@kolisachint/hoocode-agent 0.4.151 → 0.4.153
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/CHANGELOG.md +35 -0
- package/dist/core/embsearch/embsearch-progress.d.ts +33 -0
- package/dist/core/embsearch/embsearch-progress.d.ts.map +1 -0
- package/dist/core/embsearch/embsearch-progress.js +86 -0
- package/dist/core/embsearch/embsearch-progress.js.map +1 -0
- package/dist/core/embsearch/embsearch-service.d.ts +4 -0
- package/dist/core/embsearch/embsearch-service.d.ts.map +1 -1
- package/dist/core/embsearch/embsearch-service.js +6 -1
- package/dist/core/embsearch/embsearch-service.js.map +1 -1
- package/dist/core/startup-progress.d.ts +48 -0
- package/dist/core/startup-progress.d.ts.map +1 -0
- package/dist/core/startup-progress.js +50 -0
- package/dist/core/startup-progress.js.map +1 -0
- package/dist/core/system-prompt.d.ts.map +1 -1
- package/dist/core/system-prompt.js +7 -0
- package/dist/core/system-prompt.js.map +1 -1
- package/dist/core/tools/bash.d.ts.map +1 -1
- package/dist/core/tools/bash.js +1 -1
- package/dist/core/tools/bash.js.map +1 -1
- package/dist/core/tools/grep.d.ts.map +1 -1
- package/dist/core/tools/grep.js +1 -1
- package/dist/core/tools/grep.js.map +1 -1
- package/dist/core/tools/search.d.ts.map +1 -1
- package/dist/core/tools/search.js +0 -1
- package/dist/core/tools/search.js.map +1 -1
- package/dist/main.d.ts.map +1 -1
- package/dist/main.js +4 -48
- package/dist/main.js.map +1 -1
- package/dist/modes/interactive/components/footer.d.ts.map +1 -1
- package/dist/modes/interactive/components/footer.js +44 -0
- package/dist/modes/interactive/components/footer.js.map +1 -1
- package/dist/modes/interactive/interactive-mode.d.ts +3 -0
- package/dist/modes/interactive/interactive-mode.d.ts.map +1 -1
- package/dist/modes/interactive/interactive-mode.js +35 -5
- package/dist/modes/interactive/interactive-mode.js.map +1 -1
- package/examples/extensions/custom-provider-anthropic/package.json +1 -1
- package/examples/extensions/custom-provider-gitlab-duo/package.json +1 -1
- package/examples/extensions/sandbox/package.json +1 -1
- package/examples/extensions/with-deps/package.json +1 -1
- package/package.json +4 -4
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,40 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## [0.4.153] - 2026-07-23
|
|
4
|
+
|
|
5
|
+
### Added
|
|
6
|
+
|
|
7
|
+
- Startup progress bars for external tool binaries. First-run downloads of `fd`
|
|
8
|
+
and `ripgrep`, and the semantic-search index (its binary download **and** the
|
|
9
|
+
index build), now render a determinate progress bar as a transient line in
|
|
10
|
+
the footer — `received MB / total MB` with a percent, matching the voice
|
|
11
|
+
download bar — with an indeterminate byte count when the server sends no
|
|
12
|
+
`Content-Length`. The bars are fed by `ensureTool`'s existing byte-count
|
|
13
|
+
callback, run in the background so a slow first-run download never blocks
|
|
14
|
+
first paint, and each clears as its work settles. They live in a dedicated
|
|
15
|
+
`startupProgress` store, never as task-panel plan rows.
|
|
16
|
+
|
|
17
|
+
### Changed
|
|
18
|
+
|
|
19
|
+
- The startup semantic-search index no longer surfaces as a task-panel row. It
|
|
20
|
+
was created via `taskStore.create()` — indistinguishable from a main-agent
|
|
21
|
+
`TodoWrite` plan item — so its finished row lingered above the prompt for the
|
|
22
|
+
whole first turn. It now reports through the shared `startupProgress` footer
|
|
23
|
+
channel alongside the tool-download bars, dropped on `ready`/`skipped` and
|
|
24
|
+
shown as a transient notice on failure. The non-interactive stderr log path
|
|
25
|
+
is unchanged.
|
|
26
|
+
- Renamed the user-facing semantic-index startup messages from `embsearch`
|
|
27
|
+
jargon to plain language (e.g. `Building semantic search index –
|
|
28
|
+
{done}/{total} files ({pct}%)`, `Semantic search index ready ({chunkCount}
|
|
29
|
+
chunks)`, `... skipped`/`unavailable ({reason})`). Flags, settings, and
|
|
30
|
+
module paths are unchanged.
|
|
31
|
+
- The interactive startup preload no longer fetches the semantic-index binary:
|
|
32
|
+
when the index is enabled, `EmbsearchService` already downloads and builds it
|
|
33
|
+
through the same progress channel, so preloading it separately only fetched
|
|
34
|
+
the binary twice and raced.
|
|
35
|
+
|
|
36
|
+
## [0.4.152] - 2026-07-23
|
|
37
|
+
|
|
3
38
|
## [0.4.151] - 2026-07-23
|
|
4
39
|
|
|
5
40
|
## [0.4.150] - 2026-07-22
|
|
@@ -0,0 +1,33 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Routes semantic-index startup progress (`EmbsearchState`) to the right user
|
|
3
|
+
* surface: in interactive mode a transient footer line via the shared
|
|
4
|
+
* `startupProgress` store; otherwise a dim stderr log line.
|
|
5
|
+
*
|
|
6
|
+
* Extracted from main.ts so the mapping — which phase shows what, and which
|
|
7
|
+
* phases stay silent — is unit-testable without booting the whole CLI.
|
|
8
|
+
*/
|
|
9
|
+
import type { EmbsearchState } from "./embsearch-service.js";
|
|
10
|
+
/**
|
|
11
|
+
* Stable footer key for the semantic index. The binary download and the index
|
|
12
|
+
* build are one logical piece of startup work, so they share one evolving line
|
|
13
|
+
* (Downloading… → Building…) that disappears once the index is ready/skipped.
|
|
14
|
+
*/
|
|
15
|
+
export declare const SEMANTIC_INDEX_PROGRESS_KEY = "semantic-index";
|
|
16
|
+
export interface EmbsearchProgressSink {
|
|
17
|
+
/** True in the interactive TUI (route to the footer store); false logs to stderr. */
|
|
18
|
+
interactive: boolean;
|
|
19
|
+
/** Non-interactive log sink; defaults to a dim `console.error`. Injectable for tests. */
|
|
20
|
+
log?: (message: string) => void;
|
|
21
|
+
}
|
|
22
|
+
/**
|
|
23
|
+
* Apply one `EmbsearchState` update to the chosen surface.
|
|
24
|
+
*
|
|
25
|
+
* Interactive: downloading/indexing show a live footer line, unavailable shows a
|
|
26
|
+
* transient error notice, and ready/skipped drop the line. It is deliberately
|
|
27
|
+
* NOT a task-panel row — index progress is startup status, not agent work.
|
|
28
|
+
*
|
|
29
|
+
* Non-interactive: logs indexing/ready/skipped/unavailable, but stays silent on
|
|
30
|
+
* the per-chunk `downloading` ticks (they would spam the log) and on `idle`.
|
|
31
|
+
*/
|
|
32
|
+
export declare function reportEmbsearchProgress(state: EmbsearchState, sink: EmbsearchProgressSink): void;
|
|
33
|
+
//# sourceMappingURL=embsearch-progress.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"embsearch-progress.d.ts","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-progress.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAIH,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AAE7D;;;;GAIG;AACH,eAAO,MAAM,2BAA2B,mBAAmB,CAAC;AAE5D,MAAM,WAAW,qBAAqB;IACrC,qFAAqF;IACrF,WAAW,EAAE,OAAO,CAAC;IACrB,yFAAyF;IACzF,GAAG,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,KAAK,IAAI,CAAC;CAChC;AAED;;;;;;;;;GASG;AACH,wBAAgB,uBAAuB,CAAC,KAAK,EAAE,cAAc,EAAE,IAAI,EAAE,qBAAqB,GAAG,IAAI,CA2DhG","sourcesContent":["/**\n * Routes semantic-index startup progress (`EmbsearchState`) to the right user\n * surface: in interactive mode a transient footer line via the shared\n * `startupProgress` store; otherwise a dim stderr log line.\n *\n * Extracted from main.ts so the mapping — which phase shows what, and which\n * phases stay silent — is unit-testable without booting the whole CLI.\n */\n\nimport chalk from \"chalk\";\nimport { startupProgress } from \"../startup-progress.js\";\nimport type { EmbsearchState } from \"./embsearch-service.js\";\n\n/**\n * Stable footer key for the semantic index. The binary download and the index\n * build are one logical piece of startup work, so they share one evolving line\n * (Downloading… → Building…) that disappears once the index is ready/skipped.\n */\nexport const SEMANTIC_INDEX_PROGRESS_KEY = \"semantic-index\";\n\nexport interface EmbsearchProgressSink {\n\t/** True in the interactive TUI (route to the footer store); false logs to stderr. */\n\tinteractive: boolean;\n\t/** Non-interactive log sink; defaults to a dim `console.error`. Injectable for tests. */\n\tlog?: (message: string) => void;\n}\n\n/**\n * Apply one `EmbsearchState` update to the chosen surface.\n *\n * Interactive: downloading/indexing show a live footer line, unavailable shows a\n * transient error notice, and ready/skipped drop the line. It is deliberately\n * NOT a task-panel row — index progress is startup status, not agent work.\n *\n * Non-interactive: logs indexing/ready/skipped/unavailable, but stays silent on\n * the per-chunk `downloading` ticks (they would spam the log) and on `idle`.\n */\nexport function reportEmbsearchProgress(state: EmbsearchState, sink: EmbsearchProgressSink): void {\n\tif (sink.interactive) {\n\t\tswitch (state.phase) {\n\t\t\tcase \"downloading\":\n\t\t\t\tstartupProgress.set({\n\t\t\t\t\tkey: SEMANTIC_INDEX_PROGRESS_KEY,\n\t\t\t\t\tkind: \"download\",\n\t\t\t\t\tlabel: \"Semantic search index\",\n\t\t\t\t\treceivedBytes: state.receivedBytes,\n\t\t\t\t\ttotalBytes: state.totalBytes,\n\t\t\t\t});\n\t\t\t\treturn;\n\t\t\tcase \"indexing\":\n\t\t\t\tstartupProgress.set({\n\t\t\t\t\tkey: SEMANTIC_INDEX_PROGRESS_KEY,\n\t\t\t\t\tkind: \"work\",\n\t\t\t\t\tlabel: \"Building semantic search index\",\n\t\t\t\t\tdone: state.done,\n\t\t\t\t\ttotal: state.total,\n\t\t\t\t\tunit: \"files\",\n\t\t\t\t});\n\t\t\t\treturn;\n\t\t\tcase \"ready\":\n\t\t\tcase \"skipped\":\n\t\t\t\tstartupProgress.remove(SEMANTIC_INDEX_PROGRESS_KEY);\n\t\t\t\treturn;\n\t\t\tcase \"unavailable\":\n\t\t\t\tstartupProgress.set({\n\t\t\t\t\tkey: SEMANTIC_INDEX_PROGRESS_KEY,\n\t\t\t\t\tkind: \"error\",\n\t\t\t\t\tlabel: \"Semantic search index unavailable\",\n\t\t\t\t\tmessage: state.reason,\n\t\t\t\t});\n\t\t\t\treturn;\n\t\t\tcase \"idle\":\n\t\t\t\treturn;\n\t\t}\n\t}\n\n\tconst log = sink.log ?? ((message: string) => console.error(chalk.dim(message)));\n\tswitch (state.phase) {\n\t\tcase \"indexing\": {\n\t\t\tconst pct = Math.round((state.done / state.total) * 100);\n\t\t\tlog(`Building semantic search index – ${state.done}/${state.total} files (${pct}%)`);\n\t\t\treturn;\n\t\t}\n\t\tcase \"ready\":\n\t\t\tlog(`Semantic search index ready (${state.chunkCount} chunks)`);\n\t\t\treturn;\n\t\tcase \"skipped\":\n\t\t\tlog(`Semantic search index skipped (${state.reason})`);\n\t\t\treturn;\n\t\tcase \"unavailable\":\n\t\t\tlog(`Semantic search index unavailable (${state.reason})`);\n\t\t\treturn;\n\t\tdefault:\n\t\t\t// downloading (no per-chunk spam), idle: nothing to log.\n\t\t\treturn;\n\t}\n}\n"]}
|
|
@@ -0,0 +1,86 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Routes semantic-index startup progress (`EmbsearchState`) to the right user
|
|
3
|
+
* surface: in interactive mode a transient footer line via the shared
|
|
4
|
+
* `startupProgress` store; otherwise a dim stderr log line.
|
|
5
|
+
*
|
|
6
|
+
* Extracted from main.ts so the mapping — which phase shows what, and which
|
|
7
|
+
* phases stay silent — is unit-testable without booting the whole CLI.
|
|
8
|
+
*/
|
|
9
|
+
import chalk from "chalk";
|
|
10
|
+
import { startupProgress } from "../startup-progress.js";
|
|
11
|
+
/**
|
|
12
|
+
* Stable footer key for the semantic index. The binary download and the index
|
|
13
|
+
* build are one logical piece of startup work, so they share one evolving line
|
|
14
|
+
* (Downloading… → Building…) that disappears once the index is ready/skipped.
|
|
15
|
+
*/
|
|
16
|
+
export const SEMANTIC_INDEX_PROGRESS_KEY = "semantic-index";
|
|
17
|
+
/**
|
|
18
|
+
* Apply one `EmbsearchState` update to the chosen surface.
|
|
19
|
+
*
|
|
20
|
+
* Interactive: downloading/indexing show a live footer line, unavailable shows a
|
|
21
|
+
* transient error notice, and ready/skipped drop the line. It is deliberately
|
|
22
|
+
* NOT a task-panel row — index progress is startup status, not agent work.
|
|
23
|
+
*
|
|
24
|
+
* Non-interactive: logs indexing/ready/skipped/unavailable, but stays silent on
|
|
25
|
+
* the per-chunk `downloading` ticks (they would spam the log) and on `idle`.
|
|
26
|
+
*/
|
|
27
|
+
export function reportEmbsearchProgress(state, sink) {
|
|
28
|
+
if (sink.interactive) {
|
|
29
|
+
switch (state.phase) {
|
|
30
|
+
case "downloading":
|
|
31
|
+
startupProgress.set({
|
|
32
|
+
key: SEMANTIC_INDEX_PROGRESS_KEY,
|
|
33
|
+
kind: "download",
|
|
34
|
+
label: "Semantic search index",
|
|
35
|
+
receivedBytes: state.receivedBytes,
|
|
36
|
+
totalBytes: state.totalBytes,
|
|
37
|
+
});
|
|
38
|
+
return;
|
|
39
|
+
case "indexing":
|
|
40
|
+
startupProgress.set({
|
|
41
|
+
key: SEMANTIC_INDEX_PROGRESS_KEY,
|
|
42
|
+
kind: "work",
|
|
43
|
+
label: "Building semantic search index",
|
|
44
|
+
done: state.done,
|
|
45
|
+
total: state.total,
|
|
46
|
+
unit: "files",
|
|
47
|
+
});
|
|
48
|
+
return;
|
|
49
|
+
case "ready":
|
|
50
|
+
case "skipped":
|
|
51
|
+
startupProgress.remove(SEMANTIC_INDEX_PROGRESS_KEY);
|
|
52
|
+
return;
|
|
53
|
+
case "unavailable":
|
|
54
|
+
startupProgress.set({
|
|
55
|
+
key: SEMANTIC_INDEX_PROGRESS_KEY,
|
|
56
|
+
kind: "error",
|
|
57
|
+
label: "Semantic search index unavailable",
|
|
58
|
+
message: state.reason,
|
|
59
|
+
});
|
|
60
|
+
return;
|
|
61
|
+
case "idle":
|
|
62
|
+
return;
|
|
63
|
+
}
|
|
64
|
+
}
|
|
65
|
+
const log = sink.log ?? ((message) => console.error(chalk.dim(message)));
|
|
66
|
+
switch (state.phase) {
|
|
67
|
+
case "indexing": {
|
|
68
|
+
const pct = Math.round((state.done / state.total) * 100);
|
|
69
|
+
log(`Building semantic search index – ${state.done}/${state.total} files (${pct}%)`);
|
|
70
|
+
return;
|
|
71
|
+
}
|
|
72
|
+
case "ready":
|
|
73
|
+
log(`Semantic search index ready (${state.chunkCount} chunks)`);
|
|
74
|
+
return;
|
|
75
|
+
case "skipped":
|
|
76
|
+
log(`Semantic search index skipped (${state.reason})`);
|
|
77
|
+
return;
|
|
78
|
+
case "unavailable":
|
|
79
|
+
log(`Semantic search index unavailable (${state.reason})`);
|
|
80
|
+
return;
|
|
81
|
+
default:
|
|
82
|
+
// downloading (no per-chunk spam), idle: nothing to log.
|
|
83
|
+
return;
|
|
84
|
+
}
|
|
85
|
+
}
|
|
86
|
+
//# sourceMappingURL=embsearch-progress.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"embsearch-progress.js","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-progress.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,eAAe,EAAE,MAAM,wBAAwB,CAAC;AAGzD;;;;GAIG;AACH,MAAM,CAAC,MAAM,2BAA2B,GAAG,gBAAgB,CAAC;AAS5D;;;;;;;;;GASG;AACH,MAAM,UAAU,uBAAuB,CAAC,KAAqB,EAAE,IAA2B,EAAQ;IACjG,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;QACtB,QAAQ,KAAK,CAAC,KAAK,EAAE,CAAC;YACrB,KAAK,aAAa;gBACjB,eAAe,CAAC,GAAG,CAAC;oBACnB,GAAG,EAAE,2BAA2B;oBAChC,IAAI,EAAE,UAAU;oBAChB,KAAK,EAAE,uBAAuB;oBAC9B,aAAa,EAAE,KAAK,CAAC,aAAa;oBAClC,UAAU,EAAE,KAAK,CAAC,UAAU;iBAC5B,CAAC,CAAC;gBACH,OAAO;YACR,KAAK,UAAU;gBACd,eAAe,CAAC,GAAG,CAAC;oBACnB,GAAG,EAAE,2BAA2B;oBAChC,IAAI,EAAE,MAAM;oBACZ,KAAK,EAAE,gCAAgC;oBACvC,IAAI,EAAE,KAAK,CAAC,IAAI;oBAChB,KAAK,EAAE,KAAK,CAAC,KAAK;oBAClB,IAAI,EAAE,OAAO;iBACb,CAAC,CAAC;gBACH,OAAO;YACR,KAAK,OAAO,CAAC;YACb,KAAK,SAAS;gBACb,eAAe,CAAC,MAAM,CAAC,2BAA2B,CAAC,CAAC;gBACpD,OAAO;YACR,KAAK,aAAa;gBACjB,eAAe,CAAC,GAAG,CAAC;oBACnB,GAAG,EAAE,2BAA2B;oBAChC,IAAI,EAAE,OAAO;oBACb,KAAK,EAAE,mCAAmC;oBAC1C,OAAO,EAAE,KAAK,CAAC,MAAM;iBACrB,CAAC,CAAC;gBACH,OAAO;YACR,KAAK,MAAM;gBACV,OAAO;QACT,CAAC;IACF,CAAC;IAED,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC,OAAe,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;IACjF,QAAQ,KAAK,CAAC,KAAK,EAAE,CAAC;QACrB,KAAK,UAAU,EAAE,CAAC;YACjB,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC;YACzD,GAAG,CAAC,sCAAoC,KAAK,CAAC,IAAI,IAAI,KAAK,CAAC,KAAK,WAAW,GAAG,IAAI,CAAC,CAAC;YACrF,OAAO;QACR,CAAC;QACD,KAAK,OAAO;YACX,GAAG,CAAC,gCAAgC,KAAK,CAAC,UAAU,UAAU,CAAC,CAAC;YAChE,OAAO;QACR,KAAK,SAAS;YACb,GAAG,CAAC,kCAAkC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC;YACvD,OAAO;QACR,KAAK,aAAa;YACjB,GAAG,CAAC,sCAAsC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC;YAC3D,OAAO;QACR;YACC,yDAAyD;YACzD,OAAO;IACT,CAAC;AAAA,CACD","sourcesContent":["/**\n * Routes semantic-index startup progress (`EmbsearchState`) to the right user\n * surface: in interactive mode a transient footer line via the shared\n * `startupProgress` store; otherwise a dim stderr log line.\n *\n * Extracted from main.ts so the mapping — which phase shows what, and which\n * phases stay silent — is unit-testable without booting the whole CLI.\n */\n\nimport chalk from \"chalk\";\nimport { startupProgress } from \"../startup-progress.js\";\nimport type { EmbsearchState } from \"./embsearch-service.js\";\n\n/**\n * Stable footer key for the semantic index. The binary download and the index\n * build are one logical piece of startup work, so they share one evolving line\n * (Downloading… → Building…) that disappears once the index is ready/skipped.\n */\nexport const SEMANTIC_INDEX_PROGRESS_KEY = \"semantic-index\";\n\nexport interface EmbsearchProgressSink {\n\t/** True in the interactive TUI (route to the footer store); false logs to stderr. */\n\tinteractive: boolean;\n\t/** Non-interactive log sink; defaults to a dim `console.error`. Injectable for tests. */\n\tlog?: (message: string) => void;\n}\n\n/**\n * Apply one `EmbsearchState` update to the chosen surface.\n *\n * Interactive: downloading/indexing show a live footer line, unavailable shows a\n * transient error notice, and ready/skipped drop the line. It is deliberately\n * NOT a task-panel row — index progress is startup status, not agent work.\n *\n * Non-interactive: logs indexing/ready/skipped/unavailable, but stays silent on\n * the per-chunk `downloading` ticks (they would spam the log) and on `idle`.\n */\nexport function reportEmbsearchProgress(state: EmbsearchState, sink: EmbsearchProgressSink): void {\n\tif (sink.interactive) {\n\t\tswitch (state.phase) {\n\t\t\tcase \"downloading\":\n\t\t\t\tstartupProgress.set({\n\t\t\t\t\tkey: SEMANTIC_INDEX_PROGRESS_KEY,\n\t\t\t\t\tkind: \"download\",\n\t\t\t\t\tlabel: \"Semantic search index\",\n\t\t\t\t\treceivedBytes: state.receivedBytes,\n\t\t\t\t\ttotalBytes: state.totalBytes,\n\t\t\t\t});\n\t\t\t\treturn;\n\t\t\tcase \"indexing\":\n\t\t\t\tstartupProgress.set({\n\t\t\t\t\tkey: SEMANTIC_INDEX_PROGRESS_KEY,\n\t\t\t\t\tkind: \"work\",\n\t\t\t\t\tlabel: \"Building semantic search index\",\n\t\t\t\t\tdone: state.done,\n\t\t\t\t\ttotal: state.total,\n\t\t\t\t\tunit: \"files\",\n\t\t\t\t});\n\t\t\t\treturn;\n\t\t\tcase \"ready\":\n\t\t\tcase \"skipped\":\n\t\t\t\tstartupProgress.remove(SEMANTIC_INDEX_PROGRESS_KEY);\n\t\t\t\treturn;\n\t\t\tcase \"unavailable\":\n\t\t\t\tstartupProgress.set({\n\t\t\t\t\tkey: SEMANTIC_INDEX_PROGRESS_KEY,\n\t\t\t\t\tkind: \"error\",\n\t\t\t\t\tlabel: \"Semantic search index unavailable\",\n\t\t\t\t\tmessage: state.reason,\n\t\t\t\t});\n\t\t\t\treturn;\n\t\t\tcase \"idle\":\n\t\t\t\treturn;\n\t\t}\n\t}\n\n\tconst log = sink.log ?? ((message: string) => console.error(chalk.dim(message)));\n\tswitch (state.phase) {\n\t\tcase \"indexing\": {\n\t\t\tconst pct = Math.round((state.done / state.total) * 100);\n\t\t\tlog(`Building semantic search index – ${state.done}/${state.total} files (${pct}%)`);\n\t\t\treturn;\n\t\t}\n\t\tcase \"ready\":\n\t\t\tlog(`Semantic search index ready (${state.chunkCount} chunks)`);\n\t\t\treturn;\n\t\tcase \"skipped\":\n\t\t\tlog(`Semantic search index skipped (${state.reason})`);\n\t\t\treturn;\n\t\tcase \"unavailable\":\n\t\t\tlog(`Semantic search index unavailable (${state.reason})`);\n\t\t\treturn;\n\t\tdefault:\n\t\t\t// downloading (no per-chunk spam), idle: nothing to log.\n\t\t\treturn;\n\t}\n}\n"]}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"embsearch-service.d.ts","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AA2BH,MAAM,MAAM,cAAc,GACvB;IAAE,KAAK,EAAE,MAAM,CAAA;CAAE,GACjB;IAAE,KAAK,EAAE,SAAS,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,GACpC;IAAE,KAAK,EAAE,UAAU,CAAC;IAAC,IAAI,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,GAClD;IAAE,KAAK,EAAE,OAAO,CAAC;IAAC,UAAU,EAAE,MAAM,CAAA;CAAE,GACtC;IAAE,KAAK,EAAE,aAAa,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,CAAC;AAE5C,MAAM,WAAW,uBAAuB;IACvC,GAAG,EAAE,MAAM,CAAC;IACZ,gFAAgF;IAChF,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,wDAAwD;IACxD,cAAc,EAAE,MAAM,CAAC;IACvB,wDAAwD;IACxD,UAAU,CAAC,EAAE,CAAC,KAAK,EAAE,cAAc,KAAK,IAAI,CAAC;CAC7C;AAED,MAAM,WAAW,WAAW;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,gBAAiB,SAAQ,WAAW;IACpD,sFAAoF;IACpF,EAAE,EAAE,MAAM,CAAC;CACX;AAED,qBAAa,gBAAgB;IAC5B,OAAO,CAAC,QAAQ,CAAC,OAAO,CAA0B;IAClD,OAAO,CAAC,MAAM,CAA8B;IAC5C,OAAO,CAAC,IAAI,CAAwB;IACpC,OAAO,CAAC,KAAK,CAAqC;IAClD,OAAO,CAAC,QAAQ,CAAS;IAEzB,YAAY,OAAO,EAAE,uBAAuB,EAE3C;IAED,QAAQ,IAAI,cAAc,CAEzB;IAED,oFAAoF;IACpF,WAAW,IAAI,OAAO,CAErB;IAED,OAAO,CAAC,QAAQ;YAKF,aAAa;IAO3B;;;OAGG;IACG,KAAK,CAAC,MAAM,CAAC,EAAE,WAAW,GAAG,OAAO,CAAC,IAAI,CAAC,CAQ/C;YAEa,GAAG;YAmCH,iBAAiB;IAqF/B,OAAO,CAAC,WAAW;IAMnB,4DAA4D;IACtD,MAAM,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,SAAK,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC,CAE1D;IAED,wEAAwE;IAClE,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,SAAK,EAAE,IAAI,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,gBAAgB,EAAE,CAAC,CAqBpF;IAED;;;;;OAKG;IACH,kBAAkB,CACjB,GAAG,EAAE,MAAM,EACX,IAAI,EAAE,MAAM,GACV;QAAE,EAAE,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,GAAG,SAAS,CAU9E;YAEa,WAAW;IAYzB,kEAAkE;IAC5D,OAAO,IAAI,OAAO,CAAC,IAAI,CAAC,CAW7B;CACD;AAWD,wBAAgB,wBAAwB,CAAC,GAAG,EAAE,MAAM,EAAE,OAAO,EAAE,gBAAgB,GAAG,IAAI,CAMrF;AAED,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,MAAM,GAAG,gBAAgB,GAAG,SAAS,CAE7E;AAED,wBAAgB,0BAA0B,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,CAE5D","sourcesContent":["/**\n * Orchestrates semantic indexing and search for a repository.\n *\n * Lifecycle (all behind --enable-embsearchtools):\n * 1. `start()` — resolve the embsearch binary, scan the repo (ignore-aware),\n * apply the byte threshold. Under threshold → dormant. Over → spawn the\n * daemon, verify the backend is not the mock embedder, then index changed\n * files in the background in small batches, reporting progress.\n * 2. `search()` — top-k semantic query, mapping chunk ids back to\n * `path:start-end` via the sidecar metadata.\n * 3. `dispose()` — save + close the daemon.\n *\n * Every failure degrades to `unavailable` with a reason; nothing here ever\n * blocks session startup or affects grep/find.\n */\n\nimport { readFileSync } from \"fs\";\nimport { minimatch } from \"minimatch\";\nimport { ensureTool } from \"../../utils/tools-manager.js\";\nimport { chunkFile } from \"./chunker.js\";\nimport { EmbSearchClient } from \"./client.js\";\nimport {\n\temptyIndexMeta,\n\ttype FileMeta,\n\tgetEmbsearchStoreDir,\n\tgetVectorStoreDir,\n\thashContent,\n\ttype IndexMeta,\n\tloadIndexMeta,\n\tsaveIndexMeta,\n} from \"./index-meta.js\";\nimport { type RepoScanFile, scanRepo } from \"./repo-scan.js\";\n\n/** Chunks per bulk request. Small enough that a concurrent query is never\n * stuck long behind one padded batch inference. */\nconst BULK_BATCH_SIZE = 48;\n/** Yield between batches so background indexing doesn't starve the session. */\nconst BATCH_YIELD_MS = 15;\n/** The Rust mock backend's model id — semantically meaningless, never index with it. */\nconst MOCK_MODEL_ID = \"mock-hash-v1\";\n\nexport type EmbsearchState =\n\t| { phase: \"idle\" }\n\t| { phase: \"skipped\"; reason: string }\n\t| { phase: \"indexing\"; done: number; total: number }\n\t| { phase: \"ready\"; chunkCount: number }\n\t| { phase: \"unavailable\"; reason: string };\n\nexport interface EmbsearchServiceOptions {\n\tcwd: string;\n\t/** Explicit binary path (settings override). Default: \"embsearch\" from PATH. */\n\tbinaryPath?: string;\n\t/** Minimum indexable bytes before indexing kicks in. */\n\tthresholdBytes: number;\n\t/** Progress callback for UI (footer / stderr lines). */\n\tonProgress?: (state: EmbsearchState) => void;\n}\n\nexport interface SemanticHit {\n\tpath: string;\n\tstartLine: number;\n\tendLine: number;\n\tscore: number;\n}\n\nexport interface SemanticChunkHit extends SemanticHit {\n\t/** Per-build chunk id (`relpath#index`) — the fusion identity for hybrid search. */\n\tid: string;\n}\n\nexport class EmbsearchService {\n\tprivate readonly options: EmbsearchServiceOptions;\n\tprivate client: EmbSearchClient | undefined;\n\tprivate meta: IndexMeta | undefined;\n\tprivate state: EmbsearchState = { phase: \"idle\" };\n\tprivate disposed = false;\n\n\tconstructor(options: EmbsearchServiceOptions) {\n\t\tthis.options = options;\n\t}\n\n\tgetState(): EmbsearchState {\n\t\treturn this.state;\n\t}\n\n\t/** Semantic search is usable (index ready, or still building with partial data). */\n\tisAvailable(): boolean {\n\t\treturn this.state.phase === \"ready\" || this.state.phase === \"indexing\";\n\t}\n\n\tprivate setState(state: EmbsearchState): void {\n\t\tthis.state = state;\n\t\tthis.options.onProgress?.(state);\n\t}\n\n\tprivate async resolveBinary(): Promise<string | undefined> {\n\t\tif (this.options.binaryPath) {\n\t\t\treturn this.options.binaryPath;\n\t\t}\n\t\treturn await ensureTool(\"embsearch\", true);\n\t}\n\n\t/**\n\t * Scan, threshold-check, and (when needed) index in the background.\n\t * Resolves when indexing completes or the feature settles dormant.\n\t */\n\tasync start(signal?: AbortSignal): Promise<void> {\n\t\ttry {\n\t\t\tawait this.run(signal);\n\t\t} catch (e) {\n\t\t\tconst reason = e instanceof Error ? e.message : String(e);\n\t\t\tthis.setState({ phase: \"unavailable\", reason });\n\t\t\tawait this.closeClient();\n\t\t}\n\t}\n\n\tprivate async run(signal?: AbortSignal): Promise<void> {\n\t\tconst binary = await this.resolveBinary();\n\t\tif (!binary) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"unavailable\",\n\t\t\t\treason: \"embsearch binary not found (PATH or embsearchBinaryPath setting)\",\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst scan = scanRepo(this.options.cwd, signal);\n\t\tif (scan.totalBytes < this.options.thresholdBytes) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"skipped\",\n\t\t\t\treason: `repo under threshold (${scan.totalBytes} < ${this.options.thresholdBytes} bytes)`,\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst storeDir = getEmbsearchStoreDir(this.options.cwd);\n\t\tthis.client = new EmbSearchClient({ binaryPath: binary, storePath: getVectorStoreDir(storeDir) });\n\t\tawait this.client.ready();\n\n\t\tconst info = await this.client.info();\n\t\tif (info.modelId === MOCK_MODEL_ID) {\n\t\t\tthrow new Error(\"embsearch binary uses the mock embedder (not semantic); install an onnx build\");\n\t\t}\n\n\t\t// Missing/stale sidecar (format, chunker, or model changed) → clean rebuild.\n\t\tthis.meta = loadIndexMeta(storeDir, info.modelId) ?? emptyIndexMeta(this.options.cwd, info.modelId);\n\t\tthis.meta.lastUsedMs = Date.now();\n\n\t\tawait this.indexChangedFiles(scan.files, storeDir, signal);\n\t}\n\n\tprivate async indexChangedFiles(files: RepoScanFile[], storeDir: string, signal?: AbortSignal): Promise<void> {\n\t\tconst meta = this.meta!;\n\t\tconst client = this.client!;\n\n\t\t// Diff scan vs sidecar: cheap mtime+size check first, hash only on delta.\n\t\tconst toIndex: Array<{ file: RepoScanFile; content: string; hash: string }> = [];\n\t\tconst seen = new Set<string>();\n\t\tfor (const file of files) {\n\t\t\tseen.add(file.rel);\n\t\t\tconst known = meta.files[file.rel];\n\t\t\tif (known && known.mtimeMs === file.mtimeMs && known.size === file.size) continue;\n\t\t\tlet content: string;\n\t\t\ttry {\n\t\t\t\tcontent = readFileSync(file.abs, \"utf-8\");\n\t\t\t} catch {\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tconst hash = hashContent(content);\n\t\t\tif (known && known.hash === hash) {\n\t\t\t\t// Touched but unchanged — refresh stat info only.\n\t\t\t\tknown.mtimeMs = file.mtimeMs;\n\t\t\t\tknown.size = file.size;\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\ttoIndex.push({ file, content, hash });\n\t\t}\n\t\tconst toRemove = Object.keys(meta.files).filter((rel) => !seen.has(rel));\n\n\t\tif (toIndex.length === 0 && toRemove.length === 0) {\n\t\t\tsaveIndexMeta(storeDir, meta);\n\t\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t\t\treturn;\n\t\t}\n\n\t\t// Chunk changed files; count total upserts for exact progress.\n\t\tconst work: Array<{ rel: string; fileMeta: FileMeta; chunks: Array<{ id: string; text: string }> }> = [];\n\t\tlet totalChunks = 0;\n\t\tfor (const { file, content, hash } of toIndex) {\n\t\t\tconst chunks = chunkFile(file.rel, content);\n\t\t\twork.push({\n\t\t\t\trel: file.rel,\n\t\t\t\tfileMeta: {\n\t\t\t\t\tmtimeMs: file.mtimeMs,\n\t\t\t\t\tsize: file.size,\n\t\t\t\t\thash,\n\t\t\t\t\tchunks: chunks.map((c) => [c.startLine, c.endLine]),\n\t\t\t\t},\n\t\t\t\tchunks: chunks.map((c) => ({ id: c.id, text: c.text })),\n\t\t\t});\n\t\t\ttotalChunks += chunks.length;\n\t\t}\n\n\t\tthis.setState({ phase: \"indexing\", done: 0, total: totalChunks });\n\n\t\t// Drop vectors of deleted files and superseded chunk tails.\n\t\tfor (const rel of toRemove) {\n\t\t\tfor (let i = 0; i < meta.files[rel].chunks.length; i++) await client.remove(`${rel}#${i}`);\n\t\t\tdelete meta.files[rel];\n\t\t}\n\n\t\tlet done = 0;\n\t\tfor (const item of work) {\n\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\tconst oldChunkCount = meta.files[item.rel]?.chunks.length ?? 0;\n\t\t\t// Remove old chunks beyond the new count (upsert covers the rest).\n\t\t\tfor (let i = item.chunks.length; i < oldChunkCount; i++) await client.remove(`${item.rel}#${i}`);\n\n\t\t\tfor (let offset = 0; offset < item.chunks.length; offset += BULK_BATCH_SIZE) {\n\t\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\t\tconst batch = item.chunks.slice(offset, offset + BULK_BATCH_SIZE);\n\t\t\t\tawait client.bulk(batch);\n\t\t\t\tdone += batch.length;\n\t\t\t\tthis.setState({ phase: \"indexing\", done, total: totalChunks });\n\t\t\t\t// Yield so queries and the event loop stay responsive.\n\t\t\t\tawait new Promise((resolve) => setTimeout(resolve, BATCH_YIELD_MS));\n\t\t\t}\n\t\t\tmeta.files[item.rel] = item.fileMeta;\n\t\t}\n\n\t\tawait client.compact();\n\t\tawait client.save();\n\t\tsaveIndexMeta(storeDir, meta);\n\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t}\n\n\tprivate countChunks(meta: IndexMeta): number {\n\t\tlet n = 0;\n\t\tfor (const rel of Object.keys(meta.files)) n += meta.files[rel].chunks.length;\n\t\treturn n;\n\t}\n\n\t/** Top-`k` semantic hits as `path` + line range + score. */\n\tasync search(query: string, k = 10): Promise<SemanticHit[]> {\n\t\treturn await this.searchChunks(query, k);\n\t}\n\n\t/** Top-`k` semantic hits including their chunk ids, for rank fusion. */\n\tasync searchChunks(query: string, k = 10, glob?: string): Promise<SemanticChunkHit[]> {\n\t\tif (!this.client || this.client.isClosed || !this.meta) {\n\t\t\tthrow new Error(\"semantic index is not available\");\n\t\t}\n\t\tconst results = await this.client.query(query, k);\n\t\tconst hits: SemanticChunkHit[] = [];\n\t\tconst matchGlob = (rel: string): boolean => {\n\t\t\tif (!glob) return true;\n\t\t\treturn minimatch(rel, glob, { dot: true, matchBase: !glob.includes(\"/\") });\n\t\t};\n\t\tfor (const result of results) {\n\t\t\tconst sep = result.id.lastIndexOf(\"#\");\n\t\t\tif (sep === -1) continue;\n\t\t\tconst rel = result.id.slice(0, sep);\n\t\t\tif (!matchGlob(rel)) continue;\n\t\t\tconst chunkIndex = Number.parseInt(result.id.slice(sep + 1), 10);\n\t\t\tconst range = this.meta.files[rel]?.chunks[chunkIndex];\n\t\t\tif (!range) continue;\n\t\t\thits.push({ id: result.id, path: rel, startLine: range[0], endLine: range[1], score: result.score });\n\t\t}\n\t\treturn hits;\n\t}\n\n\t/**\n\t * Resolve a repo-relative path + line to its enclosing indexed chunk, or\n\t * undefined when the file/line is not covered by the index. Chunks overlap\n\t * by a few lines; the first (lowest-index) containing chunk wins so the\n\t * mapping is deterministic.\n\t */\n\tfindEnclosingChunk(\n\t\trel: string,\n\t\tline: number,\n\t): { id: string; path: string; startLine: number; endLine: number } | undefined {\n\t\tconst file = this.meta?.files[rel];\n\t\tif (!file) return undefined;\n\t\tfor (let i = 0; i < file.chunks.length; i++) {\n\t\t\tconst [startLine, endLine] = file.chunks[i];\n\t\t\tif (line >= startLine && line <= endLine) {\n\t\t\t\treturn { id: `${rel}#${i}`, path: rel, startLine, endLine };\n\t\t\t}\n\t\t}\n\t\treturn undefined;\n\t}\n\n\tprivate async closeClient(): Promise<void> {\n\t\tconst client = this.client;\n\t\tthis.client = undefined;\n\t\tif (client && !client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait client.close();\n\t\t\t} catch {\n\t\t\t\t// already dead\n\t\t\t}\n\t\t}\n\t}\n\n\t/** Persist state and shut the daemon down. Safe to call twice. */\n\tasync dispose(): Promise<void> {\n\t\tif (this.disposed) return;\n\t\tthis.disposed = true;\n\t\tif (this.client && !this.client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait this.client.save();\n\t\t\t} catch {\n\t\t\t\t// daemon may have exited; nothing to save\n\t\t\t}\n\t\t}\n\t\tawait this.closeClient();\n\t}\n}\n\n// --- Per-cwd service registry ---\n//\n// The search tool is constructed by the generic tool factory table and\n// only receives `cwd`; the service is created later during session init (flag\n// gated). This registry connects the two without threading a service instance\n// through every layer between main.ts and the tool factories.\n\nconst services = new Map<string, EmbsearchService>();\n\nexport function registerEmbsearchService(cwd: string, service: EmbsearchService): void {\n\tconst old = services.get(cwd);\n\tif (old && old !== service) {\n\t\told.dispose().catch(() => {});\n\t}\n\tservices.set(cwd, service);\n}\n\nexport function getEmbsearchService(cwd: string): EmbsearchService | undefined {\n\treturn services.get(cwd);\n}\n\nexport function unregisterEmbsearchService(cwd: string): void {\n\tservices.delete(cwd);\n}\n"]}
|
|
1
|
+
{"version":3,"file":"embsearch-service.d.ts","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AA2BH,MAAM,MAAM,cAAc,GACvB;IAAE,KAAK,EAAE,MAAM,CAAA;CAAE,GACjB;IAAE,KAAK,EAAE,SAAS,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,GACpC;IAAE,KAAK,EAAE,aAAa,CAAC;IAAC,aAAa,EAAE,MAAM,CAAC;IAAC,UAAU,EAAE,MAAM,GAAG,IAAI,CAAA;CAAE,GAC1E;IAAE,KAAK,EAAE,UAAU,CAAC;IAAC,IAAI,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,GAClD;IAAE,KAAK,EAAE,OAAO,CAAC;IAAC,UAAU,EAAE,MAAM,CAAA;CAAE,GACtC;IAAE,KAAK,EAAE,aAAa,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,CAAC;AAE5C,MAAM,WAAW,uBAAuB;IACvC,GAAG,EAAE,MAAM,CAAC;IACZ,gFAAgF;IAChF,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,wDAAwD;IACxD,cAAc,EAAE,MAAM,CAAC;IACvB,wDAAwD;IACxD,UAAU,CAAC,EAAE,CAAC,KAAK,EAAE,cAAc,KAAK,IAAI,CAAC;CAC7C;AAED,MAAM,WAAW,WAAW;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,gBAAiB,SAAQ,WAAW;IACpD,sFAAoF;IACpF,EAAE,EAAE,MAAM,CAAC;CACX;AAED,qBAAa,gBAAgB;IAC5B,OAAO,CAAC,QAAQ,CAAC,OAAO,CAA0B;IAClD,OAAO,CAAC,MAAM,CAA8B;IAC5C,OAAO,CAAC,IAAI,CAAwB;IACpC,OAAO,CAAC,KAAK,CAAqC;IAClD,OAAO,CAAC,QAAQ,CAAS;IAEzB,YAAY,OAAO,EAAE,uBAAuB,EAE3C;IAED,QAAQ,IAAI,cAAc,CAEzB;IAED,oFAAoF;IACpF,WAAW,IAAI,OAAO,CAErB;IAED,OAAO,CAAC,QAAQ;YAKF,aAAa;IAY3B;;;OAGG;IACG,KAAK,CAAC,MAAM,CAAC,EAAE,WAAW,GAAG,OAAO,CAAC,IAAI,CAAC,CAQ/C;YAEa,GAAG;YAmCH,iBAAiB;IAqF/B,OAAO,CAAC,WAAW;IAMnB,4DAA4D;IACtD,MAAM,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,SAAK,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC,CAE1D;IAED,wEAAwE;IAClE,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,SAAK,EAAE,IAAI,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,gBAAgB,EAAE,CAAC,CAqBpF;IAED;;;;;OAKG;IACH,kBAAkB,CACjB,GAAG,EAAE,MAAM,EACX,IAAI,EAAE,MAAM,GACV;QAAE,EAAE,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,GAAG,SAAS,CAU9E;YAEa,WAAW;IAYzB,kEAAkE;IAC5D,OAAO,IAAI,OAAO,CAAC,IAAI,CAAC,CAW7B;CACD;AAWD,wBAAgB,wBAAwB,CAAC,GAAG,EAAE,MAAM,EAAE,OAAO,EAAE,gBAAgB,GAAG,IAAI,CAMrF;AAED,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,MAAM,GAAG,gBAAgB,GAAG,SAAS,CAE7E;AAED,wBAAgB,0BAA0B,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,CAE5D","sourcesContent":["/**\n * Orchestrates semantic indexing and search for a repository.\n *\n * Lifecycle (all behind --enable-embsearchtools):\n * 1. `start()` — resolve the embsearch binary, scan the repo (ignore-aware),\n * apply the byte threshold. Under threshold → dormant. Over → spawn the\n * daemon, verify the backend is not the mock embedder, then index changed\n * files in the background in small batches, reporting progress.\n * 2. `search()` — top-k semantic query, mapping chunk ids back to\n * `path:start-end` via the sidecar metadata.\n * 3. `dispose()` — save + close the daemon.\n *\n * Every failure degrades to `unavailable` with a reason; nothing here ever\n * blocks session startup or affects grep/find.\n */\n\nimport { readFileSync } from \"fs\";\nimport { minimatch } from \"minimatch\";\nimport { ensureTool } from \"../../utils/tools-manager.js\";\nimport { chunkFile } from \"./chunker.js\";\nimport { EmbSearchClient } from \"./client.js\";\nimport {\n\temptyIndexMeta,\n\ttype FileMeta,\n\tgetEmbsearchStoreDir,\n\tgetVectorStoreDir,\n\thashContent,\n\ttype IndexMeta,\n\tloadIndexMeta,\n\tsaveIndexMeta,\n} from \"./index-meta.js\";\nimport { type RepoScanFile, scanRepo } from \"./repo-scan.js\";\n\n/** Chunks per bulk request. Small enough that a concurrent query is never\n * stuck long behind one padded batch inference. */\nconst BULK_BATCH_SIZE = 48;\n/** Yield between batches so background indexing doesn't starve the session. */\nconst BATCH_YIELD_MS = 15;\n/** The Rust mock backend's model id — semantically meaningless, never index with it. */\nconst MOCK_MODEL_ID = \"mock-hash-v1\";\n\nexport type EmbsearchState =\n\t| { phase: \"idle\" }\n\t| { phase: \"skipped\"; reason: string }\n\t| { phase: \"downloading\"; receivedBytes: number; totalBytes: number | null }\n\t| { phase: \"indexing\"; done: number; total: number }\n\t| { phase: \"ready\"; chunkCount: number }\n\t| { phase: \"unavailable\"; reason: string };\n\nexport interface EmbsearchServiceOptions {\n\tcwd: string;\n\t/** Explicit binary path (settings override). Default: \"embsearch\" from PATH. */\n\tbinaryPath?: string;\n\t/** Minimum indexable bytes before indexing kicks in. */\n\tthresholdBytes: number;\n\t/** Progress callback for UI (footer / stderr lines). */\n\tonProgress?: (state: EmbsearchState) => void;\n}\n\nexport interface SemanticHit {\n\tpath: string;\n\tstartLine: number;\n\tendLine: number;\n\tscore: number;\n}\n\nexport interface SemanticChunkHit extends SemanticHit {\n\t/** Per-build chunk id (`relpath#index`) — the fusion identity for hybrid search. */\n\tid: string;\n}\n\nexport class EmbsearchService {\n\tprivate readonly options: EmbsearchServiceOptions;\n\tprivate client: EmbSearchClient | undefined;\n\tprivate meta: IndexMeta | undefined;\n\tprivate state: EmbsearchState = { phase: \"idle\" };\n\tprivate disposed = false;\n\n\tconstructor(options: EmbsearchServiceOptions) {\n\t\tthis.options = options;\n\t}\n\n\tgetState(): EmbsearchState {\n\t\treturn this.state;\n\t}\n\n\t/** Semantic search is usable (index ready, or still building with partial data). */\n\tisAvailable(): boolean {\n\t\treturn this.state.phase === \"ready\" || this.state.phase === \"indexing\";\n\t}\n\n\tprivate setState(state: EmbsearchState): void {\n\t\tthis.state = state;\n\t\tthis.options.onProgress?.(state);\n\t}\n\n\tprivate async resolveBinary(): Promise<string | undefined> {\n\t\tif (this.options.binaryPath) {\n\t\t\treturn this.options.binaryPath;\n\t\t}\n\t\t// Surface the on-demand binary download through the same progress channel as\n\t\t// indexing, so the first-run fetch renders a progress bar instead of a stall.\n\t\t// A cached binary resolves without ever invoking this callback.\n\t\treturn await ensureTool(\"embsearch\", true, (receivedBytes, totalBytes) => {\n\t\t\tthis.setState({ phase: \"downloading\", receivedBytes, totalBytes });\n\t\t});\n\t}\n\n\t/**\n\t * Scan, threshold-check, and (when needed) index in the background.\n\t * Resolves when indexing completes or the feature settles dormant.\n\t */\n\tasync start(signal?: AbortSignal): Promise<void> {\n\t\ttry {\n\t\t\tawait this.run(signal);\n\t\t} catch (e) {\n\t\t\tconst reason = e instanceof Error ? e.message : String(e);\n\t\t\tthis.setState({ phase: \"unavailable\", reason });\n\t\t\tawait this.closeClient();\n\t\t}\n\t}\n\n\tprivate async run(signal?: AbortSignal): Promise<void> {\n\t\tconst binary = await this.resolveBinary();\n\t\tif (!binary) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"unavailable\",\n\t\t\t\treason: \"embsearch binary not found (PATH or embsearchBinaryPath setting)\",\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst scan = scanRepo(this.options.cwd, signal);\n\t\tif (scan.totalBytes < this.options.thresholdBytes) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"skipped\",\n\t\t\t\treason: `repo under threshold (${scan.totalBytes} < ${this.options.thresholdBytes} bytes)`,\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst storeDir = getEmbsearchStoreDir(this.options.cwd);\n\t\tthis.client = new EmbSearchClient({ binaryPath: binary, storePath: getVectorStoreDir(storeDir) });\n\t\tawait this.client.ready();\n\n\t\tconst info = await this.client.info();\n\t\tif (info.modelId === MOCK_MODEL_ID) {\n\t\t\tthrow new Error(\"embsearch binary uses the mock embedder (not semantic); install an onnx build\");\n\t\t}\n\n\t\t// Missing/stale sidecar (format, chunker, or model changed) → clean rebuild.\n\t\tthis.meta = loadIndexMeta(storeDir, info.modelId) ?? emptyIndexMeta(this.options.cwd, info.modelId);\n\t\tthis.meta.lastUsedMs = Date.now();\n\n\t\tawait this.indexChangedFiles(scan.files, storeDir, signal);\n\t}\n\n\tprivate async indexChangedFiles(files: RepoScanFile[], storeDir: string, signal?: AbortSignal): Promise<void> {\n\t\tconst meta = this.meta!;\n\t\tconst client = this.client!;\n\n\t\t// Diff scan vs sidecar: cheap mtime+size check first, hash only on delta.\n\t\tconst toIndex: Array<{ file: RepoScanFile; content: string; hash: string }> = [];\n\t\tconst seen = new Set<string>();\n\t\tfor (const file of files) {\n\t\t\tseen.add(file.rel);\n\t\t\tconst known = meta.files[file.rel];\n\t\t\tif (known && known.mtimeMs === file.mtimeMs && known.size === file.size) continue;\n\t\t\tlet content: string;\n\t\t\ttry {\n\t\t\t\tcontent = readFileSync(file.abs, \"utf-8\");\n\t\t\t} catch {\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tconst hash = hashContent(content);\n\t\t\tif (known && known.hash === hash) {\n\t\t\t\t// Touched but unchanged — refresh stat info only.\n\t\t\t\tknown.mtimeMs = file.mtimeMs;\n\t\t\t\tknown.size = file.size;\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\ttoIndex.push({ file, content, hash });\n\t\t}\n\t\tconst toRemove = Object.keys(meta.files).filter((rel) => !seen.has(rel));\n\n\t\tif (toIndex.length === 0 && toRemove.length === 0) {\n\t\t\tsaveIndexMeta(storeDir, meta);\n\t\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t\t\treturn;\n\t\t}\n\n\t\t// Chunk changed files; count total upserts for exact progress.\n\t\tconst work: Array<{ rel: string; fileMeta: FileMeta; chunks: Array<{ id: string; text: string }> }> = [];\n\t\tlet totalChunks = 0;\n\t\tfor (const { file, content, hash } of toIndex) {\n\t\t\tconst chunks = chunkFile(file.rel, content);\n\t\t\twork.push({\n\t\t\t\trel: file.rel,\n\t\t\t\tfileMeta: {\n\t\t\t\t\tmtimeMs: file.mtimeMs,\n\t\t\t\t\tsize: file.size,\n\t\t\t\t\thash,\n\t\t\t\t\tchunks: chunks.map((c) => [c.startLine, c.endLine]),\n\t\t\t\t},\n\t\t\t\tchunks: chunks.map((c) => ({ id: c.id, text: c.text })),\n\t\t\t});\n\t\t\ttotalChunks += chunks.length;\n\t\t}\n\n\t\tthis.setState({ phase: \"indexing\", done: 0, total: totalChunks });\n\n\t\t// Drop vectors of deleted files and superseded chunk tails.\n\t\tfor (const rel of toRemove) {\n\t\t\tfor (let i = 0; i < meta.files[rel].chunks.length; i++) await client.remove(`${rel}#${i}`);\n\t\t\tdelete meta.files[rel];\n\t\t}\n\n\t\tlet done = 0;\n\t\tfor (const item of work) {\n\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\tconst oldChunkCount = meta.files[item.rel]?.chunks.length ?? 0;\n\t\t\t// Remove old chunks beyond the new count (upsert covers the rest).\n\t\t\tfor (let i = item.chunks.length; i < oldChunkCount; i++) await client.remove(`${item.rel}#${i}`);\n\n\t\t\tfor (let offset = 0; offset < item.chunks.length; offset += BULK_BATCH_SIZE) {\n\t\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\t\tconst batch = item.chunks.slice(offset, offset + BULK_BATCH_SIZE);\n\t\t\t\tawait client.bulk(batch);\n\t\t\t\tdone += batch.length;\n\t\t\t\tthis.setState({ phase: \"indexing\", done, total: totalChunks });\n\t\t\t\t// Yield so queries and the event loop stay responsive.\n\t\t\t\tawait new Promise((resolve) => setTimeout(resolve, BATCH_YIELD_MS));\n\t\t\t}\n\t\t\tmeta.files[item.rel] = item.fileMeta;\n\t\t}\n\n\t\tawait client.compact();\n\t\tawait client.save();\n\t\tsaveIndexMeta(storeDir, meta);\n\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t}\n\n\tprivate countChunks(meta: IndexMeta): number {\n\t\tlet n = 0;\n\t\tfor (const rel of Object.keys(meta.files)) n += meta.files[rel].chunks.length;\n\t\treturn n;\n\t}\n\n\t/** Top-`k` semantic hits as `path` + line range + score. */\n\tasync search(query: string, k = 10): Promise<SemanticHit[]> {\n\t\treturn await this.searchChunks(query, k);\n\t}\n\n\t/** Top-`k` semantic hits including their chunk ids, for rank fusion. */\n\tasync searchChunks(query: string, k = 10, glob?: string): Promise<SemanticChunkHit[]> {\n\t\tif (!this.client || this.client.isClosed || !this.meta) {\n\t\t\tthrow new Error(\"semantic index is not available\");\n\t\t}\n\t\tconst results = await this.client.query(query, k);\n\t\tconst hits: SemanticChunkHit[] = [];\n\t\tconst matchGlob = (rel: string): boolean => {\n\t\t\tif (!glob) return true;\n\t\t\treturn minimatch(rel, glob, { dot: true, matchBase: !glob.includes(\"/\") });\n\t\t};\n\t\tfor (const result of results) {\n\t\t\tconst sep = result.id.lastIndexOf(\"#\");\n\t\t\tif (sep === -1) continue;\n\t\t\tconst rel = result.id.slice(0, sep);\n\t\t\tif (!matchGlob(rel)) continue;\n\t\t\tconst chunkIndex = Number.parseInt(result.id.slice(sep + 1), 10);\n\t\t\tconst range = this.meta.files[rel]?.chunks[chunkIndex];\n\t\t\tif (!range) continue;\n\t\t\thits.push({ id: result.id, path: rel, startLine: range[0], endLine: range[1], score: result.score });\n\t\t}\n\t\treturn hits;\n\t}\n\n\t/**\n\t * Resolve a repo-relative path + line to its enclosing indexed chunk, or\n\t * undefined when the file/line is not covered by the index. Chunks overlap\n\t * by a few lines; the first (lowest-index) containing chunk wins so the\n\t * mapping is deterministic.\n\t */\n\tfindEnclosingChunk(\n\t\trel: string,\n\t\tline: number,\n\t): { id: string; path: string; startLine: number; endLine: number } | undefined {\n\t\tconst file = this.meta?.files[rel];\n\t\tif (!file) return undefined;\n\t\tfor (let i = 0; i < file.chunks.length; i++) {\n\t\t\tconst [startLine, endLine] = file.chunks[i];\n\t\t\tif (line >= startLine && line <= endLine) {\n\t\t\t\treturn { id: `${rel}#${i}`, path: rel, startLine, endLine };\n\t\t\t}\n\t\t}\n\t\treturn undefined;\n\t}\n\n\tprivate async closeClient(): Promise<void> {\n\t\tconst client = this.client;\n\t\tthis.client = undefined;\n\t\tif (client && !client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait client.close();\n\t\t\t} catch {\n\t\t\t\t// already dead\n\t\t\t}\n\t\t}\n\t}\n\n\t/** Persist state and shut the daemon down. Safe to call twice. */\n\tasync dispose(): Promise<void> {\n\t\tif (this.disposed) return;\n\t\tthis.disposed = true;\n\t\tif (this.client && !this.client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait this.client.save();\n\t\t\t} catch {\n\t\t\t\t// daemon may have exited; nothing to save\n\t\t\t}\n\t\t}\n\t\tawait this.closeClient();\n\t}\n}\n\n// --- Per-cwd service registry ---\n//\n// The search tool is constructed by the generic tool factory table and\n// only receives `cwd`; the service is created later during session init (flag\n// gated). This registry connects the two without threading a service instance\n// through every layer between main.ts and the tool factories.\n\nconst services = new Map<string, EmbsearchService>();\n\nexport function registerEmbsearchService(cwd: string, service: EmbsearchService): void {\n\tconst old = services.get(cwd);\n\tif (old && old !== service) {\n\t\told.dispose().catch(() => {});\n\t}\n\tservices.set(cwd, service);\n}\n\nexport function getEmbsearchService(cwd: string): EmbsearchService | undefined {\n\treturn services.get(cwd);\n}\n\nexport function unregisterEmbsearchService(cwd: string): void {\n\tservices.delete(cwd);\n}\n"]}
|
|
@@ -51,7 +51,12 @@ export class EmbsearchService {
|
|
|
51
51
|
if (this.options.binaryPath) {
|
|
52
52
|
return this.options.binaryPath;
|
|
53
53
|
}
|
|
54
|
-
|
|
54
|
+
// Surface the on-demand binary download through the same progress channel as
|
|
55
|
+
// indexing, so the first-run fetch renders a progress bar instead of a stall.
|
|
56
|
+
// A cached binary resolves without ever invoking this callback.
|
|
57
|
+
return await ensureTool("embsearch", true, (receivedBytes, totalBytes) => {
|
|
58
|
+
this.setState({ phase: "downloading", receivedBytes, totalBytes });
|
|
59
|
+
});
|
|
55
60
|
}
|
|
56
61
|
/**
|
|
57
62
|
* Scan, threshold-check, and (when needed) index in the background.
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"embsearch-service.js","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,IAAI,CAAC;AAClC,OAAO,EAAE,SAAS,EAAE,MAAM,WAAW,CAAC;AACtC,OAAO,EAAE,UAAU,EAAE,MAAM,8BAA8B,CAAC;AAC1D,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAC;AAC9C,OAAO,EACN,cAAc,EAEd,oBAAoB,EACpB,iBAAiB,EACjB,WAAW,EAEX,aAAa,EACb,aAAa,GACb,MAAM,iBAAiB,CAAC;AACzB,OAAO,EAAqB,QAAQ,EAAE,MAAM,gBAAgB,CAAC;AAE7D;oDACoD;AACpD,MAAM,eAAe,GAAG,EAAE,CAAC;AAC3B,+EAA+E;AAC/E,MAAM,cAAc,GAAG,EAAE,CAAC;AAC1B,0FAAwF;AACxF,MAAM,aAAa,GAAG,cAAc,CAAC;AA+BrC,MAAM,OAAO,gBAAgB;IACX,OAAO,CAA0B;IAC1C,MAAM,CAA8B;IACpC,IAAI,CAAwB;IAC5B,KAAK,GAAmB,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;IAC1C,QAAQ,GAAG,KAAK,CAAC;IAEzB,YAAY,OAAgC,EAAE;QAC7C,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC;IAAA,CACvB;IAED,QAAQ,GAAmB;QAC1B,OAAO,IAAI,CAAC,KAAK,CAAC;IAAA,CAClB;IAED,oFAAoF;IACpF,WAAW,GAAY;QACtB,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,OAAO,IAAI,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,UAAU,CAAC;IAAA,CACvE;IAEO,QAAQ,CAAC,KAAqB,EAAQ;QAC7C,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;QACnB,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC,KAAK,CAAC,CAAC;IAAA,CACjC;IAEO,KAAK,CAAC,aAAa,GAAgC;QAC1D,IAAI,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YAC7B,OAAO,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC;QAChC,CAAC;QACD,OAAO,MAAM,UAAU,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC;IAAA,CAC3C;IAED;;;OAGG;IACH,KAAK,CAAC,KAAK,CAAC,MAAoB,EAAiB;QAChD,IAAI,CAAC;YACJ,MAAM,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;QACxB,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,MAAM,GAAG,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;YAC1D,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,aAAa,EAAE,MAAM,EAAE,CAAC,CAAC;YAChD,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC;QAC1B,CAAC;IAAA,CACD;IAEO,KAAK,CAAC,GAAG,CAAC,MAAoB,EAAiB;QACtD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,aAAa,EAAE,CAAC;QAC1C,IAAI,CAAC,MAAM,EAAE,CAAC;YACb,IAAI,CAAC,QAAQ,CAAC;gBACb,KAAK,EAAE,aAAa;gBACpB,MAAM,EAAE,kEAAkE;aAC1E,CAAC,CAAC;YACH,OAAO;QACR,CAAC;QAED,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;QAChD,IAAI,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,cAAc,EAAE,CAAC;YACnD,IAAI,CAAC,QAAQ,CAAC;gBACb,KAAK,EAAE,SAAS;gBAChB,MAAM,EAAE,yBAAyB,IAAI,CAAC,UAAU,MAAM,IAAI,CAAC,OAAO,CAAC,cAAc,SAAS;aAC1F,CAAC,CAAC;YACH,OAAO;QACR,CAAC;QAED,MAAM,QAAQ,GAAG,oBAAoB,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QACxD,IAAI,CAAC,MAAM,GAAG,IAAI,eAAe,CAAC,EAAE,UAAU,EAAE,MAAM,EAAE,SAAS,EAAE,iBAAiB,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;QAClG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;QAE1B,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;QACtC,IAAI,IAAI,CAAC,OAAO,KAAK,aAAa,EAAE,CAAC;YACpC,MAAM,IAAI,KAAK,CAAC,+EAA+E,CAAC,CAAC;QAClG,CAAC;QAED,+EAA6E;QAC7E,IAAI,CAAC,IAAI,GAAG,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,OAAO,CAAC,IAAI,cAAc,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;QACpG,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAElC,MAAM,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC;IAAA,CAC3D;IAEO,KAAK,CAAC,iBAAiB,CAAC,KAAqB,EAAE,QAAgB,EAAE,MAAoB,EAAiB;QAC7G,MAAM,IAAI,GAAG,IAAI,CAAC,IAAK,CAAC;QACxB,MAAM,MAAM,GAAG,IAAI,CAAC,MAAO,CAAC;QAE5B,0EAA0E;QAC1E,MAAM,OAAO,GAAiE,EAAE,CAAC;QACjF,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YAC1B,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnC,IAAI,KAAK,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI;gBAAE,SAAS;YAClF,IAAI,OAAe,CAAC;YACpB,IAAI,CAAC;gBACJ,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YAC3C,CAAC;YAAC,MAAM,CAAC;gBACR,SAAS;YACV,CAAC;YACD,MAAM,IAAI,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC;YAClC,IAAI,KAAK,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;gBAClC,oDAAkD;gBAClD,KAAK,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC;gBAC7B,KAAK,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;gBACvB,SAAS;YACV,CAAC;YACD,OAAO,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;QACvC,CAAC;QACD,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;QAEzE,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACnD,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;YAC9B,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACtE,OAAO;QACR,CAAC;QAED,+DAA+D;QAC/D,MAAM,IAAI,GAA4F,EAAE,CAAC;QACzG,IAAI,WAAW,GAAG,CAAC,CAAC;QACpB,KAAK,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,OAAO,EAAE,CAAC;YAC/C,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YAC5C,IAAI,CAAC,IAAI,CAAC;gBACT,GAAG,EAAE,IAAI,CAAC,GAAG;gBACb,QAAQ,EAAE;oBACT,OAAO,EAAE,IAAI,CAAC,OAAO;oBACrB,IAAI,EAAE,IAAI,CAAC,IAAI;oBACf,IAAI;oBACJ,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC;iBACnD;gBACD,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;aACvD,CAAC,CAAC;YACH,WAAW,IAAI,MAAM,CAAC,MAAM,CAAC;QAC9B,CAAC;QAED,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC,CAAC;QAElE,4DAA4D;QAC5D,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;YAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE;gBAAE,MAAM,MAAM,CAAC,MAAM,CAAC,GAAG,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC;YAC3F,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACxB,CAAC;QAED,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;YACzB,IAAI,MAAM,EAAE,OAAO,IAAI,IAAI,CAAC,QAAQ;gBAAE,OAAO;YAC7C,MAAM,aAAa,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,MAAM,IAAI,CAAC,CAAC;YAC/D,mEAAmE;YACnE,KAAK,IAAI,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,GAAG,aAAa,EAAE,CAAC,EAAE;gBAAE,MAAM,MAAM,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC;YAEjG,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,IAAI,eAAe,EAAE,CAAC;gBAC7E,IAAI,MAAM,EAAE,OAAO,IAAI,IAAI,CAAC,QAAQ;oBAAE,OAAO;gBAC7C,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,eAAe,CAAC,CAAC;gBAClE,MAAM,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gBACzB,IAAI,IAAI,KAAK,CAAC,MAAM,CAAC;gBACrB,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC,CAAC;gBAC/D,uDAAuD;gBACvD,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,cAAc,CAAC,CAAC,CAAC;YACrE,CAAC;YACD,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC;QACtC,CAAC;QAED,MAAM,MAAM,CAAC,OAAO,EAAE,CAAC;QACvB,MAAM,MAAM,CAAC,IAAI,EAAE,CAAC;QACpB,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QAC9B,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAAA,CACtE;IAEO,WAAW,CAAC,IAAe,EAAU;QAC5C,IAAI,CAAC,GAAG,CAAC,CAAC;QACV,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC;YAAE,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC;QAC9E,OAAO,CAAC,CAAC;IAAA,CACT;IAED,4DAA4D;IAC5D,KAAK,CAAC,MAAM,CAAC,KAAa,EAAE,CAAC,GAAG,EAAE,EAA0B;QAC3D,OAAO,MAAM,IAAI,CAAC,YAAY,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;IAAA,CACzC;IAED,wEAAwE;IACxE,KAAK,CAAC,YAAY,CAAC,KAAa,EAAE,CAAC,GAAG,EAAE,EAAE,IAAa,EAA+B;QACrF,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;YACxD,MAAM,IAAI,KAAK,CAAC,iCAAiC,CAAC,CAAC;QACpD,CAAC;QACD,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;QAClD,MAAM,IAAI,GAAuB,EAAE,CAAC;QACpC,MAAM,SAAS,GAAG,CAAC,GAAW,EAAW,EAAE,CAAC;YAC3C,IAAI,CAAC,IAAI;gBAAE,OAAO,IAAI,CAAC;YACvB,OAAO,SAAS,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAAA,CAC3E,CAAC;QACF,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC9B,MAAM,GAAG,GAAG,MAAM,CAAC,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;YACvC,IAAI,GAAG,KAAK,CAAC,CAAC;gBAAE,SAAS;YACzB,MAAM,GAAG,GAAG,MAAM,CAAC,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;YACpC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC;gBAAE,SAAS;YAC9B,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YACjE,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,UAAU,CAAC,CAAC;YACvD,IAAI,CAAC,KAAK;gBAAE,SAAS;YACrB,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,CAAC,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,SAAS,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC;QACtG,CAAC;QACD,OAAO,IAAI,CAAC;IAAA,CACZ;IAED;;;;;OAKG;IACH,kBAAkB,CACjB,GAAW,EACX,IAAY,EACmE;QAC/E,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;QACnC,IAAI,CAAC,IAAI;YAAE,OAAO,SAAS,CAAC;QAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7C,MAAM,CAAC,SAAS,EAAE,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;YAC5C,IAAI,IAAI,IAAI,SAAS,IAAI,IAAI,IAAI,OAAO,EAAE,CAAC;gBAC1C,OAAO,EAAE,EAAE,EAAE,GAAG,GAAG,IAAI,CAAC,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC;YAC7D,CAAC;QACF,CAAC;QACD,OAAO,SAAS,CAAC;IAAA,CACjB;IAEO,KAAK,CAAC,WAAW,GAAkB;QAC1C,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;QAC3B,IAAI,CAAC,MAAM,GAAG,SAAS,CAAC;QACxB,IAAI,MAAM,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YAChC,IAAI,CAAC;gBACJ,MAAM,MAAM,CAAC,KAAK,EAAE,CAAC;YACtB,CAAC;YAAC,MAAM,CAAC;gBACR,eAAe;YAChB,CAAC;QACF,CAAC;IAAA,CACD;IAED,kEAAkE;IAClE,KAAK,CAAC,OAAO,GAAkB;QAC9B,IAAI,IAAI,CAAC,QAAQ;YAAE,OAAO;QAC1B,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC;QACrB,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YAC1C,IAAI,CAAC;gBACJ,MAAM,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;YAC1B,CAAC;YAAC,MAAM,CAAC;gBACR,0CAA0C;YAC3C,CAAC;QACF,CAAC;QACD,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC;IAAA,CACzB;CACD;AAED,mCAAmC;AACnC,EAAE;AACF,uEAAuE;AACvE,8EAA8E;AAC9E,8EAA8E;AAC9E,8DAA8D;AAE9D,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA4B,CAAC;AAErD,MAAM,UAAU,wBAAwB,CAAC,GAAW,EAAE,OAAyB,EAAQ;IACtF,MAAM,GAAG,GAAG,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAC9B,IAAI,GAAG,IAAI,GAAG,KAAK,OAAO,EAAE,CAAC;QAC5B,GAAG,CAAC,OAAO,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAC,CAAC,CAAC,CAAC;IAC/B,CAAC;IACD,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;AAAA,CAC3B;AAED,MAAM,UAAU,mBAAmB,CAAC,GAAW,EAAgC;IAC9E,OAAO,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;AAAA,CACzB;AAED,MAAM,UAAU,0BAA0B,CAAC,GAAW,EAAQ;IAC7D,QAAQ,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;AAAA,CACrB","sourcesContent":["/**\n * Orchestrates semantic indexing and search for a repository.\n *\n * Lifecycle (all behind --enable-embsearchtools):\n * 1. `start()` — resolve the embsearch binary, scan the repo (ignore-aware),\n * apply the byte threshold. Under threshold → dormant. Over → spawn the\n * daemon, verify the backend is not the mock embedder, then index changed\n * files in the background in small batches, reporting progress.\n * 2. `search()` — top-k semantic query, mapping chunk ids back to\n * `path:start-end` via the sidecar metadata.\n * 3. `dispose()` — save + close the daemon.\n *\n * Every failure degrades to `unavailable` with a reason; nothing here ever\n * blocks session startup or affects grep/find.\n */\n\nimport { readFileSync } from \"fs\";\nimport { minimatch } from \"minimatch\";\nimport { ensureTool } from \"../../utils/tools-manager.js\";\nimport { chunkFile } from \"./chunker.js\";\nimport { EmbSearchClient } from \"./client.js\";\nimport {\n\temptyIndexMeta,\n\ttype FileMeta,\n\tgetEmbsearchStoreDir,\n\tgetVectorStoreDir,\n\thashContent,\n\ttype IndexMeta,\n\tloadIndexMeta,\n\tsaveIndexMeta,\n} from \"./index-meta.js\";\nimport { type RepoScanFile, scanRepo } from \"./repo-scan.js\";\n\n/** Chunks per bulk request. Small enough that a concurrent query is never\n * stuck long behind one padded batch inference. */\nconst BULK_BATCH_SIZE = 48;\n/** Yield between batches so background indexing doesn't starve the session. */\nconst BATCH_YIELD_MS = 15;\n/** The Rust mock backend's model id — semantically meaningless, never index with it. */\nconst MOCK_MODEL_ID = \"mock-hash-v1\";\n\nexport type EmbsearchState =\n\t| { phase: \"idle\" }\n\t| { phase: \"skipped\"; reason: string }\n\t| { phase: \"indexing\"; done: number; total: number }\n\t| { phase: \"ready\"; chunkCount: number }\n\t| { phase: \"unavailable\"; reason: string };\n\nexport interface EmbsearchServiceOptions {\n\tcwd: string;\n\t/** Explicit binary path (settings override). Default: \"embsearch\" from PATH. */\n\tbinaryPath?: string;\n\t/** Minimum indexable bytes before indexing kicks in. */\n\tthresholdBytes: number;\n\t/** Progress callback for UI (footer / stderr lines). */\n\tonProgress?: (state: EmbsearchState) => void;\n}\n\nexport interface SemanticHit {\n\tpath: string;\n\tstartLine: number;\n\tendLine: number;\n\tscore: number;\n}\n\nexport interface SemanticChunkHit extends SemanticHit {\n\t/** Per-build chunk id (`relpath#index`) — the fusion identity for hybrid search. */\n\tid: string;\n}\n\nexport class EmbsearchService {\n\tprivate readonly options: EmbsearchServiceOptions;\n\tprivate client: EmbSearchClient | undefined;\n\tprivate meta: IndexMeta | undefined;\n\tprivate state: EmbsearchState = { phase: \"idle\" };\n\tprivate disposed = false;\n\n\tconstructor(options: EmbsearchServiceOptions) {\n\t\tthis.options = options;\n\t}\n\n\tgetState(): EmbsearchState {\n\t\treturn this.state;\n\t}\n\n\t/** Semantic search is usable (index ready, or still building with partial data). */\n\tisAvailable(): boolean {\n\t\treturn this.state.phase === \"ready\" || this.state.phase === \"indexing\";\n\t}\n\n\tprivate setState(state: EmbsearchState): void {\n\t\tthis.state = state;\n\t\tthis.options.onProgress?.(state);\n\t}\n\n\tprivate async resolveBinary(): Promise<string | undefined> {\n\t\tif (this.options.binaryPath) {\n\t\t\treturn this.options.binaryPath;\n\t\t}\n\t\treturn await ensureTool(\"embsearch\", true);\n\t}\n\n\t/**\n\t * Scan, threshold-check, and (when needed) index in the background.\n\t * Resolves when indexing completes or the feature settles dormant.\n\t */\n\tasync start(signal?: AbortSignal): Promise<void> {\n\t\ttry {\n\t\t\tawait this.run(signal);\n\t\t} catch (e) {\n\t\t\tconst reason = e instanceof Error ? e.message : String(e);\n\t\t\tthis.setState({ phase: \"unavailable\", reason });\n\t\t\tawait this.closeClient();\n\t\t}\n\t}\n\n\tprivate async run(signal?: AbortSignal): Promise<void> {\n\t\tconst binary = await this.resolveBinary();\n\t\tif (!binary) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"unavailable\",\n\t\t\t\treason: \"embsearch binary not found (PATH or embsearchBinaryPath setting)\",\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst scan = scanRepo(this.options.cwd, signal);\n\t\tif (scan.totalBytes < this.options.thresholdBytes) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"skipped\",\n\t\t\t\treason: `repo under threshold (${scan.totalBytes} < ${this.options.thresholdBytes} bytes)`,\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst storeDir = getEmbsearchStoreDir(this.options.cwd);\n\t\tthis.client = new EmbSearchClient({ binaryPath: binary, storePath: getVectorStoreDir(storeDir) });\n\t\tawait this.client.ready();\n\n\t\tconst info = await this.client.info();\n\t\tif (info.modelId === MOCK_MODEL_ID) {\n\t\t\tthrow new Error(\"embsearch binary uses the mock embedder (not semantic); install an onnx build\");\n\t\t}\n\n\t\t// Missing/stale sidecar (format, chunker, or model changed) → clean rebuild.\n\t\tthis.meta = loadIndexMeta(storeDir, info.modelId) ?? emptyIndexMeta(this.options.cwd, info.modelId);\n\t\tthis.meta.lastUsedMs = Date.now();\n\n\t\tawait this.indexChangedFiles(scan.files, storeDir, signal);\n\t}\n\n\tprivate async indexChangedFiles(files: RepoScanFile[], storeDir: string, signal?: AbortSignal): Promise<void> {\n\t\tconst meta = this.meta!;\n\t\tconst client = this.client!;\n\n\t\t// Diff scan vs sidecar: cheap mtime+size check first, hash only on delta.\n\t\tconst toIndex: Array<{ file: RepoScanFile; content: string; hash: string }> = [];\n\t\tconst seen = new Set<string>();\n\t\tfor (const file of files) {\n\t\t\tseen.add(file.rel);\n\t\t\tconst known = meta.files[file.rel];\n\t\t\tif (known && known.mtimeMs === file.mtimeMs && known.size === file.size) continue;\n\t\t\tlet content: string;\n\t\t\ttry {\n\t\t\t\tcontent = readFileSync(file.abs, \"utf-8\");\n\t\t\t} catch {\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tconst hash = hashContent(content);\n\t\t\tif (known && known.hash === hash) {\n\t\t\t\t// Touched but unchanged — refresh stat info only.\n\t\t\t\tknown.mtimeMs = file.mtimeMs;\n\t\t\t\tknown.size = file.size;\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\ttoIndex.push({ file, content, hash });\n\t\t}\n\t\tconst toRemove = Object.keys(meta.files).filter((rel) => !seen.has(rel));\n\n\t\tif (toIndex.length === 0 && toRemove.length === 0) {\n\t\t\tsaveIndexMeta(storeDir, meta);\n\t\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t\t\treturn;\n\t\t}\n\n\t\t// Chunk changed files; count total upserts for exact progress.\n\t\tconst work: Array<{ rel: string; fileMeta: FileMeta; chunks: Array<{ id: string; text: string }> }> = [];\n\t\tlet totalChunks = 0;\n\t\tfor (const { file, content, hash } of toIndex) {\n\t\t\tconst chunks = chunkFile(file.rel, content);\n\t\t\twork.push({\n\t\t\t\trel: file.rel,\n\t\t\t\tfileMeta: {\n\t\t\t\t\tmtimeMs: file.mtimeMs,\n\t\t\t\t\tsize: file.size,\n\t\t\t\t\thash,\n\t\t\t\t\tchunks: chunks.map((c) => [c.startLine, c.endLine]),\n\t\t\t\t},\n\t\t\t\tchunks: chunks.map((c) => ({ id: c.id, text: c.text })),\n\t\t\t});\n\t\t\ttotalChunks += chunks.length;\n\t\t}\n\n\t\tthis.setState({ phase: \"indexing\", done: 0, total: totalChunks });\n\n\t\t// Drop vectors of deleted files and superseded chunk tails.\n\t\tfor (const rel of toRemove) {\n\t\t\tfor (let i = 0; i < meta.files[rel].chunks.length; i++) await client.remove(`${rel}#${i}`);\n\t\t\tdelete meta.files[rel];\n\t\t}\n\n\t\tlet done = 0;\n\t\tfor (const item of work) {\n\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\tconst oldChunkCount = meta.files[item.rel]?.chunks.length ?? 0;\n\t\t\t// Remove old chunks beyond the new count (upsert covers the rest).\n\t\t\tfor (let i = item.chunks.length; i < oldChunkCount; i++) await client.remove(`${item.rel}#${i}`);\n\n\t\t\tfor (let offset = 0; offset < item.chunks.length; offset += BULK_BATCH_SIZE) {\n\t\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\t\tconst batch = item.chunks.slice(offset, offset + BULK_BATCH_SIZE);\n\t\t\t\tawait client.bulk(batch);\n\t\t\t\tdone += batch.length;\n\t\t\t\tthis.setState({ phase: \"indexing\", done, total: totalChunks });\n\t\t\t\t// Yield so queries and the event loop stay responsive.\n\t\t\t\tawait new Promise((resolve) => setTimeout(resolve, BATCH_YIELD_MS));\n\t\t\t}\n\t\t\tmeta.files[item.rel] = item.fileMeta;\n\t\t}\n\n\t\tawait client.compact();\n\t\tawait client.save();\n\t\tsaveIndexMeta(storeDir, meta);\n\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t}\n\n\tprivate countChunks(meta: IndexMeta): number {\n\t\tlet n = 0;\n\t\tfor (const rel of Object.keys(meta.files)) n += meta.files[rel].chunks.length;\n\t\treturn n;\n\t}\n\n\t/** Top-`k` semantic hits as `path` + line range + score. */\n\tasync search(query: string, k = 10): Promise<SemanticHit[]> {\n\t\treturn await this.searchChunks(query, k);\n\t}\n\n\t/** Top-`k` semantic hits including their chunk ids, for rank fusion. */\n\tasync searchChunks(query: string, k = 10, glob?: string): Promise<SemanticChunkHit[]> {\n\t\tif (!this.client || this.client.isClosed || !this.meta) {\n\t\t\tthrow new Error(\"semantic index is not available\");\n\t\t}\n\t\tconst results = await this.client.query(query, k);\n\t\tconst hits: SemanticChunkHit[] = [];\n\t\tconst matchGlob = (rel: string): boolean => {\n\t\t\tif (!glob) return true;\n\t\t\treturn minimatch(rel, glob, { dot: true, matchBase: !glob.includes(\"/\") });\n\t\t};\n\t\tfor (const result of results) {\n\t\t\tconst sep = result.id.lastIndexOf(\"#\");\n\t\t\tif (sep === -1) continue;\n\t\t\tconst rel = result.id.slice(0, sep);\n\t\t\tif (!matchGlob(rel)) continue;\n\t\t\tconst chunkIndex = Number.parseInt(result.id.slice(sep + 1), 10);\n\t\t\tconst range = this.meta.files[rel]?.chunks[chunkIndex];\n\t\t\tif (!range) continue;\n\t\t\thits.push({ id: result.id, path: rel, startLine: range[0], endLine: range[1], score: result.score });\n\t\t}\n\t\treturn hits;\n\t}\n\n\t/**\n\t * Resolve a repo-relative path + line to its enclosing indexed chunk, or\n\t * undefined when the file/line is not covered by the index. Chunks overlap\n\t * by a few lines; the first (lowest-index) containing chunk wins so the\n\t * mapping is deterministic.\n\t */\n\tfindEnclosingChunk(\n\t\trel: string,\n\t\tline: number,\n\t): { id: string; path: string; startLine: number; endLine: number } | undefined {\n\t\tconst file = this.meta?.files[rel];\n\t\tif (!file) return undefined;\n\t\tfor (let i = 0; i < file.chunks.length; i++) {\n\t\t\tconst [startLine, endLine] = file.chunks[i];\n\t\t\tif (line >= startLine && line <= endLine) {\n\t\t\t\treturn { id: `${rel}#${i}`, path: rel, startLine, endLine };\n\t\t\t}\n\t\t}\n\t\treturn undefined;\n\t}\n\n\tprivate async closeClient(): Promise<void> {\n\t\tconst client = this.client;\n\t\tthis.client = undefined;\n\t\tif (client && !client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait client.close();\n\t\t\t} catch {\n\t\t\t\t// already dead\n\t\t\t}\n\t\t}\n\t}\n\n\t/** Persist state and shut the daemon down. Safe to call twice. */\n\tasync dispose(): Promise<void> {\n\t\tif (this.disposed) return;\n\t\tthis.disposed = true;\n\t\tif (this.client && !this.client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait this.client.save();\n\t\t\t} catch {\n\t\t\t\t// daemon may have exited; nothing to save\n\t\t\t}\n\t\t}\n\t\tawait this.closeClient();\n\t}\n}\n\n// --- Per-cwd service registry ---\n//\n// The search tool is constructed by the generic tool factory table and\n// only receives `cwd`; the service is created later during session init (flag\n// gated). This registry connects the two without threading a service instance\n// through every layer between main.ts and the tool factories.\n\nconst services = new Map<string, EmbsearchService>();\n\nexport function registerEmbsearchService(cwd: string, service: EmbsearchService): void {\n\tconst old = services.get(cwd);\n\tif (old && old !== service) {\n\t\told.dispose().catch(() => {});\n\t}\n\tservices.set(cwd, service);\n}\n\nexport function getEmbsearchService(cwd: string): EmbsearchService | undefined {\n\treturn services.get(cwd);\n}\n\nexport function unregisterEmbsearchService(cwd: string): void {\n\tservices.delete(cwd);\n}\n"]}
|
|
1
|
+
{"version":3,"file":"embsearch-service.js","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,IAAI,CAAC;AAClC,OAAO,EAAE,SAAS,EAAE,MAAM,WAAW,CAAC;AACtC,OAAO,EAAE,UAAU,EAAE,MAAM,8BAA8B,CAAC;AAC1D,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAC;AAC9C,OAAO,EACN,cAAc,EAEd,oBAAoB,EACpB,iBAAiB,EACjB,WAAW,EAEX,aAAa,EACb,aAAa,GACb,MAAM,iBAAiB,CAAC;AACzB,OAAO,EAAqB,QAAQ,EAAE,MAAM,gBAAgB,CAAC;AAE7D;oDACoD;AACpD,MAAM,eAAe,GAAG,EAAE,CAAC;AAC3B,+EAA+E;AAC/E,MAAM,cAAc,GAAG,EAAE,CAAC;AAC1B,0FAAwF;AACxF,MAAM,aAAa,GAAG,cAAc,CAAC;AAgCrC,MAAM,OAAO,gBAAgB;IACX,OAAO,CAA0B;IAC1C,MAAM,CAA8B;IACpC,IAAI,CAAwB;IAC5B,KAAK,GAAmB,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;IAC1C,QAAQ,GAAG,KAAK,CAAC;IAEzB,YAAY,OAAgC,EAAE;QAC7C,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC;IAAA,CACvB;IAED,QAAQ,GAAmB;QAC1B,OAAO,IAAI,CAAC,KAAK,CAAC;IAAA,CAClB;IAED,oFAAoF;IACpF,WAAW,GAAY;QACtB,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,OAAO,IAAI,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,UAAU,CAAC;IAAA,CACvE;IAEO,QAAQ,CAAC,KAAqB,EAAQ;QAC7C,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;QACnB,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC,KAAK,CAAC,CAAC;IAAA,CACjC;IAEO,KAAK,CAAC,aAAa,GAAgC;QAC1D,IAAI,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YAC7B,OAAO,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC;QAChC,CAAC;QACD,6EAA6E;QAC7E,8EAA8E;QAC9E,gEAAgE;QAChE,OAAO,MAAM,UAAU,CAAC,WAAW,EAAE,IAAI,EAAE,CAAC,aAAa,EAAE,UAAU,EAAE,EAAE,CAAC;YACzE,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,aAAa,EAAE,aAAa,EAAE,UAAU,EAAE,CAAC,CAAC;QAAA,CACnE,CAAC,CAAC;IAAA,CACH;IAED;;;OAGG;IACH,KAAK,CAAC,KAAK,CAAC,MAAoB,EAAiB;QAChD,IAAI,CAAC;YACJ,MAAM,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;QACxB,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,MAAM,GAAG,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;YAC1D,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,aAAa,EAAE,MAAM,EAAE,CAAC,CAAC;YAChD,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC;QAC1B,CAAC;IAAA,CACD;IAEO,KAAK,CAAC,GAAG,CAAC,MAAoB,EAAiB;QACtD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,aAAa,EAAE,CAAC;QAC1C,IAAI,CAAC,MAAM,EAAE,CAAC;YACb,IAAI,CAAC,QAAQ,CAAC;gBACb,KAAK,EAAE,aAAa;gBACpB,MAAM,EAAE,kEAAkE;aAC1E,CAAC,CAAC;YACH,OAAO;QACR,CAAC;QAED,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;QAChD,IAAI,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,cAAc,EAAE,CAAC;YACnD,IAAI,CAAC,QAAQ,CAAC;gBACb,KAAK,EAAE,SAAS;gBAChB,MAAM,EAAE,yBAAyB,IAAI,CAAC,UAAU,MAAM,IAAI,CAAC,OAAO,CAAC,cAAc,SAAS;aAC1F,CAAC,CAAC;YACH,OAAO;QACR,CAAC;QAED,MAAM,QAAQ,GAAG,oBAAoB,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QACxD,IAAI,CAAC,MAAM,GAAG,IAAI,eAAe,CAAC,EAAE,UAAU,EAAE,MAAM,EAAE,SAAS,EAAE,iBAAiB,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;QAClG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;QAE1B,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;QACtC,IAAI,IAAI,CAAC,OAAO,KAAK,aAAa,EAAE,CAAC;YACpC,MAAM,IAAI,KAAK,CAAC,+EAA+E,CAAC,CAAC;QAClG,CAAC;QAED,+EAA6E;QAC7E,IAAI,CAAC,IAAI,GAAG,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,OAAO,CAAC,IAAI,cAAc,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;QACpG,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAElC,MAAM,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC;IAAA,CAC3D;IAEO,KAAK,CAAC,iBAAiB,CAAC,KAAqB,EAAE,QAAgB,EAAE,MAAoB,EAAiB;QAC7G,MAAM,IAAI,GAAG,IAAI,CAAC,IAAK,CAAC;QACxB,MAAM,MAAM,GAAG,IAAI,CAAC,MAAO,CAAC;QAE5B,0EAA0E;QAC1E,MAAM,OAAO,GAAiE,EAAE,CAAC;QACjF,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YAC1B,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnC,IAAI,KAAK,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI;gBAAE,SAAS;YAClF,IAAI,OAAe,CAAC;YACpB,IAAI,CAAC;gBACJ,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YAC3C,CAAC;YAAC,MAAM,CAAC;gBACR,SAAS;YACV,CAAC;YACD,MAAM,IAAI,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC;YAClC,IAAI,KAAK,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;gBAClC,oDAAkD;gBAClD,KAAK,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC;gBAC7B,KAAK,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;gBACvB,SAAS;YACV,CAAC;YACD,OAAO,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;QACvC,CAAC;QACD,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;QAEzE,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACnD,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;YAC9B,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACtE,OAAO;QACR,CAAC;QAED,+DAA+D;QAC/D,MAAM,IAAI,GAA4F,EAAE,CAAC;QACzG,IAAI,WAAW,GAAG,CAAC,CAAC;QACpB,KAAK,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,OAAO,EAAE,CAAC;YAC/C,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YAC5C,IAAI,CAAC,IAAI,CAAC;gBACT,GAAG,EAAE,IAAI,CAAC,GAAG;gBACb,QAAQ,EAAE;oBACT,OAAO,EAAE,IAAI,CAAC,OAAO;oBACrB,IAAI,EAAE,IAAI,CAAC,IAAI;oBACf,IAAI;oBACJ,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC;iBACnD;gBACD,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;aACvD,CAAC,CAAC;YACH,WAAW,IAAI,MAAM,CAAC,MAAM,CAAC;QAC9B,CAAC;QAED,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC,CAAC;QAElE,4DAA4D;QAC5D,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;YAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE;gBAAE,MAAM,MAAM,CAAC,MAAM,CAAC,GAAG,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC;YAC3F,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACxB,CAAC;QAED,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;YACzB,IAAI,MAAM,EAAE,OAAO,IAAI,IAAI,CAAC,QAAQ;gBAAE,OAAO;YAC7C,MAAM,aAAa,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,MAAM,IAAI,CAAC,CAAC;YAC/D,mEAAmE;YACnE,KAAK,IAAI,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,GAAG,aAAa,EAAE,CAAC,EAAE;gBAAE,MAAM,MAAM,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC;YAEjG,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,IAAI,eAAe,EAAE,CAAC;gBAC7E,IAAI,MAAM,EAAE,OAAO,IAAI,IAAI,CAAC,QAAQ;oBAAE,OAAO;gBAC7C,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,eAAe,CAAC,CAAC;gBAClE,MAAM,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gBACzB,IAAI,IAAI,KAAK,CAAC,MAAM,CAAC;gBACrB,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC,CAAC;gBAC/D,uDAAuD;gBACvD,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,cAAc,CAAC,CAAC,CAAC;YACrE,CAAC;YACD,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC;QACtC,CAAC;QAED,MAAM,MAAM,CAAC,OAAO,EAAE,CAAC;QACvB,MAAM,MAAM,CAAC,IAAI,EAAE,CAAC;QACpB,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QAC9B,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAAA,CACtE;IAEO,WAAW,CAAC,IAAe,EAAU;QAC5C,IAAI,CAAC,GAAG,CAAC,CAAC;QACV,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC;YAAE,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC;QAC9E,OAAO,CAAC,CAAC;IAAA,CACT;IAED,4DAA4D;IAC5D,KAAK,CAAC,MAAM,CAAC,KAAa,EAAE,CAAC,GAAG,EAAE,EAA0B;QAC3D,OAAO,MAAM,IAAI,CAAC,YAAY,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;IAAA,CACzC;IAED,wEAAwE;IACxE,KAAK,CAAC,YAAY,CAAC,KAAa,EAAE,CAAC,GAAG,EAAE,EAAE,IAAa,EAA+B;QACrF,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;YACxD,MAAM,IAAI,KAAK,CAAC,iCAAiC,CAAC,CAAC;QACpD,CAAC;QACD,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;QAClD,MAAM,IAAI,GAAuB,EAAE,CAAC;QACpC,MAAM,SAAS,GAAG,CAAC,GAAW,EAAW,EAAE,CAAC;YAC3C,IAAI,CAAC,IAAI;gBAAE,OAAO,IAAI,CAAC;YACvB,OAAO,SAAS,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAAA,CAC3E,CAAC;QACF,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC9B,MAAM,GAAG,GAAG,MAAM,CAAC,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;YACvC,IAAI,GAAG,KAAK,CAAC,CAAC;gBAAE,SAAS;YACzB,MAAM,GAAG,GAAG,MAAM,CAAC,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;YACpC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC;gBAAE,SAAS;YAC9B,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YACjE,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,UAAU,CAAC,CAAC;YACvD,IAAI,CAAC,KAAK;gBAAE,SAAS;YACrB,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,CAAC,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,SAAS,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC;QACtG,CAAC;QACD,OAAO,IAAI,CAAC;IAAA,CACZ;IAED;;;;;OAKG;IACH,kBAAkB,CACjB,GAAW,EACX,IAAY,EACmE;QAC/E,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;QACnC,IAAI,CAAC,IAAI;YAAE,OAAO,SAAS,CAAC;QAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7C,MAAM,CAAC,SAAS,EAAE,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;YAC5C,IAAI,IAAI,IAAI,SAAS,IAAI,IAAI,IAAI,OAAO,EAAE,CAAC;gBAC1C,OAAO,EAAE,EAAE,EAAE,GAAG,GAAG,IAAI,CAAC,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC;YAC7D,CAAC;QACF,CAAC;QACD,OAAO,SAAS,CAAC;IAAA,CACjB;IAEO,KAAK,CAAC,WAAW,GAAkB;QAC1C,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;QAC3B,IAAI,CAAC,MAAM,GAAG,SAAS,CAAC;QACxB,IAAI,MAAM,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YAChC,IAAI,CAAC;gBACJ,MAAM,MAAM,CAAC,KAAK,EAAE,CAAC;YACtB,CAAC;YAAC,MAAM,CAAC;gBACR,eAAe;YAChB,CAAC;QACF,CAAC;IAAA,CACD;IAED,kEAAkE;IAClE,KAAK,CAAC,OAAO,GAAkB;QAC9B,IAAI,IAAI,CAAC,QAAQ;YAAE,OAAO;QAC1B,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC;QACrB,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YAC1C,IAAI,CAAC;gBACJ,MAAM,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;YAC1B,CAAC;YAAC,MAAM,CAAC;gBACR,0CAA0C;YAC3C,CAAC;QACF,CAAC;QACD,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC;IAAA,CACzB;CACD;AAED,mCAAmC;AACnC,EAAE;AACF,uEAAuE;AACvE,8EAA8E;AAC9E,8EAA8E;AAC9E,8DAA8D;AAE9D,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA4B,CAAC;AAErD,MAAM,UAAU,wBAAwB,CAAC,GAAW,EAAE,OAAyB,EAAQ;IACtF,MAAM,GAAG,GAAG,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAC9B,IAAI,GAAG,IAAI,GAAG,KAAK,OAAO,EAAE,CAAC;QAC5B,GAAG,CAAC,OAAO,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAC,CAAC,CAAC,CAAC;IAC/B,CAAC;IACD,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;AAAA,CAC3B;AAED,MAAM,UAAU,mBAAmB,CAAC,GAAW,EAAgC;IAC9E,OAAO,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;AAAA,CACzB;AAED,MAAM,UAAU,0BAA0B,CAAC,GAAW,EAAQ;IAC7D,QAAQ,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;AAAA,CACrB","sourcesContent":["/**\n * Orchestrates semantic indexing and search for a repository.\n *\n * Lifecycle (all behind --enable-embsearchtools):\n * 1. `start()` — resolve the embsearch binary, scan the repo (ignore-aware),\n * apply the byte threshold. Under threshold → dormant. Over → spawn the\n * daemon, verify the backend is not the mock embedder, then index changed\n * files in the background in small batches, reporting progress.\n * 2. `search()` — top-k semantic query, mapping chunk ids back to\n * `path:start-end` via the sidecar metadata.\n * 3. `dispose()` — save + close the daemon.\n *\n * Every failure degrades to `unavailable` with a reason; nothing here ever\n * blocks session startup or affects grep/find.\n */\n\nimport { readFileSync } from \"fs\";\nimport { minimatch } from \"minimatch\";\nimport { ensureTool } from \"../../utils/tools-manager.js\";\nimport { chunkFile } from \"./chunker.js\";\nimport { EmbSearchClient } from \"./client.js\";\nimport {\n\temptyIndexMeta,\n\ttype FileMeta,\n\tgetEmbsearchStoreDir,\n\tgetVectorStoreDir,\n\thashContent,\n\ttype IndexMeta,\n\tloadIndexMeta,\n\tsaveIndexMeta,\n} from \"./index-meta.js\";\nimport { type RepoScanFile, scanRepo } from \"./repo-scan.js\";\n\n/** Chunks per bulk request. Small enough that a concurrent query is never\n * stuck long behind one padded batch inference. */\nconst BULK_BATCH_SIZE = 48;\n/** Yield between batches so background indexing doesn't starve the session. */\nconst BATCH_YIELD_MS = 15;\n/** The Rust mock backend's model id — semantically meaningless, never index with it. */\nconst MOCK_MODEL_ID = \"mock-hash-v1\";\n\nexport type EmbsearchState =\n\t| { phase: \"idle\" }\n\t| { phase: \"skipped\"; reason: string }\n\t| { phase: \"downloading\"; receivedBytes: number; totalBytes: number | null }\n\t| { phase: \"indexing\"; done: number; total: number }\n\t| { phase: \"ready\"; chunkCount: number }\n\t| { phase: \"unavailable\"; reason: string };\n\nexport interface EmbsearchServiceOptions {\n\tcwd: string;\n\t/** Explicit binary path (settings override). Default: \"embsearch\" from PATH. */\n\tbinaryPath?: string;\n\t/** Minimum indexable bytes before indexing kicks in. */\n\tthresholdBytes: number;\n\t/** Progress callback for UI (footer / stderr lines). */\n\tonProgress?: (state: EmbsearchState) => void;\n}\n\nexport interface SemanticHit {\n\tpath: string;\n\tstartLine: number;\n\tendLine: number;\n\tscore: number;\n}\n\nexport interface SemanticChunkHit extends SemanticHit {\n\t/** Per-build chunk id (`relpath#index`) — the fusion identity for hybrid search. */\n\tid: string;\n}\n\nexport class EmbsearchService {\n\tprivate readonly options: EmbsearchServiceOptions;\n\tprivate client: EmbSearchClient | undefined;\n\tprivate meta: IndexMeta | undefined;\n\tprivate state: EmbsearchState = { phase: \"idle\" };\n\tprivate disposed = false;\n\n\tconstructor(options: EmbsearchServiceOptions) {\n\t\tthis.options = options;\n\t}\n\n\tgetState(): EmbsearchState {\n\t\treturn this.state;\n\t}\n\n\t/** Semantic search is usable (index ready, or still building with partial data). */\n\tisAvailable(): boolean {\n\t\treturn this.state.phase === \"ready\" || this.state.phase === \"indexing\";\n\t}\n\n\tprivate setState(state: EmbsearchState): void {\n\t\tthis.state = state;\n\t\tthis.options.onProgress?.(state);\n\t}\n\n\tprivate async resolveBinary(): Promise<string | undefined> {\n\t\tif (this.options.binaryPath) {\n\t\t\treturn this.options.binaryPath;\n\t\t}\n\t\t// Surface the on-demand binary download through the same progress channel as\n\t\t// indexing, so the first-run fetch renders a progress bar instead of a stall.\n\t\t// A cached binary resolves without ever invoking this callback.\n\t\treturn await ensureTool(\"embsearch\", true, (receivedBytes, totalBytes) => {\n\t\t\tthis.setState({ phase: \"downloading\", receivedBytes, totalBytes });\n\t\t});\n\t}\n\n\t/**\n\t * Scan, threshold-check, and (when needed) index in the background.\n\t * Resolves when indexing completes or the feature settles dormant.\n\t */\n\tasync start(signal?: AbortSignal): Promise<void> {\n\t\ttry {\n\t\t\tawait this.run(signal);\n\t\t} catch (e) {\n\t\t\tconst reason = e instanceof Error ? e.message : String(e);\n\t\t\tthis.setState({ phase: \"unavailable\", reason });\n\t\t\tawait this.closeClient();\n\t\t}\n\t}\n\n\tprivate async run(signal?: AbortSignal): Promise<void> {\n\t\tconst binary = await this.resolveBinary();\n\t\tif (!binary) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"unavailable\",\n\t\t\t\treason: \"embsearch binary not found (PATH or embsearchBinaryPath setting)\",\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst scan = scanRepo(this.options.cwd, signal);\n\t\tif (scan.totalBytes < this.options.thresholdBytes) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"skipped\",\n\t\t\t\treason: `repo under threshold (${scan.totalBytes} < ${this.options.thresholdBytes} bytes)`,\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst storeDir = getEmbsearchStoreDir(this.options.cwd);\n\t\tthis.client = new EmbSearchClient({ binaryPath: binary, storePath: getVectorStoreDir(storeDir) });\n\t\tawait this.client.ready();\n\n\t\tconst info = await this.client.info();\n\t\tif (info.modelId === MOCK_MODEL_ID) {\n\t\t\tthrow new Error(\"embsearch binary uses the mock embedder (not semantic); install an onnx build\");\n\t\t}\n\n\t\t// Missing/stale sidecar (format, chunker, or model changed) → clean rebuild.\n\t\tthis.meta = loadIndexMeta(storeDir, info.modelId) ?? emptyIndexMeta(this.options.cwd, info.modelId);\n\t\tthis.meta.lastUsedMs = Date.now();\n\n\t\tawait this.indexChangedFiles(scan.files, storeDir, signal);\n\t}\n\n\tprivate async indexChangedFiles(files: RepoScanFile[], storeDir: string, signal?: AbortSignal): Promise<void> {\n\t\tconst meta = this.meta!;\n\t\tconst client = this.client!;\n\n\t\t// Diff scan vs sidecar: cheap mtime+size check first, hash only on delta.\n\t\tconst toIndex: Array<{ file: RepoScanFile; content: string; hash: string }> = [];\n\t\tconst seen = new Set<string>();\n\t\tfor (const file of files) {\n\t\t\tseen.add(file.rel);\n\t\t\tconst known = meta.files[file.rel];\n\t\t\tif (known && known.mtimeMs === file.mtimeMs && known.size === file.size) continue;\n\t\t\tlet content: string;\n\t\t\ttry {\n\t\t\t\tcontent = readFileSync(file.abs, \"utf-8\");\n\t\t\t} catch {\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tconst hash = hashContent(content);\n\t\t\tif (known && known.hash === hash) {\n\t\t\t\t// Touched but unchanged — refresh stat info only.\n\t\t\t\tknown.mtimeMs = file.mtimeMs;\n\t\t\t\tknown.size = file.size;\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\ttoIndex.push({ file, content, hash });\n\t\t}\n\t\tconst toRemove = Object.keys(meta.files).filter((rel) => !seen.has(rel));\n\n\t\tif (toIndex.length === 0 && toRemove.length === 0) {\n\t\t\tsaveIndexMeta(storeDir, meta);\n\t\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t\t\treturn;\n\t\t}\n\n\t\t// Chunk changed files; count total upserts for exact progress.\n\t\tconst work: Array<{ rel: string; fileMeta: FileMeta; chunks: Array<{ id: string; text: string }> }> = [];\n\t\tlet totalChunks = 0;\n\t\tfor (const { file, content, hash } of toIndex) {\n\t\t\tconst chunks = chunkFile(file.rel, content);\n\t\t\twork.push({\n\t\t\t\trel: file.rel,\n\t\t\t\tfileMeta: {\n\t\t\t\t\tmtimeMs: file.mtimeMs,\n\t\t\t\t\tsize: file.size,\n\t\t\t\t\thash,\n\t\t\t\t\tchunks: chunks.map((c) => [c.startLine, c.endLine]),\n\t\t\t\t},\n\t\t\t\tchunks: chunks.map((c) => ({ id: c.id, text: c.text })),\n\t\t\t});\n\t\t\ttotalChunks += chunks.length;\n\t\t}\n\n\t\tthis.setState({ phase: \"indexing\", done: 0, total: totalChunks });\n\n\t\t// Drop vectors of deleted files and superseded chunk tails.\n\t\tfor (const rel of toRemove) {\n\t\t\tfor (let i = 0; i < meta.files[rel].chunks.length; i++) await client.remove(`${rel}#${i}`);\n\t\t\tdelete meta.files[rel];\n\t\t}\n\n\t\tlet done = 0;\n\t\tfor (const item of work) {\n\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\tconst oldChunkCount = meta.files[item.rel]?.chunks.length ?? 0;\n\t\t\t// Remove old chunks beyond the new count (upsert covers the rest).\n\t\t\tfor (let i = item.chunks.length; i < oldChunkCount; i++) await client.remove(`${item.rel}#${i}`);\n\n\t\t\tfor (let offset = 0; offset < item.chunks.length; offset += BULK_BATCH_SIZE) {\n\t\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\t\tconst batch = item.chunks.slice(offset, offset + BULK_BATCH_SIZE);\n\t\t\t\tawait client.bulk(batch);\n\t\t\t\tdone += batch.length;\n\t\t\t\tthis.setState({ phase: \"indexing\", done, total: totalChunks });\n\t\t\t\t// Yield so queries and the event loop stay responsive.\n\t\t\t\tawait new Promise((resolve) => setTimeout(resolve, BATCH_YIELD_MS));\n\t\t\t}\n\t\t\tmeta.files[item.rel] = item.fileMeta;\n\t\t}\n\n\t\tawait client.compact();\n\t\tawait client.save();\n\t\tsaveIndexMeta(storeDir, meta);\n\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t}\n\n\tprivate countChunks(meta: IndexMeta): number {\n\t\tlet n = 0;\n\t\tfor (const rel of Object.keys(meta.files)) n += meta.files[rel].chunks.length;\n\t\treturn n;\n\t}\n\n\t/** Top-`k` semantic hits as `path` + line range + score. */\n\tasync search(query: string, k = 10): Promise<SemanticHit[]> {\n\t\treturn await this.searchChunks(query, k);\n\t}\n\n\t/** Top-`k` semantic hits including their chunk ids, for rank fusion. */\n\tasync searchChunks(query: string, k = 10, glob?: string): Promise<SemanticChunkHit[]> {\n\t\tif (!this.client || this.client.isClosed || !this.meta) {\n\t\t\tthrow new Error(\"semantic index is not available\");\n\t\t}\n\t\tconst results = await this.client.query(query, k);\n\t\tconst hits: SemanticChunkHit[] = [];\n\t\tconst matchGlob = (rel: string): boolean => {\n\t\t\tif (!glob) return true;\n\t\t\treturn minimatch(rel, glob, { dot: true, matchBase: !glob.includes(\"/\") });\n\t\t};\n\t\tfor (const result of results) {\n\t\t\tconst sep = result.id.lastIndexOf(\"#\");\n\t\t\tif (sep === -1) continue;\n\t\t\tconst rel = result.id.slice(0, sep);\n\t\t\tif (!matchGlob(rel)) continue;\n\t\t\tconst chunkIndex = Number.parseInt(result.id.slice(sep + 1), 10);\n\t\t\tconst range = this.meta.files[rel]?.chunks[chunkIndex];\n\t\t\tif (!range) continue;\n\t\t\thits.push({ id: result.id, path: rel, startLine: range[0], endLine: range[1], score: result.score });\n\t\t}\n\t\treturn hits;\n\t}\n\n\t/**\n\t * Resolve a repo-relative path + line to its enclosing indexed chunk, or\n\t * undefined when the file/line is not covered by the index. Chunks overlap\n\t * by a few lines; the first (lowest-index) containing chunk wins so the\n\t * mapping is deterministic.\n\t */\n\tfindEnclosingChunk(\n\t\trel: string,\n\t\tline: number,\n\t): { id: string; path: string; startLine: number; endLine: number } | undefined {\n\t\tconst file = this.meta?.files[rel];\n\t\tif (!file) return undefined;\n\t\tfor (let i = 0; i < file.chunks.length; i++) {\n\t\t\tconst [startLine, endLine] = file.chunks[i];\n\t\t\tif (line >= startLine && line <= endLine) {\n\t\t\t\treturn { id: `${rel}#${i}`, path: rel, startLine, endLine };\n\t\t\t}\n\t\t}\n\t\treturn undefined;\n\t}\n\n\tprivate async closeClient(): Promise<void> {\n\t\tconst client = this.client;\n\t\tthis.client = undefined;\n\t\tif (client && !client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait client.close();\n\t\t\t} catch {\n\t\t\t\t// already dead\n\t\t\t}\n\t\t}\n\t}\n\n\t/** Persist state and shut the daemon down. Safe to call twice. */\n\tasync dispose(): Promise<void> {\n\t\tif (this.disposed) return;\n\t\tthis.disposed = true;\n\t\tif (this.client && !this.client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait this.client.save();\n\t\t\t} catch {\n\t\t\t\t// daemon may have exited; nothing to save\n\t\t\t}\n\t\t}\n\t\tawait this.closeClient();\n\t}\n}\n\n// --- Per-cwd service registry ---\n//\n// The search tool is constructed by the generic tool factory table and\n// only receives `cwd`; the service is created later during session init (flag\n// gated). This registry connects the two without threading a service instance\n// through every layer between main.ts and the tool factories.\n\nconst services = new Map<string, EmbsearchService>();\n\nexport function registerEmbsearchService(cwd: string, service: EmbsearchService): void {\n\tconst old = services.get(cwd);\n\tif (old && old !== service) {\n\t\told.dispose().catch(() => {});\n\t}\n\tservices.set(cwd, service);\n}\n\nexport function getEmbsearchService(cwd: string): EmbsearchService | undefined {\n\treturn services.get(cwd);\n}\n\nexport function unregisterEmbsearchService(cwd: string): void {\n\tservices.delete(cwd);\n}\n"]}
|
|
@@ -0,0 +1,48 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Process-wide store for transient startup progress: first-run downloads of the
|
|
3
|
+
* external tool binaries (fd, rg, the semantic-index engine) and the semantic
|
|
4
|
+
* index build itself.
|
|
5
|
+
*
|
|
6
|
+
* Modeled on `taskStore` — a singleton the footer reads directly and re-renders
|
|
7
|
+
* on change — but deliberately separate from it: this is short-lived startup
|
|
8
|
+
* status, not agent work, so it must never render as a task-panel plan row.
|
|
9
|
+
* Entries are keyed so several concurrent downloads each own a stable footer
|
|
10
|
+
* line; a caller removes its key when the work settles and the line disappears.
|
|
11
|
+
*/
|
|
12
|
+
export type StartupProgress = {
|
|
13
|
+
readonly key: string;
|
|
14
|
+
kind: "download";
|
|
15
|
+
label: string;
|
|
16
|
+
receivedBytes: number;
|
|
17
|
+
totalBytes: number | null;
|
|
18
|
+
} | {
|
|
19
|
+
readonly key: string;
|
|
20
|
+
kind: "work";
|
|
21
|
+
label: string;
|
|
22
|
+
done: number;
|
|
23
|
+
total: number;
|
|
24
|
+
unit: string;
|
|
25
|
+
} | {
|
|
26
|
+
readonly key: string;
|
|
27
|
+
kind: "error";
|
|
28
|
+
label: string;
|
|
29
|
+
message: string;
|
|
30
|
+
};
|
|
31
|
+
type Listener = () => void;
|
|
32
|
+
declare class StartupProgressStore {
|
|
33
|
+
private readonly entries;
|
|
34
|
+
private readonly listeners;
|
|
35
|
+
/** Insert or replace the entry for its key. */
|
|
36
|
+
set(entry: StartupProgress): void;
|
|
37
|
+
/** Drop a key's line once its work settled (download done, index ready/skipped). */
|
|
38
|
+
remove(key: string): void;
|
|
39
|
+
/** Wipe everything — called when the first user turn starts (transient startup status). */
|
|
40
|
+
clear(): void;
|
|
41
|
+
list(): readonly StartupProgress[];
|
|
42
|
+
subscribe(listener: Listener): () => void;
|
|
43
|
+
private emit;
|
|
44
|
+
}
|
|
45
|
+
/** Shared, process-wide startup-progress store. */
|
|
46
|
+
export declare const startupProgress: StartupProgressStore;
|
|
47
|
+
export {};
|
|
48
|
+
//# sourceMappingURL=startup-progress.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"startup-progress.d.ts","sourceRoot":"","sources":["../../src/core/startup-progress.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,MAAM,MAAM,eAAe,GACxB;IAAE,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,UAAU,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,aAAa,EAAE,MAAM,CAAC;IAAC,UAAU,EAAE,MAAM,GAAG,IAAI,CAAA;CAAE,GAC3G;IAAE,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,GAChG;IAAE,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,OAAO,CAAC;IAAC,KAAK,EAAE,MAAM,CAAC;IAAC,OAAO,EAAE,MAAM,CAAA;CAAE,CAAC;AAE3E,KAAK,QAAQ,GAAG,MAAM,IAAI,CAAC;AAE3B,cAAM,oBAAoB;IAGzB,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAsC;IAC9D,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAuB;IAEjD,+CAA+C;IAC/C,GAAG,CAAC,KAAK,EAAE,eAAe,GAAG,IAAI,CAGhC;IAED,oFAAoF;IACpF,MAAM,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,CAExB;IAED,6FAA2F;IAC3F,KAAK,IAAI,IAAI,CAIZ;IAED,IAAI,IAAI,SAAS,eAAe,EAAE,CAEjC;IAED,SAAS,CAAC,QAAQ,EAAE,QAAQ,GAAG,MAAM,IAAI,CAKxC;IAED,OAAO,CAAC,IAAI;CAGZ;AAED,mDAAmD;AACnD,eAAO,MAAM,eAAe,sBAA6B,CAAC","sourcesContent":["/**\n * Process-wide store for transient startup progress: first-run downloads of the\n * external tool binaries (fd, rg, the semantic-index engine) and the semantic\n * index build itself.\n *\n * Modeled on `taskStore` — a singleton the footer reads directly and re-renders\n * on change — but deliberately separate from it: this is short-lived startup\n * status, not agent work, so it must never render as a task-panel plan row.\n * Entries are keyed so several concurrent downloads each own a stable footer\n * line; a caller removes its key when the work settles and the line disappears.\n */\n\nexport type StartupProgress =\n\t| { readonly key: string; kind: \"download\"; label: string; receivedBytes: number; totalBytes: number | null }\n\t| { readonly key: string; kind: \"work\"; label: string; done: number; total: number; unit: string }\n\t| { readonly key: string; kind: \"error\"; label: string; message: string };\n\ntype Listener = () => void;\n\nclass StartupProgressStore {\n\t// Map preserves insertion order, so lines stay in the order work started\n\t// (fd, rg, then the index) instead of jumping as byte counts update.\n\tprivate readonly entries = new Map<string, StartupProgress>();\n\tprivate readonly listeners = new Set<Listener>();\n\n\t/** Insert or replace the entry for its key. */\n\tset(entry: StartupProgress): void {\n\t\tthis.entries.set(entry.key, entry);\n\t\tthis.emit();\n\t}\n\n\t/** Drop a key's line once its work settled (download done, index ready/skipped). */\n\tremove(key: string): void {\n\t\tif (this.entries.delete(key)) this.emit();\n\t}\n\n\t/** Wipe everything — called when the first user turn starts (transient startup status). */\n\tclear(): void {\n\t\tif (this.entries.size === 0) return;\n\t\tthis.entries.clear();\n\t\tthis.emit();\n\t}\n\n\tlist(): readonly StartupProgress[] {\n\t\treturn [...this.entries.values()];\n\t}\n\n\tsubscribe(listener: Listener): () => void {\n\t\tthis.listeners.add(listener);\n\t\treturn () => {\n\t\t\tthis.listeners.delete(listener);\n\t\t};\n\t}\n\n\tprivate emit(): void {\n\t\tfor (const listener of this.listeners) listener();\n\t}\n}\n\n/** Shared, process-wide startup-progress store. */\nexport const startupProgress = new StartupProgressStore();\n"]}
|
|
@@ -0,0 +1,50 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Process-wide store for transient startup progress: first-run downloads of the
|
|
3
|
+
* external tool binaries (fd, rg, the semantic-index engine) and the semantic
|
|
4
|
+
* index build itself.
|
|
5
|
+
*
|
|
6
|
+
* Modeled on `taskStore` — a singleton the footer reads directly and re-renders
|
|
7
|
+
* on change — but deliberately separate from it: this is short-lived startup
|
|
8
|
+
* status, not agent work, so it must never render as a task-panel plan row.
|
|
9
|
+
* Entries are keyed so several concurrent downloads each own a stable footer
|
|
10
|
+
* line; a caller removes its key when the work settles and the line disappears.
|
|
11
|
+
*/
|
|
12
|
+
class StartupProgressStore {
|
|
13
|
+
// Map preserves insertion order, so lines stay in the order work started
|
|
14
|
+
// (fd, rg, then the index) instead of jumping as byte counts update.
|
|
15
|
+
entries = new Map();
|
|
16
|
+
listeners = new Set();
|
|
17
|
+
/** Insert or replace the entry for its key. */
|
|
18
|
+
set(entry) {
|
|
19
|
+
this.entries.set(entry.key, entry);
|
|
20
|
+
this.emit();
|
|
21
|
+
}
|
|
22
|
+
/** Drop a key's line once its work settled (download done, index ready/skipped). */
|
|
23
|
+
remove(key) {
|
|
24
|
+
if (this.entries.delete(key))
|
|
25
|
+
this.emit();
|
|
26
|
+
}
|
|
27
|
+
/** Wipe everything — called when the first user turn starts (transient startup status). */
|
|
28
|
+
clear() {
|
|
29
|
+
if (this.entries.size === 0)
|
|
30
|
+
return;
|
|
31
|
+
this.entries.clear();
|
|
32
|
+
this.emit();
|
|
33
|
+
}
|
|
34
|
+
list() {
|
|
35
|
+
return [...this.entries.values()];
|
|
36
|
+
}
|
|
37
|
+
subscribe(listener) {
|
|
38
|
+
this.listeners.add(listener);
|
|
39
|
+
return () => {
|
|
40
|
+
this.listeners.delete(listener);
|
|
41
|
+
};
|
|
42
|
+
}
|
|
43
|
+
emit() {
|
|
44
|
+
for (const listener of this.listeners)
|
|
45
|
+
listener();
|
|
46
|
+
}
|
|
47
|
+
}
|
|
48
|
+
/** Shared, process-wide startup-progress store. */
|
|
49
|
+
export const startupProgress = new StartupProgressStore();
|
|
50
|
+
//# sourceMappingURL=startup-progress.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"startup-progress.js","sourceRoot":"","sources":["../../src/core/startup-progress.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AASH,MAAM,oBAAoB;IACzB,yEAAyE;IACzE,qEAAqE;IACpD,OAAO,GAAG,IAAI,GAAG,EAA2B,CAAC;IAC7C,SAAS,GAAG,IAAI,GAAG,EAAY,CAAC;IAEjD,+CAA+C;IAC/C,GAAG,CAAC,KAAsB,EAAQ;QACjC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QACnC,IAAI,CAAC,IAAI,EAAE,CAAC;IAAA,CACZ;IAED,oFAAoF;IACpF,MAAM,CAAC,GAAW,EAAQ;QACzB,IAAI,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC;YAAE,IAAI,CAAC,IAAI,EAAE,CAAC;IAAA,CAC1C;IAED,6FAA2F;IAC3F,KAAK,GAAS;QACb,IAAI,IAAI,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC;YAAE,OAAO;QACpC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC;QACrB,IAAI,CAAC,IAAI,EAAE,CAAC;IAAA,CACZ;IAED,IAAI,GAA+B;QAClC,OAAO,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IAAA,CAClC;IAED,SAAS,CAAC,QAAkB,EAAc;QACzC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QAC7B,OAAO,GAAG,EAAE,CAAC;YACZ,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QAAA,CAChC,CAAC;IAAA,CACF;IAEO,IAAI,GAAS;QACpB,KAAK,MAAM,QAAQ,IAAI,IAAI,CAAC,SAAS;YAAE,QAAQ,EAAE,CAAC;IAAA,CAClD;CACD;AAED,mDAAmD;AACnD,MAAM,CAAC,MAAM,eAAe,GAAG,IAAI,oBAAoB,EAAE,CAAC","sourcesContent":["/**\n * Process-wide store for transient startup progress: first-run downloads of the\n * external tool binaries (fd, rg, the semantic-index engine) and the semantic\n * index build itself.\n *\n * Modeled on `taskStore` — a singleton the footer reads directly and re-renders\n * on change — but deliberately separate from it: this is short-lived startup\n * status, not agent work, so it must never render as a task-panel plan row.\n * Entries are keyed so several concurrent downloads each own a stable footer\n * line; a caller removes its key when the work settles and the line disappears.\n */\n\nexport type StartupProgress =\n\t| { readonly key: string; kind: \"download\"; label: string; receivedBytes: number; totalBytes: number | null }\n\t| { readonly key: string; kind: \"work\"; label: string; done: number; total: number; unit: string }\n\t| { readonly key: string; kind: \"error\"; label: string; message: string };\n\ntype Listener = () => void;\n\nclass StartupProgressStore {\n\t// Map preserves insertion order, so lines stay in the order work started\n\t// (fd, rg, then the index) instead of jumping as byte counts update.\n\tprivate readonly entries = new Map<string, StartupProgress>();\n\tprivate readonly listeners = new Set<Listener>();\n\n\t/** Insert or replace the entry for its key. */\n\tset(entry: StartupProgress): void {\n\t\tthis.entries.set(entry.key, entry);\n\t\tthis.emit();\n\t}\n\n\t/** Drop a key's line once its work settled (download done, index ready/skipped). */\n\tremove(key: string): void {\n\t\tif (this.entries.delete(key)) this.emit();\n\t}\n\n\t/** Wipe everything — called when the first user turn starts (transient startup status). */\n\tclear(): void {\n\t\tif (this.entries.size === 0) return;\n\t\tthis.entries.clear();\n\t\tthis.emit();\n\t}\n\n\tlist(): readonly StartupProgress[] {\n\t\treturn [...this.entries.values()];\n\t}\n\n\tsubscribe(listener: Listener): () => void {\n\t\tthis.listeners.add(listener);\n\t\treturn () => {\n\t\t\tthis.listeners.delete(listener);\n\t\t};\n\t}\n\n\tprivate emit(): void {\n\t\tfor (const listener of this.listeners) listener();\n\t}\n}\n\n/** Shared, process-wide startup-progress store. */\nexport const startupProgress = new StartupProgressStore();\n"]}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"system-prompt.d.ts","sourceRoot":"","sources":["../../src/core/system-prompt.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,EAAE,KAAK,eAAe,EAAkB,MAAM,wBAAwB,CAAC;AAE9E,OAAO,EAAyB,KAAK,KAAK,EAAE,MAAM,aAAa,CAAC;AAEhE,MAAM,WAAW,wBAAwB;IACxC,+CAA+C;IAC/C,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,6FAA6F;IAC7F,aAAa,CAAC,EAAE,MAAM,EAAE,CAAC;IACzB,0DAA0D;IAC1D,YAAY,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACtC,qFAAqF;IACrF,gBAAgB,CAAC,EAAE,MAAM,EAAE,CAAC;IAC5B,uCAAuC;IACvC,kBAAkB,CAAC,EAAE,MAAM,CAAC;IAC5B,yBAAyB;IACzB,GAAG,EAAE,MAAM,CAAC;IACZ,gCAAgC;IAChC,YAAY,CAAC,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IACxD,yBAAyB;IACzB,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC;IACjB;;;;OAIG;IACH,MAAM,CAAC,EAAE,eAAe,EAAE,CAAC;CAC3B;AAED,kEAAkE;AAClE,wBAAgB,iBAAiB,CAAC,OAAO,EAAE,wBAAwB,GAAG,MAAM,
|
|
1
|
+
{"version":3,"file":"system-prompt.d.ts","sourceRoot":"","sources":["../../src/core/system-prompt.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,EAAE,KAAK,eAAe,EAAkB,MAAM,wBAAwB,CAAC;AAE9E,OAAO,EAAyB,KAAK,KAAK,EAAE,MAAM,aAAa,CAAC;AAEhE,MAAM,WAAW,wBAAwB;IACxC,+CAA+C;IAC/C,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,6FAA6F;IAC7F,aAAa,CAAC,EAAE,MAAM,EAAE,CAAC;IACzB,0DAA0D;IAC1D,YAAY,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACtC,qFAAqF;IACrF,gBAAgB,CAAC,EAAE,MAAM,EAAE,CAAC;IAC5B,uCAAuC;IACvC,kBAAkB,CAAC,EAAE,MAAM,CAAC;IAC5B,yBAAyB;IACzB,GAAG,EAAE,MAAM,CAAC;IACZ,gCAAgC;IAChC,YAAY,CAAC,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IACxD,yBAAyB;IACzB,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC;IACjB;;;;OAIG;IACH,MAAM,CAAC,EAAE,eAAe,EAAE,CAAC;CAC3B;AAED,kEAAkE;AAClE,wBAAgB,iBAAiB,CAAC,OAAO,EAAE,wBAAwB,GAAG,MAAM,CA8K3E","sourcesContent":["/**\n * System prompt construction and project context loading\n */\n\nimport { type AgentDefinition, TASK_TOOL_NAME } from \"./agent-frontmatter.js\";\nimport { formatAgentsForPrompt } from \"./agent-registry.js\";\nimport { formatSkillsForPrompt, type Skill } from \"./skills.js\";\n\nexport interface BuildSystemPromptOptions {\n\t/** Custom system prompt (replaces default). */\n\tcustomPrompt?: string;\n\t/** Tools to include in prompt. Default: [read, bash, edit, write, search, grep, find, ls] */\n\tselectedTools?: string[];\n\t/** Optional one-line tool snippets keyed by tool name. */\n\ttoolSnippets?: Record<string, string>;\n\t/** Additional guideline bullets appended to the default system prompt guidelines. */\n\tpromptGuidelines?: string[];\n\t/** Text to append to system prompt. */\n\tappendSystemPrompt?: string;\n\t/** Working directory. */\n\tcwd: string;\n\t/** Pre-loaded context files. */\n\tcontextFiles?: Array<{ path: string; content: string }>;\n\t/** Pre-loaded skills. */\n\tskills?: Skill[];\n\t/**\n\t * Available agents for delegation, emitted as `<available_agents>` XML.\n\t * Only populated when the Task tool is active so the model knows which\n\t * agents exist without re-reading the agent registry each turn.\n\t */\n\tagents?: AgentDefinition[];\n}\n\n/** Build the system prompt with tools, guidelines, and context */\nexport function buildSystemPrompt(options: BuildSystemPromptOptions): string {\n\tconst {\n\t\tcustomPrompt,\n\t\tselectedTools,\n\t\ttoolSnippets,\n\t\tpromptGuidelines,\n\t\tappendSystemPrompt,\n\t\tcwd,\n\t\tcontextFiles: providedContextFiles,\n\t\tskills: providedSkills,\n\t\tagents: providedAgents,\n\t} = options;\n\tconst resolvedCwd = cwd;\n\tconst promptCwd = resolvedCwd.replace(/\\\\/g, \"/\");\n\n\tconst now = new Date();\n\tconst year = now.getFullYear();\n\tconst month = String(now.getMonth() + 1).padStart(2, \"0\");\n\tconst day = String(now.getDate()).padStart(2, \"0\");\n\tconst date = `${year}-${month}-${day}`;\n\n\tconst appendSection = appendSystemPrompt ? `\\n\\n${appendSystemPrompt}` : \"\";\n\n\tconst contextFiles = providedContextFiles ?? [];\n\tconst skills = providedSkills ?? [];\n\tconst agents = providedAgents ?? [];\n\n\tif (customPrompt) {\n\t\tlet prompt = customPrompt;\n\n\t\tif (appendSection) {\n\t\t\tprompt += appendSection;\n\t\t}\n\n\t\t// Append project context files\n\t\tif (contextFiles.length > 0) {\n\t\t\tprompt += \"\\n\\n# Project Context\\n\\n\";\n\t\t\tprompt += \"Project-specific instructions and guidelines:\\n\\n\";\n\t\t\tfor (const { path: filePath, content } of contextFiles) {\n\t\t\t\tprompt += `## ${filePath}\\n\\n${content}\\n\\n`;\n\t\t\t}\n\t\t}\n\n\t\t// Append skills section (only if read tool is available)\n\t\tconst hasRead = !selectedTools || selectedTools.includes(\"read\");\n\t\tif (hasRead && skills.length > 0) {\n\t\t\tprompt += formatSkillsForPrompt(skills);\n\t\t}\n\n\t\t// Append agents section (only when Task tool is active)\n\t\tconst hasTask = !selectedTools || selectedTools.includes(TASK_TOOL_NAME);\n\t\tif (hasTask && agents.length > 0) {\n\t\t\tprompt += formatAgentsForPrompt(agents);\n\t\t}\n\n\t\t// Add date and working directory last\n\t\tprompt += `\\nCurrent date: ${date}`;\n\t\tprompt += `\\nCurrent working directory: ${promptCwd}`;\n\n\t\treturn prompt;\n\t}\n\n\t// Build tools list based on selected tools.\n\t// A tool appears in Available tools only when the caller provides a one-line snippet.\n\tconst tools = selectedTools || [\"read\", \"bash\", \"edit\", \"write\", \"search\", \"grep\", \"find\", \"ls\"];\n\tconst visibleTools = tools.filter((name) => !!toolSnippets?.[name]);\n\tconst toolsList =\n\t\tvisibleTools.length > 0 ? visibleTools.map((name) => `- ${name}: ${toolSnippets![name]}`).join(\"\\n\") : \"(none)\";\n\n\t// Build guidelines based on which tools are actually available\n\tconst guidelinesList: string[] = [];\n\tconst guidelinesSet = new Set<string>();\n\tconst addGuideline = (guideline: string): void => {\n\t\tif (guidelinesSet.has(guideline)) {\n\t\t\treturn;\n\t\t}\n\t\tguidelinesSet.add(guideline);\n\t\tguidelinesList.push(guideline);\n\t};\n\n\tconst hasBash = tools.includes(\"bash\");\n\tconst hasSearch = tools.includes(\"search\");\n\tconst hasGrep = tools.includes(\"grep\");\n\tconst hasFind = tools.includes(\"find\");\n\tconst hasLs = tools.includes(\"ls\");\n\tconst hasRead = tools.includes(\"read\");\n\n\t// File exploration guidelines. Name only the tools that are actually\n\t// registered (the condition used to OR the three but hardcode all three\n\t// names, advertising tools that might not exist) and map each to its job so\n\t// the model picks the right one instead of defaulting to its bash habit.\n\tconst explore: string[] = [];\n\tif (hasSearch) explore.push(\"search (find where code lives by concept or identifier)\");\n\tif (hasGrep) explore.push(\"grep (exact line/regex search)\");\n\tif (hasFind) explore.push(\"find (locate files by name/glob)\");\n\tif (hasLs) explore.push(\"ls (list directory contents)\");\n\tif (explore.length > 0) {\n\t\taddGuideline(\n\t\t\t`For file exploration use the dedicated tools — ${explore.join(\", \")} — instead of bash; they are faster, and respect .gitignore where applicable`,\n\t\t);\n\t} else if (hasBash) {\n\t\taddGuideline(\"Use bash for file exploration (ls, rg/grep, find)\");\n\t}\n\n\t// Single source of truth for the search↔grep decision. Gated on both tools\n\t// being active so we never reference a tool that isn't registered — grep.ts\n\t// and search.ts intentionally no longer cross-reference each other, since a\n\t// tool factory can't know what else is in the bundle.\n\tif (hasSearch && hasGrep) {\n\t\taddGuideline(\n\t\t\t\"Between search and grep: search finds where code lives by concept, behavior, or half-known name (ranked results); grep enumerates exact matching lines, regexes, and counts (output proportional to matches)\",\n\t\t);\n\t}\n\n\tfor (const guideline of promptGuidelines ?? []) {\n\t\tconst normalized = guideline.trim();\n\t\tif (normalized.length > 0) {\n\t\t\taddGuideline(normalized);\n\t\t}\n\t}\n\n\t// Always include these\n\taddGuideline(\"Be concise in your responses\");\n\taddGuideline(\"No preamble or postamble; do not restate the task or summarize what you just did\");\n\taddGuideline('Do not add closers like \"Let me know\" or \"Hope this helps\"');\n\taddGuideline(\n\t\t\"Do not narrate routine tool calls or results — the permission gate already shows them; speak when you have the answer or need a decision\",\n\t);\n\taddGuideline(\n\t\t\"Match the surrounding code's conventions for comments, docstrings, and types — do not add or strip them by default\",\n\t);\n\taddGuideline(\"Show file paths clearly when working with files\");\n\n\tconst guidelines = guidelinesList.map((g) => `- ${g}`).join(\"\\n\");\n\n\tlet prompt = `You are an expert coding assistant operating inside hoocode, a coding agent harness. You help users by reading files, executing commands, editing code, and writing new files.\n\nAvailable tools:\n${toolsList}\n\nIn addition to the tools above, you may have access to other custom tools depending on the project.\n\nGuidelines:\n${guidelines}`;\n\n\tif (appendSection) {\n\t\tprompt += appendSection;\n\t}\n\n\t// Append project context files\n\tif (contextFiles.length > 0) {\n\t\tprompt += \"\\n\\n# Project Context\\n\\n\";\n\t\tprompt += \"Project-specific instructions and guidelines:\\n\\n\";\n\t\tfor (const { path: filePath, content } of contextFiles) {\n\t\t\tprompt += `## ${filePath}\\n\\n${content}\\n\\n`;\n\t\t}\n\t}\n\n\t// Append skills section (only if read tool is available)\n\tif (hasRead && skills.length > 0) {\n\t\tprompt += formatSkillsForPrompt(skills);\n\t}\n\n\t// Append agents section (only when Task tool is active)\n\tconst hasTask = tools.includes(TASK_TOOL_NAME);\n\tif (hasTask && agents.length > 0) {\n\t\tprompt += formatAgentsForPrompt(agents);\n\t}\n\n\t// Add date and working directory last\n\tprompt += `\\nCurrent date: ${date}`;\n\tprompt += `\\nCurrent working directory: ${promptCwd}`;\n\n\treturn prompt;\n}\n"]}
|
|
@@ -86,6 +86,13 @@ export function buildSystemPrompt(options) {
|
|
|
86
86
|
else if (hasBash) {
|
|
87
87
|
addGuideline("Use bash for file exploration (ls, rg/grep, find)");
|
|
88
88
|
}
|
|
89
|
+
// Single source of truth for the search↔grep decision. Gated on both tools
|
|
90
|
+
// being active so we never reference a tool that isn't registered — grep.ts
|
|
91
|
+
// and search.ts intentionally no longer cross-reference each other, since a
|
|
92
|
+
// tool factory can't know what else is in the bundle.
|
|
93
|
+
if (hasSearch && hasGrep) {
|
|
94
|
+
addGuideline("Between search and grep: search finds where code lives by concept, behavior, or half-known name (ranked results); grep enumerates exact matching lines, regexes, and counts (output proportional to matches)");
|
|
95
|
+
}
|
|
89
96
|
for (const guideline of promptGuidelines ?? []) {
|
|
90
97
|
const normalized = guideline.trim();
|
|
91
98
|
if (normalized.length > 0) {
|