmented text lives only there. An author's own quoted phrase, and everything else, pass\n// through byte-identical.\nexport function segmentMatch(terms: string): string {\n  if (!HAS_RUN.test(terms)) return terms;\n  let out = '';\n  let quoted = false;\n  const pieces = terms.split(RUN); // split keeps captured runs at odd indices\n  for (let i = 0; i < pieces.length; i++) {\n    if (i % 2 === 0) {\n      for (const ch of pieces[i]) if (ch === '\"') quoted = !quoted;\n      out += pieces[i];\n      continue;\n    }\n    if (quoted) {\n      out += pieces[i]; // an author's phrase is matched as written\n      continue;\n    }\n    const g = graphemes(pieces[i]);\n    const phrase = g.length > 1 ? `\"${g.join(' ')}\"` : g[0];\n    const m = out.match(QUALIFIER);\n    if (m) {\n      out = `${out.slice(0, m.index)}${m[1]}${m[2]}${m[3]}_seg:${phrase}`;\n    } else {\n      out += phrase;\n    }\n  }\n  return out;\n}\n"],"names":["segmentField","segmentMatch","SCRIPTS","RUN","RegExp","HAS_RUN","GRAPHEME","BARRIER","graphemes","run","match","text","test","out","replace","join","trim","QUALIFIER","terms","quoted","pieces","split","i","length","ch","g","phrase","m","slice","index"],"mappings":";;;;;;;;;;;QA6BgBA;eAAAA;;QAcAC;eAAAA;;;AA3ChB,8DAA8D;AAC9D,EAAE;AACF,8FAA8F;AAC9F,4FAA4F;AAC5F,2FAA2F;AAC3F,4FAA4F;AAC5F,kFAAkF;AAClF,iFAAiF;AAEjF,wEAAwE;AACxE,kFAAkF;AAClF,IAAMC,UAAU;AAChB,2FAA2F;AAC3F,gDAAgD;AAChD,IAAMC,MAAM,IAAIC,OAAO,AAAC,QAAe,OAARF,SAAQ,gBAAc;AACrD,IAAMG,UAAU,IAAID,OAAO,AAAC,IAAW,OAARF,SAAQ,MAAI;AAC3C,yFAAyF;AACzF,IAAMI,WAAW;AACjB,oFAAoF;AACpF,kFAAkF;AAClF,IAAMC,UAAU;AAEhB,SAASC,UAAUC,GAAW;QACrBA;IAAP,QAAOA,aAAAA,IAAIC,KAAK,CAACJ,uBAAVG,wBAAAA,aAAuB,EAAE;AAClC;AAKO,SAAST,aAAaW,IAAY;IACvC,IAAI,CAACN,QAAQO,IAAI,CAACD,OAAO,OAAO;IAChC,IAAME,MAAMF,KAAKG,OAAO,CAACX,KAAK,SAACM;eAAQ,AAAC,IAAcD,OAAXD,SAAQ,KAA+BA,OAA5BC,UAAUC,KAAKM,IAAI,CAAC,MAAK,KAAW,OAARR,SAAQ;;IAC1F,OAAOM,IAAIC,OAAO,CAAC,QAAQ,KAAKE,IAAI;AACtC;AAEA,0FAA0F;AAC1F,wFAAwF;AACxF,IAAMC,YAAY;AAMX,SAAShB,aAAaiB,KAAa;IACxC,IAAI,CAACb,QAAQO,IAAI,CAACM,QAAQ,OAAOA;IACjC,IAAIL,MAAM;IACV,IAAIM,SAAS;IACb,IAAMC,SAASF,MAAMG,KAAK,CAAClB,MAAM,2CAA2C;IAC5E,IAAK,IAAImB,IAAI,GAAGA,IAAIF,OAAOG,MAAM,EAAED,IAAK;QACtC,IAAIA,IAAI,MAAM,GAAG;gBACV,kCAAA,2BAAA;;gBAAL,QAAK,YAAYF,MAAM,CAACE,EAAE,qBAArB,SAAA,6BAAA,QAAA,yBAAA;oBAAA,IAAME,KAAN;oBAAuB,IAAIA,OAAO,KAAKL,SAAS,CAACA;;;gBAAjD;gBAAA;;;yBAAA,6BAAA;wBAAA;;;wBAAA;8BAAA;;;;YACLN,OAAOO,MAAM,CAACE,EAAE;YAChB;QACF;QACA,IAAIH,QAAQ;YACVN,OAAOO,MAAM,CAACE,EAAE,EAAE,2CAA2C;YAC7D;QACF;QACA,IAAMG,IAAIjB,UAAUY,MAAM,CAACE,EAAE;QAC7B,IAAMI,SAASD,EAAEF,MAAM,GAAG,IAAI,AAAC,IAAe,OAAZE,EAAEV,IAAI,CAAC,MAAK,OAAKU,CAAC,CAAC,EAAE;QACvD,IAAME,IAAId,IAAIH,KAAK,CAACO;QACpB,IAAIU,GAAG;YACLd,MAAM,AAAC,GAA0Bc,OAAxBd,IAAIe,KAAK,CAAC,GAAGD,EAAEE,KAAK,GAAWF,OAAPA,CAAC,CAAC,EAAE,EAAUA,OAAPA,CAAC,CAAC,EAAE,EAAeD,OAAZC,CAAC,CAAC,EAAE,EAAC,SAAc,OAAPD;QAC7D,OAAO;YACLb,OAAOa;QACT;IACF;IACA,OAAOb;AACT"}

package/dist/cjs/sql-functions.js.map

@@ -1 +0,0 @@
-{"version":3,"sources":["/Users/kevin/Dev/OpenSource/ai/sensemaking/src/sql-functions.ts"],"sourcesContent":["import posix from 'node:path/posix';\nimport type { DatabaseSync } from 'node:sqlite';\nimport { segmentMatch } from './segment.ts';\n\n// has(field, value): JSON-array field -> membership, string field -> substring, NULL -> false.\nexport function registerFunctions(db: DatabaseSync, segmenting: boolean): void {\n  db.function('has', { deterministic: true, varargs: false }, (field: unknown, value: unknown): number => {\n    if (field === null || field === undefined) return 0;\n\n    const needle = String(value);\n\n    if (typeof field === 'string') {\n      if (field.startsWith('[')) {\n        try {\n          const parsed = JSON.parse(field);\n          if (Array.isArray(parsed)) {\n            return parsed.some((item) => String(item) === needle) ? 1 : 0;\n          }\n        } catch {}\n      }\n      return field.includes(needle) ? 1 : 0;\n    }\n\n    return String(field).includes(needle) ? 1 : 0;\n  });\n\n  // Unix basename(path, suffix?): the filename, minus suffix when it ends with one. SQLite\n  // itself has no filename function, and the LIKE tricks that fake one also match folder names.\n  db.function('basename', { deterministic: true, varargs: true }, (path: unknown, suffix?: unknown): string | null => {\n    if (path === null || path === undefined) return null;\n    const name = posix.basename(String(path));\n    const tail = suffix === null || suffix === undefined ? '' : String(suffix);\n    // POSIX: a suffix identical to the whole name is not removed (node's own basename strips it).\n    return tail !== '' && tail !== name && name.endsWith(tail) ? name.slice(0, -tail.length) : name;\n  });\n\n  // Raw `content MATCH '<unspaced text>'` cannot be rewritten behind the author's back, and\n  // matches nothing, so the same transform is available to hand-written SQL by name. `segmenting`\n  // is the same config predicate commands.ts's search() gates on (contentTokenize(cfg) ===\n  // undefined), passed in by open() rather than read back from persisted state.\n  db.function('segment', { deterministic: true, varargs: false }, (terms: unknown): string => {\n    const text = terms === null || terms === undefined ? '' : String(terms);\n    return segmenting ? segmentMatch(text) : text;\n  });\n}\n"],"names":["registerFunctions","db","segmenting","function","deterministic","varargs","field","value","undefined","needle","String","startsWith","parsed","JSON","parse","Array","isArray","some","item","includes","path","suffix","name","posix","basename","tail","endsWith","slice","length","terms","text","segmentMatch"],"mappings":";;;;+BAKgBA;;;eAAAA;;;4DALE;yBAEW;;;;;;AAGtB,SAASA,kBAAkBC,EAAgB,EAAEC,UAAmB;IACrED,GAAGE,QAAQ,CAAC,OAAO;QAAEC,eAAe;QAAMC,SAAS;IAAM,GAAG,SAACC,OAAgBC;QAC3E,IAAID,UAAU,QAAQA,UAAUE,WAAW,OAAO;QAElD,IAAMC,SAASC,OAAOH;QAEtB,IAAI,OAAOD,UAAU,UAAU;YAC7B,IAAIA,MAAMK,UAAU,CAAC,MAAM;gBACzB,IAAI;oBACF,IAAMC,SAASC,KAAKC,KAAK,CAACR;oBAC1B,IAAIS,MAAMC,OAAO,CAACJ,SAAS;wBACzB,OAAOA,OAAOK,IAAI,CAAC,SAACC;mCAASR,OAAOQ,UAAUT;6BAAU,IAAI;oBAC9D;gBACF,EAAE,eAAM,CAAC;YACX;YACA,OAAOH,MAAMa,QAAQ,CAACV,UAAU,IAAI;QACtC;QAEA,OAAOC,OAAOJ,OAAOa,QAAQ,CAACV,UAAU,IAAI;IAC9C;IAEA,yFAAyF;IACzF,8FAA8F;IAC9FR,GAAGE,QAAQ,CAAC,YAAY;QAAEC,eAAe;QAAMC,SAAS;IAAK,GAAG,SAACe,MAAeC;QAC9E,IAAID,SAAS,QAAQA,SAASZ,WAAW,OAAO;QAChD,IAAMc,OAAOC,cAAK,CAACC,QAAQ,CAACd,OAAOU;QACnC,IAAMK,OAAOJ,WAAW,QAAQA,WAAWb,YAAY,KAAKE,OAAOW;QACnE,8FAA8F;QAC9F,OAAOI,SAAS,MAAMA,SAASH,QAAQA,KAAKI,QAAQ,CAACD,QAAQH,KAAKK,KAAK,CAAC,GAAG,CAACF,KAAKG,MAAM,IAAIN;IAC7F;IAEA,0FAA0F;IAC1F,gGAAgG;IAChG,yFAAyF;IACzF,8EAA8E;IAC9ErB,GAAGE,QAAQ,CAAC,WAAW;QAAEC,eAAe;QAAMC,SAAS;IAAM,GAAG,SAACwB;QAC/D,IAAMC,OAAOD,UAAU,QAAQA,UAAUrB,YAAY,KAAKE,OAAOmB;QACjE,OAAO3B,aAAa6B,IAAAA,uBAAY,EAACD,QAAQA;IAC3C;AACF"}