@alextis59/athena 1.0.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +15 -0
- package/README.md +147 -0
- package/dist/agent/ask.d.ts +7 -0
- package/dist/agent/ask.d.ts.map +1 -0
- package/dist/agent/ask.js +35 -0
- package/dist/agent/ask.js.map +1 -0
- package/dist/agent/citations.d.ts +27 -0
- package/dist/agent/citations.d.ts.map +1 -0
- package/dist/agent/citations.js +190 -0
- package/dist/agent/citations.js.map +1 -0
- package/dist/agent/fake-provider.d.ts +9 -0
- package/dist/agent/fake-provider.d.ts.map +1 -0
- package/dist/agent/fake-provider.js +51 -0
- package/dist/agent/fake-provider.js.map +1 -0
- package/dist/agent/index.d.ts +10 -0
- package/dist/agent/index.d.ts.map +1 -0
- package/dist/agent/index.js +6 -0
- package/dist/agent/index.js.map +1 -0
- package/dist/agent/loop.d.ts +11 -0
- package/dist/agent/loop.d.ts.map +1 -0
- package/dist/agent/loop.js +156 -0
- package/dist/agent/loop.js.map +1 -0
- package/dist/agent/providers.d.ts +17 -0
- package/dist/agent/providers.d.ts.map +1 -0
- package/dist/agent/providers.js +171 -0
- package/dist/agent/providers.js.map +1 -0
- package/dist/agent/tools.d.ts +8 -0
- package/dist/agent/tools.d.ts.map +1 -0
- package/dist/agent/tools.js +233 -0
- package/dist/agent/tools.js.map +1 -0
- package/dist/agent/types.d.ts +84 -0
- package/dist/agent/types.d.ts.map +1 -0
- package/dist/agent/types.js +2 -0
- package/dist/agent/types.js.map +1 -0
- package/dist/cli/help.d.ts +2 -0
- package/dist/cli/help.d.ts.map +1 -0
- package/dist/cli/help.js +28 -0
- package/dist/cli/help.js.map +1 -0
- package/dist/cli/main.d.ts +3 -0
- package/dist/cli/main.d.ts.map +1 -0
- package/dist/cli/main.js +5 -0
- package/dist/cli/main.js.map +1 -0
- package/dist/cli/options.d.ts +21 -0
- package/dist/cli/options.d.ts.map +1 -0
- package/dist/cli/options.js +97 -0
- package/dist/cli/options.js.map +1 -0
- package/dist/cli/run.d.ts +21 -0
- package/dist/cli/run.d.ts.map +1 -0
- package/dist/cli/run.js +144 -0
- package/dist/cli/run.js.map +1 -0
- package/dist/config/index.d.ts +4 -0
- package/dist/config/index.d.ts.map +1 -0
- package/dist/config/index.js +3 -0
- package/dist/config/index.js.map +1 -0
- package/dist/config/loader.d.ts +14 -0
- package/dist/config/loader.d.ts.map +1 -0
- package/dist/config/loader.js +352 -0
- package/dist/config/loader.js.map +1 -0
- package/dist/config/schema.d.ts +6 -0
- package/dist/config/schema.d.ts.map +1 -0
- package/dist/config/schema.js +72 -0
- package/dist/config/schema.js.map +1 -0
- package/dist/config/types.d.ts +53 -0
- package/dist/config/types.d.ts.map +1 -0
- package/dist/config/types.js +2 -0
- package/dist/config/types.js.map +1 -0
- package/dist/doctor/doctor.d.ts +47 -0
- package/dist/doctor/doctor.d.ts.map +1 -0
- package/dist/doctor/doctor.js +159 -0
- package/dist/doctor/doctor.js.map +1 -0
- package/dist/embeddings/index.d.ts +117 -0
- package/dist/embeddings/index.d.ts.map +1 -0
- package/dist/embeddings/index.js +515 -0
- package/dist/embeddings/index.js.map +1 -0
- package/dist/errors.d.ts +19 -0
- package/dist/errors.d.ts.map +1 -0
- package/dist/errors.js +37 -0
- package/dist/errors.js.map +1 -0
- package/dist/index.d.ts +17 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +9 -0
- package/dist/index.js.map +1 -0
- package/dist/indexer/discovery.d.ts +10 -0
- package/dist/indexer/discovery.d.ts.map +1 -0
- package/dist/indexer/discovery.js +38 -0
- package/dist/indexer/discovery.js.map +1 -0
- package/dist/indexer/hash.d.ts +3 -0
- package/dist/indexer/hash.d.ts.map +1 -0
- package/dist/indexer/hash.js +11 -0
- package/dist/indexer/hash.js.map +1 -0
- package/dist/indexer/index.d.ts +7 -0
- package/dist/indexer/index.d.ts.map +1 -0
- package/dist/indexer/index.js +5 -0
- package/dist/indexer/index.js.map +1 -0
- package/dist/indexer/kinds.d.ts +4 -0
- package/dist/indexer/kinds.d.ts.map +1 -0
- package/dist/indexer/kinds.js +38 -0
- package/dist/indexer/kinds.js.map +1 -0
- package/dist/indexer/run.d.ts +42 -0
- package/dist/indexer/run.d.ts.map +1 -0
- package/dist/indexer/run.js +563 -0
- package/dist/indexer/run.js.map +1 -0
- package/dist/logging/logger.d.ts +11 -0
- package/dist/logging/logger.d.ts.map +1 -0
- package/dist/logging/logger.js +18 -0
- package/dist/logging/logger.js.map +1 -0
- package/dist/mcp/index.d.ts +35 -0
- package/dist/mcp/index.d.ts.map +1 -0
- package/dist/mcp/index.js +392 -0
- package/dist/mcp/index.js.map +1 -0
- package/dist/parsers/chunking.d.ts +4 -0
- package/dist/parsers/chunking.d.ts.map +1 -0
- package/dist/parsers/chunking.js +112 -0
- package/dist/parsers/chunking.js.map +1 -0
- package/dist/parsers/index.d.ts +8 -0
- package/dist/parsers/index.d.ts.map +1 -0
- package/dist/parsers/index.js +5 -0
- package/dist/parsers/index.js.map +1 -0
- package/dist/parsers/line-parser.d.ts +13 -0
- package/dist/parsers/line-parser.d.ts.map +1 -0
- package/dist/parsers/line-parser.js +242 -0
- package/dist/parsers/line-parser.js.map +1 -0
- package/dist/parsers/normalize.d.ts +4 -0
- package/dist/parsers/normalize.d.ts.map +1 -0
- package/dist/parsers/normalize.js +17 -0
- package/dist/parsers/normalize.js.map +1 -0
- package/dist/parsers/parser.d.ts +14 -0
- package/dist/parsers/parser.d.ts.map +1 -0
- package/dist/parsers/parser.js +77 -0
- package/dist/parsers/parser.js.map +1 -0
- package/dist/parsers/pdf2llm.d.ts +14 -0
- package/dist/parsers/pdf2llm.d.ts.map +1 -0
- package/dist/parsers/pdf2llm.js +215 -0
- package/dist/parsers/pdf2llm.js.map +1 -0
- package/dist/parsers/types.d.ts +48 -0
- package/dist/parsers/types.d.ts.map +1 -0
- package/dist/parsers/types.js +2 -0
- package/dist/parsers/types.js.map +1 -0
- package/dist/retrieval/index.d.ts +4 -0
- package/dist/retrieval/index.d.ts.map +1 -0
- package/dist/retrieval/index.js +3 -0
- package/dist/retrieval/index.js.map +1 -0
- package/dist/retrieval/schemas.d.ts +127 -0
- package/dist/retrieval/schemas.d.ts.map +1 -0
- package/dist/retrieval/schemas.js +65 -0
- package/dist/retrieval/schemas.js.map +1 -0
- package/dist/retrieval/tools.d.ts +31 -0
- package/dist/retrieval/tools.d.ts.map +1 -0
- package/dist/retrieval/tools.js +803 -0
- package/dist/retrieval/tools.js.map +1 -0
- package/dist/retrieval/types.d.ts +104 -0
- package/dist/retrieval/types.d.ts.map +1 -0
- package/dist/retrieval/types.js +2 -0
- package/dist/retrieval/types.js.map +1 -0
- package/dist/server/index.d.ts +3 -0
- package/dist/server/index.d.ts.map +1 -0
- package/dist/server/index.js +2 -0
- package/dist/server/index.js.map +1 -0
- package/dist/server/server.d.ts +49 -0
- package/dist/server/server.d.ts.map +1 -0
- package/dist/server/server.js +517 -0
- package/dist/server/server.js.map +1 -0
- package/dist/shared/types.d.ts +98 -0
- package/dist/shared/types.d.ts.map +1 -0
- package/dist/shared/types.js +2 -0
- package/dist/shared/types.js.map +1 -0
- package/dist/storage/database.d.ts +18 -0
- package/dist/storage/database.d.ts.map +1 -0
- package/dist/storage/database.js +66 -0
- package/dist/storage/database.js.map +1 -0
- package/dist/storage/index.d.ts +8 -0
- package/dist/storage/index.d.ts.map +1 -0
- package/dist/storage/index.js +5 -0
- package/dist/storage/index.js.map +1 -0
- package/dist/storage/migrations.d.ts +15 -0
- package/dist/storage/migrations.d.ts.map +1 -0
- package/dist/storage/migrations.js +170 -0
- package/dist/storage/migrations.js.map +1 -0
- package/dist/storage/paths.d.ts +15 -0
- package/dist/storage/paths.d.ts.map +1 -0
- package/dist/storage/paths.js +27 -0
- package/dist/storage/paths.js.map +1 -0
- package/dist/storage/schema.d.ts +11 -0
- package/dist/storage/schema.d.ts.map +1 -0
- package/dist/storage/schema.js +40 -0
- package/dist/storage/schema.js.map +1 -0
- package/dist/ui/index.d.ts +2 -0
- package/dist/ui/index.d.ts.map +1 -0
- package/dist/ui/index.js +2 -0
- package/dist/ui/index.js.map +1 -0
- package/dist/ui/static-app.d.ts +2 -0
- package/dist/ui/static-app.d.ts.map +1 -0
- package/dist/ui/static-app.js +735 -0
- package/dist/ui/static-app.js.map +1 -0
- package/docs/adr/0001-sqlite-driver.md +23 -0
- package/docs/agentic-rag-specifications.md +659 -0
- package/docs/agentic-rag-study.md +1056 -0
- package/docs/implementation-plan.md +1425 -0
- package/docs/mcp-integration.md +154 -0
- package/docs/validation-corpus.md +71 -0
- package/package.json +79 -0
|
@@ -0,0 +1,35 @@
|
|
|
1
|
+
import { RetrievalToolRuntime } from "../retrieval/index.js";
|
|
2
|
+
export interface McpServerOptions {
|
|
3
|
+
root: string;
|
|
4
|
+
dbPath?: string;
|
|
5
|
+
input: AsyncIterable<string | Uint8Array>;
|
|
6
|
+
output: McpWritable;
|
|
7
|
+
}
|
|
8
|
+
export interface McpWritable {
|
|
9
|
+
write(chunk: string | Uint8Array): unknown;
|
|
10
|
+
}
|
|
11
|
+
type JsonRpcId = string | number;
|
|
12
|
+
interface JsonRpcResponse {
|
|
13
|
+
jsonrpc: "2.0";
|
|
14
|
+
id: JsonRpcId | null;
|
|
15
|
+
result?: unknown;
|
|
16
|
+
error?: {
|
|
17
|
+
code: number;
|
|
18
|
+
message: string;
|
|
19
|
+
data?: unknown;
|
|
20
|
+
};
|
|
21
|
+
}
|
|
22
|
+
interface McpToolDefinition {
|
|
23
|
+
name: McpToolName;
|
|
24
|
+
description: string;
|
|
25
|
+
inputSchema: unknown;
|
|
26
|
+
annotations: {
|
|
27
|
+
readOnlyHint: true;
|
|
28
|
+
};
|
|
29
|
+
}
|
|
30
|
+
type McpToolName = "searchDocs" | "readSourceRange" | "readPdfSource" | "grepDocs" | "listDocs" | "relatedChunks";
|
|
31
|
+
export declare const mcpToolDefinitions: McpToolDefinition[];
|
|
32
|
+
export declare function runMcpServer(options: McpServerOptions): Promise<void>;
|
|
33
|
+
export declare function handleMcpMessage(message: unknown, runtime: RetrievalToolRuntime): Promise<JsonRpcResponse[]>;
|
|
34
|
+
export {};
|
|
35
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/mcp/index.ts"],"names":[],"mappings":"AAIA,OAAO,EAAE,oBAAoB,EAAe,MAAM,uBAAuB,CAAC;AAe1E,MAAM,WAAW,gBAAgB;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,aAAa,CAAC,MAAM,GAAG,UAAU,CAAC,CAAC;IAC1C,MAAM,EAAE,WAAW,CAAC;CACrB;AAED,MAAM,WAAW,WAAW;IAC1B,KAAK,CAAC,KAAK,EAAE,MAAM,GAAG,UAAU,GAAG,OAAO,CAAC;CAC5C;AAED,KAAK,SAAS,GAAG,MAAM,GAAG,MAAM,CAAC;AAejC,UAAU,eAAe;IACvB,OAAO,EAAE,KAAK,CAAC;IACf,EAAE,EAAE,SAAS,GAAG,IAAI,CAAC;IACrB,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,KAAK,CAAC,EAAE;QACN,IAAI,EAAE,MAAM,CAAC;QACb,OAAO,EAAE,MAAM,CAAC;QAChB,IAAI,CAAC,EAAE,OAAO,CAAC;KAChB,CAAC;CACH;AAED,UAAU,iBAAiB;IACzB,IAAI,EAAE,WAAW,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,WAAW,EAAE,OAAO,CAAC;IACrB,WAAW,EAAE;QACX,YAAY,EAAE,IAAI,CAAC;KACpB,CAAC;CACH;AAED,KAAK,WAAW,GACZ,YAAY,GACZ,iBAAiB,GACjB,eAAe,GACf,UAAU,GACV,UAAU,GACV,eAAe,CAAC;AAEpB,eAAO,MAAM,kBAAkB,EAAE,iBAAiB,EA2CjD,CAAC;AAIF,wBAAsB,YAAY,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,IAAI,CAAC,CAmB3E;AAED,wBAAsB,gBAAgB,CACpC,OAAO,EAAE,OAAO,EAChB,OAAO,EAAE,oBAAoB,GAC5B,OAAO,CAAC,eAAe,EAAE,CAAC,CAc5B"}
|
|
@@ -0,0 +1,392 @@
|
|
|
1
|
+
import path from "node:path";
|
|
2
|
+
import { UserFacingError } from "../errors.js";
|
|
3
|
+
import { isSupportedSourceKind } from "../indexer/kinds.js";
|
|
4
|
+
import { RetrievalToolRuntime, toolSchemas } from "../retrieval/index.js";
|
|
5
|
+
const mcpProtocolVersion = "2025-03-26";
|
|
6
|
+
const serverVersion = "1.0.0";
|
|
7
|
+
export const mcpToolDefinitions = [
|
|
8
|
+
{
|
|
9
|
+
name: "searchDocs",
|
|
10
|
+
description: "Search indexed documentation chunks with Athena hybrid retrieval and citation locators.",
|
|
11
|
+
inputSchema: toolSchemas.searchDocs,
|
|
12
|
+
annotations: { readOnlyHint: true }
|
|
13
|
+
},
|
|
14
|
+
{
|
|
15
|
+
name: "readSourceRange",
|
|
16
|
+
description: "Read a repository-relative text source range by line number for citation-backed answers.",
|
|
17
|
+
inputSchema: toolSchemas.readSourceRange,
|
|
18
|
+
annotations: { readOnlyHint: true }
|
|
19
|
+
},
|
|
20
|
+
{
|
|
21
|
+
name: "readPdfSource",
|
|
22
|
+
description: "Read generated PDF Markdown plus page, region, confidence, warning, and sidecar metadata.",
|
|
23
|
+
inputSchema: toolSchemas.readPdfSource,
|
|
24
|
+
annotations: { readOnlyHint: true }
|
|
25
|
+
},
|
|
26
|
+
{
|
|
27
|
+
name: "grepDocs",
|
|
28
|
+
description: "Search indexed repository documentation lines with a literal or bounded regular expression.",
|
|
29
|
+
inputSchema: toolSchemas.grepDocs,
|
|
30
|
+
annotations: { readOnlyHint: true }
|
|
31
|
+
},
|
|
32
|
+
{
|
|
33
|
+
name: "listDocs",
|
|
34
|
+
description: "List indexed documentation sources by optional path prefix and source kind.",
|
|
35
|
+
inputSchema: toolSchemas.listDocs,
|
|
36
|
+
annotations: { readOnlyHint: true }
|
|
37
|
+
},
|
|
38
|
+
{
|
|
39
|
+
name: "relatedChunks",
|
|
40
|
+
description: "Read neighboring chunks around an already discovered chunk without mutating the index.",
|
|
41
|
+
inputSchema: toolSchemas.relatedChunks,
|
|
42
|
+
annotations: { readOnlyHint: true }
|
|
43
|
+
}
|
|
44
|
+
];
|
|
45
|
+
const mcpToolNames = new Set(mcpToolDefinitions.map((tool) => tool.name));
|
|
46
|
+
export async function runMcpServer(options) {
|
|
47
|
+
const runtime = new RetrievalToolRuntime(options.root, options.dbPath);
|
|
48
|
+
try {
|
|
49
|
+
for await (const message of readJsonRpcMessages(options.input)) {
|
|
50
|
+
if (isJsonRpcResponse(message)) {
|
|
51
|
+
options.output.write(`${JSON.stringify(message)}\n`);
|
|
52
|
+
continue;
|
|
53
|
+
}
|
|
54
|
+
const responses = await handleMcpMessage(message, runtime);
|
|
55
|
+
for (const response of responses) {
|
|
56
|
+
options.output.write(`${JSON.stringify(response)}\n`);
|
|
57
|
+
}
|
|
58
|
+
}
|
|
59
|
+
}
|
|
60
|
+
finally {
|
|
61
|
+
runtime.close();
|
|
62
|
+
}
|
|
63
|
+
}
|
|
64
|
+
export async function handleMcpMessage(message, runtime) {
|
|
65
|
+
if (Array.isArray(message)) {
|
|
66
|
+
const responses = await Promise.all(message.map(async (entry) => await handleSingleMcpMessage(entry, runtime)));
|
|
67
|
+
return responses.filter((response) => response !== null);
|
|
68
|
+
}
|
|
69
|
+
const response = await handleSingleMcpMessage(message, runtime);
|
|
70
|
+
return response === null ? [] : [response];
|
|
71
|
+
}
|
|
72
|
+
async function handleSingleMcpMessage(message, runtime) {
|
|
73
|
+
if (!isRecord(message) || message.jsonrpc !== "2.0") {
|
|
74
|
+
return jsonRpcError(null, -32600, "Invalid JSON-RPC request.");
|
|
75
|
+
}
|
|
76
|
+
if (typeof message.method !== "string") {
|
|
77
|
+
return jsonRpcError(readRequestId(message), -32600, "Missing method.");
|
|
78
|
+
}
|
|
79
|
+
if (!("id" in message)) {
|
|
80
|
+
handleMcpNotification({
|
|
81
|
+
jsonrpc: "2.0",
|
|
82
|
+
method: message.method,
|
|
83
|
+
...(message.params === undefined ? {} : { params: message.params })
|
|
84
|
+
});
|
|
85
|
+
return null;
|
|
86
|
+
}
|
|
87
|
+
const id = readRequestId(message);
|
|
88
|
+
if (id === null) {
|
|
89
|
+
return jsonRpcError(null, -32600, "JSON-RPC request id must be a string or number.");
|
|
90
|
+
}
|
|
91
|
+
try {
|
|
92
|
+
return jsonRpcResult(id, await handleMcpRequest({
|
|
93
|
+
jsonrpc: "2.0",
|
|
94
|
+
id,
|
|
95
|
+
method: message.method,
|
|
96
|
+
...(message.params === undefined ? {} : { params: message.params })
|
|
97
|
+
}, runtime));
|
|
98
|
+
}
|
|
99
|
+
catch (error) {
|
|
100
|
+
return toJsonRpcError(id, error);
|
|
101
|
+
}
|
|
102
|
+
}
|
|
103
|
+
function handleMcpNotification(message) {
|
|
104
|
+
switch (message.method) {
|
|
105
|
+
case "notifications/initialized":
|
|
106
|
+
case "notifications/cancelled":
|
|
107
|
+
return;
|
|
108
|
+
default:
|
|
109
|
+
return;
|
|
110
|
+
}
|
|
111
|
+
}
|
|
112
|
+
async function handleMcpRequest(request, runtime) {
|
|
113
|
+
switch (request.method) {
|
|
114
|
+
case "initialize":
|
|
115
|
+
return createInitializeResult();
|
|
116
|
+
case "ping":
|
|
117
|
+
return {};
|
|
118
|
+
case "tools/list":
|
|
119
|
+
return { tools: mcpToolDefinitions };
|
|
120
|
+
case "tools/call":
|
|
121
|
+
return await callMcpTool(runtime, request.params);
|
|
122
|
+
default:
|
|
123
|
+
throw new JsonRpcProtocolError(-32601, `Unknown MCP method: ${request.method}`);
|
|
124
|
+
}
|
|
125
|
+
}
|
|
126
|
+
function createInitializeResult() {
|
|
127
|
+
return {
|
|
128
|
+
protocolVersion: mcpProtocolVersion,
|
|
129
|
+
capabilities: {
|
|
130
|
+
tools: {
|
|
131
|
+
listChanged: false
|
|
132
|
+
}
|
|
133
|
+
},
|
|
134
|
+
serverInfo: {
|
|
135
|
+
name: "athena",
|
|
136
|
+
version: serverVersion
|
|
137
|
+
},
|
|
138
|
+
instructions: "Athena exposes read-only, citation-aware documentation retrieval tools. External agents consume retrieval context but do not own Athena's agent loop."
|
|
139
|
+
};
|
|
140
|
+
}
|
|
141
|
+
async function callMcpTool(runtime, params) {
|
|
142
|
+
const input = requireRecord(params, "tools/call params");
|
|
143
|
+
const name = requireString(input.name, "name");
|
|
144
|
+
if (!isMcpToolName(name)) {
|
|
145
|
+
throw new JsonRpcProtocolError(-32602, `Unknown tool: ${name}`);
|
|
146
|
+
}
|
|
147
|
+
try {
|
|
148
|
+
const result = await executeMcpTool(runtime, name, input.arguments ?? {});
|
|
149
|
+
return {
|
|
150
|
+
content: [
|
|
151
|
+
{
|
|
152
|
+
type: "text",
|
|
153
|
+
text: JSON.stringify({
|
|
154
|
+
result,
|
|
155
|
+
trace: serializeRetrievalTrace(runtime.trace)
|
|
156
|
+
}, null, 2)
|
|
157
|
+
}
|
|
158
|
+
],
|
|
159
|
+
isError: false
|
|
160
|
+
};
|
|
161
|
+
}
|
|
162
|
+
catch (error) {
|
|
163
|
+
return {
|
|
164
|
+
content: [
|
|
165
|
+
{
|
|
166
|
+
type: "text",
|
|
167
|
+
text: error instanceof Error ? error.message : String(error)
|
|
168
|
+
}
|
|
169
|
+
],
|
|
170
|
+
isError: true
|
|
171
|
+
};
|
|
172
|
+
}
|
|
173
|
+
}
|
|
174
|
+
async function executeMcpTool(runtime, name, args) {
|
|
175
|
+
switch (name) {
|
|
176
|
+
case "searchDocs":
|
|
177
|
+
return runtime.searchDocs(readSearchDocsInput(args));
|
|
178
|
+
case "readSourceRange":
|
|
179
|
+
return await runtime.readSourceRange(readSourceRangeInput(args));
|
|
180
|
+
case "readPdfSource":
|
|
181
|
+
return runtime.readPdfSource(readPdfSourceInput(args));
|
|
182
|
+
case "grepDocs":
|
|
183
|
+
return await runtime.grepDocs(readGrepDocsInput(args));
|
|
184
|
+
case "listDocs":
|
|
185
|
+
return runtime.listDocs(readListDocsInput(args));
|
|
186
|
+
case "relatedChunks":
|
|
187
|
+
return runtime.relatedChunks(readRelatedChunksInput(args));
|
|
188
|
+
}
|
|
189
|
+
}
|
|
190
|
+
function serializeRetrievalTrace(trace) {
|
|
191
|
+
return {
|
|
192
|
+
seenChunkIds: [...trace.seenChunkIds],
|
|
193
|
+
readLocators: trace.readLocators,
|
|
194
|
+
searchScores: trace.searchScores
|
|
195
|
+
};
|
|
196
|
+
}
|
|
197
|
+
function readSearchDocsInput(value) {
|
|
198
|
+
const input = requireRecord(value, "searchDocs arguments");
|
|
199
|
+
return {
|
|
200
|
+
query: requireString(input.query, "query"),
|
|
201
|
+
...optionalLimit(input.limit, "limit", 1, 50),
|
|
202
|
+
...optionalPathPrefix(input.pathPrefix),
|
|
203
|
+
...optionalKind(input.kind)
|
|
204
|
+
};
|
|
205
|
+
}
|
|
206
|
+
function readSourceRangeInput(value) {
|
|
207
|
+
const input = requireRecord(value, "readSourceRange arguments");
|
|
208
|
+
return {
|
|
209
|
+
path: requireRelativePath(input.path, "path"),
|
|
210
|
+
startLine: requireInteger(input.startLine, "startLine", 1),
|
|
211
|
+
endLine: requireInteger(input.endLine, "endLine", 1)
|
|
212
|
+
};
|
|
213
|
+
}
|
|
214
|
+
function readPdfSourceInput(value) {
|
|
215
|
+
const input = requireRecord(value, "readPdfSource arguments");
|
|
216
|
+
return {
|
|
217
|
+
path: requireRelativePath(input.path, "path"),
|
|
218
|
+
...optionalLimit(input.page, "page", 1, Number.MAX_SAFE_INTEGER),
|
|
219
|
+
...optionalString(input.regionId, "regionId")
|
|
220
|
+
};
|
|
221
|
+
}
|
|
222
|
+
function readGrepDocsInput(value) {
|
|
223
|
+
const input = requireRecord(value, "grepDocs arguments");
|
|
224
|
+
return {
|
|
225
|
+
pattern: requireString(input.pattern, "pattern"),
|
|
226
|
+
...optionalBoolean(input.regex, "regex"),
|
|
227
|
+
...optionalLimit(input.limit, "limit", 1, 100),
|
|
228
|
+
...optionalPathPrefix(input.pathPrefix),
|
|
229
|
+
...optionalKind(input.kind)
|
|
230
|
+
};
|
|
231
|
+
}
|
|
232
|
+
function readListDocsInput(value) {
|
|
233
|
+
const input = requireRecord(value, "listDocs arguments");
|
|
234
|
+
return {
|
|
235
|
+
...optionalLimit(input.limit, "limit", 1, 500),
|
|
236
|
+
...optionalPathPrefix(input.pathPrefix),
|
|
237
|
+
...optionalKind(input.kind)
|
|
238
|
+
};
|
|
239
|
+
}
|
|
240
|
+
function readRelatedChunksInput(value) {
|
|
241
|
+
const input = requireRecord(value, "relatedChunks arguments");
|
|
242
|
+
return {
|
|
243
|
+
chunkId: requireString(input.chunkId, "chunkId"),
|
|
244
|
+
...optionalLimit(input.before, "before", 0, 10),
|
|
245
|
+
...optionalLimit(input.after, "after", 0, 10)
|
|
246
|
+
};
|
|
247
|
+
}
|
|
248
|
+
async function* readJsonRpcMessages(input) {
|
|
249
|
+
let buffered = "";
|
|
250
|
+
for await (const chunk of input) {
|
|
251
|
+
buffered += chunk.toString();
|
|
252
|
+
while (true) {
|
|
253
|
+
const newlineIndex = buffered.indexOf("\n");
|
|
254
|
+
if (newlineIndex === -1) {
|
|
255
|
+
break;
|
|
256
|
+
}
|
|
257
|
+
const line = buffered.slice(0, newlineIndex).trim();
|
|
258
|
+
buffered = buffered.slice(newlineIndex + 1);
|
|
259
|
+
if (line === "") {
|
|
260
|
+
continue;
|
|
261
|
+
}
|
|
262
|
+
yield parseJsonRpcLine(line);
|
|
263
|
+
}
|
|
264
|
+
}
|
|
265
|
+
const trailing = buffered.trim();
|
|
266
|
+
if (trailing !== "") {
|
|
267
|
+
yield parseJsonRpcLine(trailing);
|
|
268
|
+
}
|
|
269
|
+
}
|
|
270
|
+
function parseJsonRpcLine(line) {
|
|
271
|
+
try {
|
|
272
|
+
return JSON.parse(line);
|
|
273
|
+
}
|
|
274
|
+
catch (error) {
|
|
275
|
+
return jsonRpcError(null, -32700, "Parse error.", {
|
|
276
|
+
message: error instanceof Error ? error.message : String(error)
|
|
277
|
+
});
|
|
278
|
+
}
|
|
279
|
+
}
|
|
280
|
+
function isJsonRpcResponse(value) {
|
|
281
|
+
return (isRecord(value) &&
|
|
282
|
+
value.jsonrpc === "2.0" &&
|
|
283
|
+
"id" in value &&
|
|
284
|
+
("result" in value || "error" in value) &&
|
|
285
|
+
!("method" in value));
|
|
286
|
+
}
|
|
287
|
+
function jsonRpcResult(id, result) {
|
|
288
|
+
return {
|
|
289
|
+
jsonrpc: "2.0",
|
|
290
|
+
id,
|
|
291
|
+
result
|
|
292
|
+
};
|
|
293
|
+
}
|
|
294
|
+
function jsonRpcError(id, code, message, data) {
|
|
295
|
+
return {
|
|
296
|
+
jsonrpc: "2.0",
|
|
297
|
+
id,
|
|
298
|
+
error: {
|
|
299
|
+
code,
|
|
300
|
+
message,
|
|
301
|
+
...(data === undefined ? {} : { data })
|
|
302
|
+
}
|
|
303
|
+
};
|
|
304
|
+
}
|
|
305
|
+
function toJsonRpcError(id, error) {
|
|
306
|
+
if (error instanceof JsonRpcProtocolError) {
|
|
307
|
+
return jsonRpcError(id, error.code, error.message);
|
|
308
|
+
}
|
|
309
|
+
return jsonRpcError(id, -32603, error instanceof Error ? error.message : String(error));
|
|
310
|
+
}
|
|
311
|
+
function readRequestId(message) {
|
|
312
|
+
const id = message.id;
|
|
313
|
+
return typeof id === "string" || typeof id === "number" ? id : null;
|
|
314
|
+
}
|
|
315
|
+
function isMcpToolName(value) {
|
|
316
|
+
return mcpToolNames.has(value);
|
|
317
|
+
}
|
|
318
|
+
function requireRecord(value, fieldName) {
|
|
319
|
+
if (!isRecord(value)) {
|
|
320
|
+
throw new UserFacingError(`${fieldName} must be an object.`);
|
|
321
|
+
}
|
|
322
|
+
return value;
|
|
323
|
+
}
|
|
324
|
+
function requireString(value, fieldName) {
|
|
325
|
+
if (typeof value !== "string" || value.trim() === "") {
|
|
326
|
+
throw new UserFacingError(`${fieldName} must be a non-empty string.`);
|
|
327
|
+
}
|
|
328
|
+
return value;
|
|
329
|
+
}
|
|
330
|
+
function requireRelativePath(value, fieldName) {
|
|
331
|
+
const pathValue = requireString(value, fieldName);
|
|
332
|
+
if (pathValue.includes("..") || path.isAbsolute(pathValue)) {
|
|
333
|
+
throw new UserFacingError(`${fieldName} must be repository-relative.`);
|
|
334
|
+
}
|
|
335
|
+
return pathValue;
|
|
336
|
+
}
|
|
337
|
+
function requireInteger(value, fieldName, minimum, maximum = Number.MAX_SAFE_INTEGER) {
|
|
338
|
+
if (typeof value !== "number" ||
|
|
339
|
+
!Number.isInteger(value) ||
|
|
340
|
+
value < minimum ||
|
|
341
|
+
value > maximum) {
|
|
342
|
+
throw new UserFacingError(`${fieldName} must be an integer between ${String(minimum)} and ${String(maximum)}.`);
|
|
343
|
+
}
|
|
344
|
+
return value;
|
|
345
|
+
}
|
|
346
|
+
function optionalLimit(value, fieldName, minimum, maximum) {
|
|
347
|
+
return value === undefined
|
|
348
|
+
? {}
|
|
349
|
+
: { [fieldName]: requireInteger(value, fieldName, minimum, maximum) };
|
|
350
|
+
}
|
|
351
|
+
function optionalString(value, fieldName) {
|
|
352
|
+
return value === undefined
|
|
353
|
+
? {}
|
|
354
|
+
: { [fieldName]: requireString(value, fieldName) };
|
|
355
|
+
}
|
|
356
|
+
function optionalBoolean(value, fieldName) {
|
|
357
|
+
if (value === undefined) {
|
|
358
|
+
return {};
|
|
359
|
+
}
|
|
360
|
+
if (typeof value !== "boolean") {
|
|
361
|
+
throw new UserFacingError(`${fieldName} must be a boolean.`);
|
|
362
|
+
}
|
|
363
|
+
return { [fieldName]: value };
|
|
364
|
+
}
|
|
365
|
+
function optionalPathPrefix(value) {
|
|
366
|
+
if (value === undefined) {
|
|
367
|
+
return {};
|
|
368
|
+
}
|
|
369
|
+
return { pathPrefix: requireRelativePath(value, "pathPrefix") };
|
|
370
|
+
}
|
|
371
|
+
function optionalKind(value) {
|
|
372
|
+
if (value === undefined) {
|
|
373
|
+
return {};
|
|
374
|
+
}
|
|
375
|
+
const kind = requireString(value, "kind");
|
|
376
|
+
if (!isSupportedSourceKind(kind)) {
|
|
377
|
+
throw new UserFacingError(`Unsupported source kind: ${kind}.`);
|
|
378
|
+
}
|
|
379
|
+
return { kind };
|
|
380
|
+
}
|
|
381
|
+
function isRecord(value) {
|
|
382
|
+
return typeof value === "object" && value !== null;
|
|
383
|
+
}
|
|
384
|
+
class JsonRpcProtocolError extends Error {
|
|
385
|
+
code;
|
|
386
|
+
constructor(code, message) {
|
|
387
|
+
super(message);
|
|
388
|
+
this.code = code;
|
|
389
|
+
this.name = "JsonRpcProtocolError";
|
|
390
|
+
}
|
|
391
|
+
}
|
|
392
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/mcp/index.ts"],"names":[],"mappings":"AAAA,OAAO,IAAI,MAAM,WAAW,CAAC;AAE7B,OAAO,EAAE,eAAe,EAAE,MAAM,cAAc,CAAC;AAC/C,OAAO,EAAE,qBAAqB,EAAE,MAAM,qBAAqB,CAAC;AAC5D,OAAO,EAAE,oBAAoB,EAAE,WAAW,EAAE,MAAM,uBAAuB,CAAC;AAY1E,MAAM,kBAAkB,GAAG,YAAY,CAAC;AACxC,MAAM,aAAa,GAAG,OAAO,CAAC;AAwD9B,MAAM,CAAC,MAAM,kBAAkB,GAAwB;IACrD;QACE,IAAI,EAAE,YAAY;QAClB,WAAW,EACT,yFAAyF;QAC3F,WAAW,EAAE,WAAW,CAAC,UAAU;QACnC,WAAW,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE;KACpC;IACD;QACE,IAAI,EAAE,iBAAiB;QACvB,WAAW,EACT,0FAA0F;QAC5F,WAAW,EAAE,WAAW,CAAC,eAAe;QACxC,WAAW,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE;KACpC;IACD;QACE,IAAI,EAAE,eAAe;QACrB,WAAW,EACT,2FAA2F;QAC7F,WAAW,EAAE,WAAW,CAAC,aAAa;QACtC,WAAW,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE;KACpC;IACD;QACE,IAAI,EAAE,UAAU;QAChB,WAAW,EACT,6FAA6F;QAC/F,WAAW,EAAE,WAAW,CAAC,QAAQ;QACjC,WAAW,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE;KACpC;IACD;QACE,IAAI,EAAE,UAAU;QAChB,WAAW,EACT,6EAA6E;QAC/E,WAAW,EAAE,WAAW,CAAC,QAAQ;QACjC,WAAW,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE;KACpC;IACD;QACE,IAAI,EAAE,eAAe;QACrB,WAAW,EACT,wFAAwF;QAC1F,WAAW,EAAE,WAAW,CAAC,aAAa;QACtC,WAAW,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE;KACpC;CACF,CAAC;AAEF,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,kBAAkB,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;AAE1E,MAAM,CAAC,KAAK,UAAU,YAAY,CAAC,OAAyB;IAC1D,MAAM,OAAO,GAAG,IAAI,oBAAoB,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;IAEvE,IAAI,CAAC;QACH,IAAI,KAAK,EAAE,MAAM,OAAO,IAAI,mBAAmB,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;YAC/D,IAAI,iBAAiB,CAAC,OAAO,CAAC,EAAE,CAAC;gBAC/B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;gBACrD,SAAS;YACX,CAAC;YAED,MAAM,SAAS,GAAG,MAAM,gBAAgB,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;YAE3D,KAAK,MAAM,QAAQ,IAAI,SAAS,EAAE,CAAC;gBACjC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;YACxD,CAAC;QACH,CAAC;IACH,CAAC;YAAS,CAAC;QACT,OAAO,CAAC,KAAK,EAAE,CAAC;IAClB,CAAC;AACH,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,gBAAgB,CACpC,OAAgB,EAChB,OAA6B;IAE7B,IAAI,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC3B,MAAM,SAAS,GAAG,MAAM,OAAO,CAAC,GAAG,CACjC,OAAO,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,MAAM,sBAAsB,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAC3E,CAAC;QAEF,OAAO,SAAS,CAAC,MAAM,CACrB,CAAC,QAAQ,EAA+B,EAAE,CAAC,QAAQ,KAAK,IAAI,CAC7D,CAAC;IACJ,CAAC;IAED,MAAM,QAAQ,GAAG,MAAM,sBAAsB,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;IAEhE,OAAO,QAAQ,KAAK,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC;AAC7C,CAAC;AAED,KAAK,UAAU,sBAAsB,CACnC,OAAgB,EAChB,OAA6B;IAE7B,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,OAAO,CAAC,OAAO,KAAK,KAAK,EAAE,CAAC;QACpD,OAAO,YAAY,CAAC,IAAI,EAAE,CAAC,KAAK,EAAE,2BAA2B,CAAC,CAAC;IACjE,CAAC;IAED,IAAI,OAAO,OAAO,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;QACvC,OAAO,YAAY,CAAC,aAAa,CAAC,OAAO,CAAC,EAAE,CAAC,KAAK,EAAE,iBAAiB,CAAC,CAAC;IACzE,CAAC;IAED,IAAI,CAAC,CAAC,IAAI,IAAI,OAAO,CAAC,EAAE,CAAC;QACvB,qBAAqB,CAAC;YACpB,OAAO,EAAE,KAAK;YACd,MAAM,EAAE,OAAO,CAAC,MAAM;YACtB,GAAG,CAAC,OAAO,CAAC,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC;SACpE,CAAC,CAAC;QACH,OAAO,IAAI,CAAC;IACd,CAAC;IAED,MAAM,EAAE,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC;IAElC,IAAI,EAAE,KAAK,IAAI,EAAE,CAAC;QAChB,OAAO,YAAY,CACjB,IAAI,EACJ,CAAC,KAAK,EACN,iDAAiD,CAClD,CAAC;IACJ,CAAC;IAED,IAAI,CAAC;QACH,OAAO,aAAa,CAClB,EAAE,EACF,MAAM,gBAAgB,CACpB;YACE,OAAO,EAAE,KAAK;YACd,EAAE;YACF,MAAM,EAAE,OAAO,CAAC,MAAM;YACtB,GAAG,CAAC,OAAO,CAAC,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC;SACpE,EACD,OAAO,CACR,CACF,CAAC;IACJ,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,OAAO,cAAc,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC;IACnC,CAAC;AACH,CAAC;AAED,SAAS,qBAAqB,CAAC,OAA4B;IACzD,QAAQ,OAAO,CAAC,MAAM,EAAE,CAAC;QACvB,KAAK,2BAA2B,CAAC;QACjC,KAAK,yBAAyB;YAC5B,OAAO;QACT;YACE,OAAO;IACX,CAAC;AACH,CAAC;AAED,KAAK,UAAU,gBAAgB,CAC7B,OAAuB,EACvB,OAA6B;IAE7B,QAAQ,OAAO,CAAC,MAAM,EAAE,CAAC;QACvB,KAAK,YAAY;YACf,OAAO,sBAAsB,EAAE,CAAC;QAClC,KAAK,MAAM;YACT,OAAO,EAAE,CAAC;QACZ,KAAK,YAAY;YACf,OAAO,EAAE,KAAK,EAAE,kBAAkB,EAAE,CAAC;QACvC,KAAK,YAAY;YACf,OAAO,MAAM,WAAW,CAAC,OAAO,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;QACpD;YACE,MAAM,IAAI,oBAAoB,CAC5B,CAAC,KAAK,EACN,uBAAuB,OAAO,CAAC,MAAM,EAAE,CACxC,CAAC;IACN,CAAC;AACH,CAAC;AAED,SAAS,sBAAsB;IAC7B,OAAO;QACL,eAAe,EAAE,kBAAkB;QACnC,YAAY,EAAE;YACZ,KAAK,EAAE;gBACL,WAAW,EAAE,KAAK;aACnB;SACF;QACD,UAAU,EAAE;YACV,IAAI,EAAE,QAAQ;YACd,OAAO,EAAE,aAAa;SACvB;QACD,YAAY,EACV,uJAAuJ;KAC1J,CAAC;AACJ,CAAC;AAED,KAAK,UAAU,WAAW,CACxB,OAA6B,EAC7B,MAAe;IAEf,MAAM,KAAK,GAAG,aAAa,CAAC,MAAM,EAAE,mBAAmB,CAAC,CAAC;IACzD,MAAM,IAAI,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IAE/C,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC;QACzB,MAAM,IAAI,oBAAoB,CAAC,CAAC,KAAK,EAAE,iBAAiB,IAAI,EAAE,CAAC,CAAC;IAClE,CAAC;IAED,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,MAAM,cAAc,CAAC,OAAO,EAAE,IAAI,EAAE,KAAK,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC;QAE1E,OAAO;YACL,OAAO,EAAE;gBACP;oBACE,IAAI,EAAE,MAAM;oBACZ,IAAI,EAAE,IAAI,CAAC,SAAS,CAClB;wBACE,MAAM;wBACN,KAAK,EAAE,uBAAuB,CAAC,OAAO,CAAC,KAAK,CAAC;qBAC9C,EACD,IAAI,EACJ,CAAC,CACF;iBACF;aACF;YACD,OAAO,EAAE,KAAK;SACf,CAAC;IACJ,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,OAAO;YACL,OAAO,EAAE;gBACP;oBACE,IAAI,EAAE,MAAM;oBACZ,IAAI,EAAE,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC;iBAC7D;aACF;YACD,OAAO,EAAE,IAAI;SACd,CAAC;IACJ,CAAC;AACH,CAAC;AAED,KAAK,UAAU,cAAc,CAC3B,OAA6B,EAC7B,IAAiB,EACjB,IAAa;IAEb,QAAQ,IAAI,EAAE,CAAC;QACb,KAAK,YAAY;YACf,OAAO,OAAO,CAAC,UAAU,CAAC,mBAAmB,CAAC,IAAI,CAAC,CAAC,CAAC;QACvD,KAAK,iBAAiB;YACpB,OAAO,MAAM,OAAO,CAAC,eAAe,CAAC,oBAAoB,CAAC,IAAI,CAAC,CAAC,CAAC;QACnE,KAAK,eAAe;YAClB,OAAO,OAAO,CAAC,aAAa,CAAC,kBAAkB,CAAC,IAAI,CAAC,CAAC,CAAC;QACzD,KAAK,UAAU;YACb,OAAO,MAAM,OAAO,CAAC,QAAQ,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC;QACzD,KAAK,UAAU;YACb,OAAO,OAAO,CAAC,QAAQ,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC;QACnD,KAAK,eAAe;YAClB,OAAO,OAAO,CAAC,aAAa,CAAC,sBAAsB,CAAC,IAAI,CAAC,CAAC,CAAC;IAC/D,CAAC;AACH,CAAC;AAED,SAAS,uBAAuB,CAAC,KAAqB;IACpD,OAAO;QACL,YAAY,EAAE,CAAC,GAAG,KAAK,CAAC,YAAY,CAAC;QACrC,YAAY,EAAE,KAAK,CAAC,YAAY;QAChC,YAAY,EAAE,KAAK,CAAC,YAAY;KACjC,CAAC;AACJ,CAAC;AAED,SAAS,mBAAmB,CAAC,KAAc;IACzC,MAAM,KAAK,GAAG,aAAa,CAAC,KAAK,EAAE,sBAAsB,CAAC,CAAC;IAE3D,OAAO;QACL,KAAK,EAAE,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,CAAC;QAC1C,GAAG,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,EAAE,CAAC,EAAE,EAAE,CAAC;QAC7C,GAAG,kBAAkB,CAAC,KAAK,CAAC,UAAU,CAAC;QACvC,GAAG,YAAY,CAAC,KAAK,CAAC,IAAI,CAAC;KAC5B,CAAC;AACJ,CAAC;AAED,SAAS,oBAAoB,CAAC,KAAc;IAC1C,MAAM,KAAK,GAAG,aAAa,CAAC,KAAK,EAAE,2BAA2B,CAAC,CAAC;IAEhE,OAAO;QACL,IAAI,EAAE,mBAAmB,CAAC,KAAK,CAAC,IAAI,EAAE,MAAM,CAAC;QAC7C,SAAS,EAAE,cAAc,CAAC,KAAK,CAAC,SAAS,EAAE,WAAW,EAAE,CAAC,CAAC;QAC1D,OAAO,EAAE,cAAc,CAAC,KAAK,CAAC,OAAO,EAAE,SAAS,EAAE,CAAC,CAAC;KACrD,CAAC;AACJ,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAc;IACxC,MAAM,KAAK,GAAG,aAAa,CAAC,KAAK,EAAE,yBAAyB,CAAC,CAAC;IAE9D,OAAO;QACL,IAAI,EAAE,mBAAmB,CAAC,KAAK,CAAC,IAAI,EAAE,MAAM,CAAC;QAC7C,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,CAAC,gBAAgB,CAAC;QAChE,GAAG,cAAc,CAAC,KAAK,CAAC,QAAQ,EAAE,UAAU,CAAC;KAC9C,CAAC;AACJ,CAAC;AAED,SAAS,iBAAiB,CAAC,KAAc;IACvC,MAAM,KAAK,GAAG,aAAa,CAAC,KAAK,EAAE,oBAAoB,CAAC,CAAC;IAEzD,OAAO;QACL,OAAO,EAAE,aAAa,CAAC,KAAK,CAAC,OAAO,EAAE,SAAS,CAAC;QAChD,GAAG,eAAe,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,CAAC;QACxC,GAAG,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,EAAE,CAAC,EAAE,GAAG,CAAC;QAC9C,GAAG,kBAAkB,CAAC,KAAK,CAAC,UAAU,CAAC;QACvC,GAAG,YAAY,CAAC,KAAK,CAAC,IAAI,CAAC;KAC5B,CAAC;AACJ,CAAC;AAED,SAAS,iBAAiB,CAAC,KAAc;IACvC,MAAM,KAAK,GAAG,aAAa,CAAC,KAAK,EAAE,oBAAoB,CAAC,CAAC;IAEzD,OAAO;QACL,GAAG,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,EAAE,CAAC,EAAE,GAAG,CAAC;QAC9C,GAAG,kBAAkB,CAAC,KAAK,CAAC,UAAU,CAAC;QACvC,GAAG,YAAY,CAAC,KAAK,CAAC,IAAI,CAAC;KAC5B,CAAC;AACJ,CAAC;AAED,SAAS,sBAAsB,CAAC,KAAc;IAC5C,MAAM,KAAK,GAAG,aAAa,CAAC,KAAK,EAAE,yBAAyB,CAAC,CAAC;IAE9D,OAAO;QACL,OAAO,EAAE,aAAa,CAAC,KAAK,CAAC,OAAO,EAAE,SAAS,CAAC;QAChD,GAAG,aAAa,CAAC,KAAK,CAAC,MAAM,EAAE,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;QAC/C,GAAG,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,EAAE,CAAC,EAAE,EAAE,CAAC;KAC9C,CAAC;AACJ,CAAC;AAED,KAAK,SAAS,CAAC,CAAC,mBAAmB,CACjC,KAAyC;IAEzC,IAAI,QAAQ,GAAG,EAAE,CAAC;IAElB,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,KAAK,EAAE,CAAC;QAChC,QAAQ,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;QAE7B,OAAO,IAAI,EAAE,CAAC;YACZ,MAAM,YAAY,GAAG,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAE5C,IAAI,YAAY,KAAK,CAAC,CAAC,EAAE,CAAC;gBACxB,MAAM;YACR,CAAC;YAED,MAAM,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC,IAAI,EAAE,CAAC;YACpD,QAAQ,GAAG,QAAQ,CAAC,KAAK,CAAC,YAAY,GAAG,CAAC,CAAC,CAAC;YAE5C,IAAI,IAAI,KAAK,EAAE,EAAE,CAAC;gBAChB,SAAS;YACX,CAAC;YAED,MAAM,gBAAgB,CAAC,IAAI,CAAC,CAAC;QAC/B,CAAC;IACH,CAAC;IAED,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,EAAE,CAAC;IAEjC,IAAI,QAAQ,KAAK,EAAE,EAAE,CAAC;QACpB,MAAM,gBAAgB,CAAC,QAAQ,CAAC,CAAC;IACnC,CAAC;AACH,CAAC;AAED,SAAS,gBAAgB,CAAC,IAAY;IACpC,IAAI,CAAC;QACH,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAY,CAAC;IACrC,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,OAAO,YAAY,CAAC,IAAI,EAAE,CAAC,KAAK,EAAE,cAAc,EAAE;YAChD,OAAO,EAAE,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC;SAChE,CAAC,CAAC;IACL,CAAC;AACH,CAAC;AAED,SAAS,iBAAiB,CAAC,KAAc;IACvC,OAAO,CACL,QAAQ,CAAC,KAAK,CAAC;QACf,KAAK,CAAC,OAAO,KAAK,KAAK;QACvB,IAAI,IAAI,KAAK;QACb,CAAC,QAAQ,IAAI,KAAK,IAAI,OAAO,IAAI,KAAK,CAAC;QACvC,CAAC,CAAC,QAAQ,IAAI,KAAK,CAAC,CACrB,CAAC;AACJ,CAAC;AAED,SAAS,aAAa,CAAC,EAAa,EAAE,MAAe;IACnD,OAAO;QACL,OAAO,EAAE,KAAK;QACd,EAAE;QACF,MAAM;KACP,CAAC;AACJ,CAAC;AAED,SAAS,YAAY,CACnB,EAAoB,EACpB,IAAY,EACZ,OAAe,EACf,IAAc;IAEd,OAAO;QACL,OAAO,EAAE,KAAK;QACd,EAAE;QACF,KAAK,EAAE;YACL,IAAI;YACJ,OAAO;YACP,GAAG,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;SACxC;KACF,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CAAC,EAAa,EAAE,KAAc;IACnD,IAAI,KAAK,YAAY,oBAAoB,EAAE,CAAC;QAC1C,OAAO,YAAY,CAAC,EAAE,EAAE,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC;IACrD,CAAC;IAED,OAAO,YAAY,CACjB,EAAE,EACF,CAAC,KAAK,EACN,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CACvD,CAAC;AACJ,CAAC;AAED,SAAS,aAAa,CAAC,OAAgC;IACrD,MAAM,EAAE,GAAG,OAAO,CAAC,EAAE,CAAC;IAEtB,OAAO,OAAO,EAAE,KAAK,QAAQ,IAAI,OAAO,EAAE,KAAK,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;AACtE,CAAC;AAED,SAAS,aAAa,CAAC,KAAa;IAClC,OAAO,YAAY,CAAC,GAAG,CAAC,KAAoB,CAAC,CAAC;AAChD,CAAC;AAED,SAAS,aAAa,CACpB,KAAc,EACd,SAAiB;IAEjB,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;QACrB,MAAM,IAAI,eAAe,CAAC,GAAG,SAAS,qBAAqB,CAAC,CAAC;IAC/D,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,aAAa,CAAC,KAAc,EAAE,SAAiB;IACtD,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;QACrD,MAAM,IAAI,eAAe,CAAC,GAAG,SAAS,8BAA8B,CAAC,CAAC;IACxE,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,mBAAmB,CAAC,KAAc,EAAE,SAAiB;IAC5D,MAAM,SAAS,GAAG,aAAa,CAAC,KAAK,EAAE,SAAS,CAAC,CAAC;IAElD,IAAI,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,CAAC;QAC3D,MAAM,IAAI,eAAe,CAAC,GAAG,SAAS,+BAA+B,CAAC,CAAC;IACzE,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,cAAc,CACrB,KAAc,EACd,SAAiB,EACjB,OAAe,EACf,OAAO,GAAG,MAAM,CAAC,gBAAgB;IAEjC,IACE,OAAO,KAAK,KAAK,QAAQ;QACzB,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC;QACxB,KAAK,GAAG,OAAO;QACf,KAAK,GAAG,OAAO,EACf,CAAC;QACD,MAAM,IAAI,eAAe,CACvB,GAAG,SAAS,+BAA+B,MAAM,CAAC,OAAO,CAAC,QAAQ,MAAM,CAAC,OAAO,CAAC,GAAG,CACrF,CAAC;IACJ,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,aAAa,CACpB,KAAc,EACd,SAAiB,EACjB,OAAe,EACf,OAAe;IAEf,OAAO,KAAK,KAAK,SAAS;QACxB,CAAC,CAAC,EAAE;QACJ,CAAC,CAAC,EAAE,CAAC,SAAS,CAAC,EAAE,cAAc,CAAC,KAAK,EAAE,SAAS,EAAE,OAAO,EAAE,OAAO,CAAC,EAAE,CAAC;AAC1E,CAAC;AAED,SAAS,cAAc,CACrB,KAAc,EACd,SAAiB;IAEjB,OAAO,KAAK,KAAK,SAAS;QACxB,CAAC,CAAC,EAAE;QACJ,CAAC,CAAC,EAAE,CAAC,SAAS,CAAC,EAAE,aAAa,CAAC,KAAK,EAAE,SAAS,CAAC,EAAE,CAAC;AACvD,CAAC;AAED,SAAS,eAAe,CACtB,KAAc,EACd,SAAiB;IAEjB,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,IAAI,OAAO,KAAK,KAAK,SAAS,EAAE,CAAC;QAC/B,MAAM,IAAI,eAAe,CAAC,GAAG,SAAS,qBAAqB,CAAC,CAAC;IAC/D,CAAC;IAED,OAAO,EAAE,CAAC,SAAS,CAAC,EAAE,KAAK,EAAE,CAAC;AAChC,CAAC;AAED,SAAS,kBAAkB,CACzB,KAAc;IAEd,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,OAAO,EAAE,UAAU,EAAE,mBAAmB,CAAC,KAAK,EAAE,YAAY,CAAC,EAAE,CAAC;AAClE,CAAC;AAED,SAAS,YAAY,CAAC,KAAc;IAClC,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,IAAI,GAAG,aAAa,CAAC,KAAK,EAAE,MAAM,CAAe,CAAC;IAExD,IAAI,CAAC,qBAAqB,CAAC,IAAI,CAAC,EAAE,CAAC;QACjC,MAAM,IAAI,eAAe,CAAC,4BAA4B,IAAI,GAAG,CAAC,CAAC;IACjE,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,CAAC;AAClB,CAAC;AAED,SAAS,QAAQ,CAAC,KAAc;IAC9B,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC;AACrD,CAAC;AAED,MAAM,oBAAqB,SAAQ,KAAK;IAEpB;IADlB,YACkB,IAAY,EAC5B,OAAe;QAEf,KAAK,CAAC,OAAO,CAAC,CAAC;QAHC,SAAI,GAAJ,IAAI,CAAQ;QAI5B,IAAI,CAAC,IAAI,GAAG,sBAAsB,CAAC;IACrC,CAAC;CACF"}
|
|
@@ -0,0 +1,4 @@
|
|
|
1
|
+
import type { ChunkDraft, ChunkingOptions, ParsedDocument } from "./types.js";
|
|
2
|
+
export declare const defaultChunkingOptions: ChunkingOptions;
|
|
3
|
+
export declare function chunkParsedDocument(document: ParsedDocument, options?: ChunkingOptions): ChunkDraft[];
|
|
4
|
+
//# sourceMappingURL=chunking.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"chunking.d.ts","sourceRoot":"","sources":["../../src/parsers/chunking.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EACV,UAAU,EACV,eAAe,EAEf,cAAc,EACf,MAAM,YAAY,CAAC;AAEpB,eAAO,MAAM,sBAAsB,EAAE,eAIpC,CAAC;AAEF,wBAAgB,mBAAmB,CACjC,QAAQ,EAAE,cAAc,EACxB,OAAO,GAAE,eAAwC,GAChD,UAAU,EAAE,CAyBd"}
|
|
@@ -0,0 +1,112 @@
|
|
|
1
|
+
import { estimateTokens, normalizeText } from "./normalize.js";
|
|
2
|
+
export const defaultChunkingOptions = {
|
|
3
|
+
targetTokens: 350,
|
|
4
|
+
overlapTokens: 50,
|
|
5
|
+
hardMaxTokens: 700
|
|
6
|
+
};
|
|
7
|
+
export function chunkParsedDocument(document, options = defaultChunkingOptions) {
|
|
8
|
+
const chunks = [];
|
|
9
|
+
let currentBlocks = [];
|
|
10
|
+
for (const block of document.blocks) {
|
|
11
|
+
const blockParts = splitOversizedBlock(block, options.hardMaxTokens);
|
|
12
|
+
for (const blockPart of blockParts) {
|
|
13
|
+
const nextTokenCount = estimateBlocks(currentBlocks) + estimateTokens(blockPart.text);
|
|
14
|
+
if (currentBlocks.length > 0 && nextTokenCount > options.targetTokens) {
|
|
15
|
+
chunks.push(createChunk(chunks.length, currentBlocks));
|
|
16
|
+
currentBlocks = trailingOverlap(currentBlocks, options.overlapTokens);
|
|
17
|
+
}
|
|
18
|
+
currentBlocks.push(blockPart);
|
|
19
|
+
}
|
|
20
|
+
}
|
|
21
|
+
if (currentBlocks.length > 0) {
|
|
22
|
+
chunks.push(createChunk(chunks.length, currentBlocks));
|
|
23
|
+
}
|
|
24
|
+
return chunks;
|
|
25
|
+
}
|
|
26
|
+
function splitOversizedBlock(block, hardMaxTokens) {
|
|
27
|
+
if (estimateTokens(block.text) <= hardMaxTokens) {
|
|
28
|
+
return [block];
|
|
29
|
+
}
|
|
30
|
+
const lines = block.text.split("\n");
|
|
31
|
+
const lineCount = Math.max(lines.length, 1);
|
|
32
|
+
const parts = [];
|
|
33
|
+
let currentLines = [];
|
|
34
|
+
let currentStartLine = block.startLine;
|
|
35
|
+
for (const [index, line] of lines.entries()) {
|
|
36
|
+
const projected = [...currentLines, line].join("\n");
|
|
37
|
+
if (currentLines.length > 0 && estimateTokens(projected) > hardMaxTokens) {
|
|
38
|
+
parts.push(createBlockPart(block, currentLines, currentStartLine));
|
|
39
|
+
currentLines = [];
|
|
40
|
+
currentStartLine = block.startLine + index;
|
|
41
|
+
}
|
|
42
|
+
currentLines.push(line);
|
|
43
|
+
}
|
|
44
|
+
if (currentLines.length > 0) {
|
|
45
|
+
parts.push(createBlockPart(block, currentLines, currentStartLine));
|
|
46
|
+
}
|
|
47
|
+
return parts.length === 0
|
|
48
|
+
? [
|
|
49
|
+
{
|
|
50
|
+
...block,
|
|
51
|
+
endLine: block.startLine + lineCount - 1
|
|
52
|
+
}
|
|
53
|
+
]
|
|
54
|
+
: parts;
|
|
55
|
+
}
|
|
56
|
+
function createBlockPart(block, lines, startLine) {
|
|
57
|
+
const text = lines.join("\n");
|
|
58
|
+
return {
|
|
59
|
+
...block,
|
|
60
|
+
text,
|
|
61
|
+
normalizedText: normalizeText(text),
|
|
62
|
+
startLine,
|
|
63
|
+
endLine: startLine + lines.length - 1
|
|
64
|
+
};
|
|
65
|
+
}
|
|
66
|
+
function trailingOverlap(blocks, overlapTokens) {
|
|
67
|
+
const overlap = [];
|
|
68
|
+
let tokenCount = 0;
|
|
69
|
+
for (let index = blocks.length - 1; index >= 0; index -= 1) {
|
|
70
|
+
const block = blocks[index];
|
|
71
|
+
if (block === undefined) {
|
|
72
|
+
continue;
|
|
73
|
+
}
|
|
74
|
+
const blockTokens = estimateTokens(block.text);
|
|
75
|
+
if (overlap.length > 0 && tokenCount + blockTokens > overlapTokens) {
|
|
76
|
+
break;
|
|
77
|
+
}
|
|
78
|
+
overlap.unshift(block);
|
|
79
|
+
tokenCount += blockTokens;
|
|
80
|
+
}
|
|
81
|
+
return overlap;
|
|
82
|
+
}
|
|
83
|
+
function createChunk(ordinal, blocks) {
|
|
84
|
+
const text = blocks.map((block) => block.text).join("\n\n");
|
|
85
|
+
const headingPath = [...(blocks.at(-1)?.headingPath ?? [])];
|
|
86
|
+
const metadata = mergeBlockMetadata(blocks);
|
|
87
|
+
return {
|
|
88
|
+
ordinal,
|
|
89
|
+
text,
|
|
90
|
+
normalizedText: normalizeText(text),
|
|
91
|
+
headingPath,
|
|
92
|
+
startLine: blocks[0]?.startLine ?? 1,
|
|
93
|
+
endLine: blocks.at(-1)?.endLine ?? blocks[0]?.endLine ?? 1,
|
|
94
|
+
tokenEstimate: estimateTokens(text),
|
|
95
|
+
...(metadata === undefined ? {} : { metadata })
|
|
96
|
+
};
|
|
97
|
+
}
|
|
98
|
+
function estimateBlocks(blocks) {
|
|
99
|
+
return blocks.reduce((total, block) => total + estimateTokens(block.text), 0);
|
|
100
|
+
}
|
|
101
|
+
function mergeBlockMetadata(blocks) {
|
|
102
|
+
const metadata = {};
|
|
103
|
+
for (const block of blocks) {
|
|
104
|
+
for (const [key, value] of Object.entries(block.metadata ?? {})) {
|
|
105
|
+
if (metadata[key] === undefined) {
|
|
106
|
+
metadata[key] = value;
|
|
107
|
+
}
|
|
108
|
+
}
|
|
109
|
+
}
|
|
110
|
+
return Object.keys(metadata).length === 0 ? undefined : metadata;
|
|
111
|
+
}
|
|
112
|
+
//# sourceMappingURL=chunking.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"chunking.js","sourceRoot":"","sources":["../../src/parsers/chunking.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAQ/D,MAAM,CAAC,MAAM,sBAAsB,GAAoB;IACrD,YAAY,EAAE,GAAG;IACjB,aAAa,EAAE,EAAE;IACjB,aAAa,EAAE,GAAG;CACnB,CAAC;AAEF,MAAM,UAAU,mBAAmB,CACjC,QAAwB,EACxB,UAA2B,sBAAsB;IAEjD,MAAM,MAAM,GAAiB,EAAE,CAAC;IAChC,IAAI,aAAa,GAAkB,EAAE,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,QAAQ,CAAC,MAAM,EAAE,CAAC;QACpC,MAAM,UAAU,GAAG,mBAAmB,CAAC,KAAK,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC;QAErE,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;YACnC,MAAM,cAAc,GAClB,cAAc,CAAC,aAAa,CAAC,GAAG,cAAc,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;YAEjE,IAAI,aAAa,CAAC,MAAM,GAAG,CAAC,IAAI,cAAc,GAAG,OAAO,CAAC,YAAY,EAAE,CAAC;gBACtE,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,MAAM,EAAE,aAAa,CAAC,CAAC,CAAC;gBACvD,aAAa,GAAG,eAAe,CAAC,aAAa,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC;YACxE,CAAC;YAED,aAAa,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAChC,CAAC;IACH,CAAC;IAED,IAAI,aAAa,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC7B,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,MAAM,EAAE,aAAa,CAAC,CAAC,CAAC;IACzD,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,mBAAmB,CAC1B,KAAkB,EAClB,aAAqB;IAErB,IAAI,cAAc,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,aAAa,EAAE,CAAC;QAChD,OAAO,CAAC,KAAK,CAAC,CAAC;IACjB,CAAC;IAED,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACrC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;IAC5C,MAAM,KAAK,GAAkB,EAAE,CAAC;IAChC,IAAI,YAAY,GAAa,EAAE,CAAC;IAChC,IAAI,gBAAgB,GAAG,KAAK,CAAC,SAAS,CAAC;IAEvC,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,KAAK,CAAC,OAAO,EAAE,EAAE,CAAC;QAC5C,MAAM,SAAS,GAAG,CAAC,GAAG,YAAY,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAErD,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,IAAI,cAAc,CAAC,SAAS,CAAC,GAAG,aAAa,EAAE,CAAC;YACzE,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,KAAK,EAAE,YAAY,EAAE,gBAAgB,CAAC,CAAC,CAAC;YACnE,YAAY,GAAG,EAAE,CAAC;YAClB,gBAAgB,GAAG,KAAK,CAAC,SAAS,GAAG,KAAK,CAAC;QAC7C,CAAC;QAED,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC1B,CAAC;IAED,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,KAAK,EAAE,YAAY,EAAE,gBAAgB,CAAC,CAAC,CAAC;IACrE,CAAC;IAED,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC;QACvB,CAAC,CAAC;YACE;gBACE,GAAG,KAAK;gBACR,OAAO,EAAE,KAAK,CAAC,SAAS,GAAG,SAAS,GAAG,CAAC;aACzC;SACF;QACH,CAAC,CAAC,KAAK,CAAC;AACZ,CAAC;AAED,SAAS,eAAe,CACtB,KAAkB,EAClB,KAAe,EACf,SAAiB;IAEjB,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAE9B,OAAO;QACL,GAAG,KAAK;QACR,IAAI;QACJ,cAAc,EAAE,aAAa,CAAC,IAAI,CAAC;QACnC,SAAS;QACT,OAAO,EAAE,SAAS,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC;KACtC,CAAC;AACJ,CAAC;AAED,SAAS,eAAe,CACtB,MAAqB,EACrB,aAAqB;IAErB,MAAM,OAAO,GAAkB,EAAE,CAAC;IAClC,IAAI,UAAU,GAAG,CAAC,CAAC;IAEnB,KAAK,IAAI,KAAK,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,KAAK,IAAI,CAAC,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QAC3D,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;QAE5B,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,SAAS;QACX,CAAC;QAED,MAAM,WAAW,GAAG,cAAc,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAE/C,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,UAAU,GAAG,WAAW,GAAG,aAAa,EAAE,CAAC;YACnE,MAAM;QACR,CAAC;QAED,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;QACvB,UAAU,IAAI,WAAW,CAAC;IAC5B,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,WAAW,CAAC,OAAe,EAAE,MAAqB;IACzD,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAC5D,MAAM,WAAW,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,WAAW,IAAI,EAAE,CAAC,CAAC,CAAC;IAC5D,MAAM,QAAQ,GAAG,kBAAkB,CAAC,MAAM,CAAC,CAAC;IAE5C,OAAO;QACL,OAAO;QACP,IAAI;QACJ,cAAc,EAAE,aAAa,CAAC,IAAI,CAAC;QACnC,WAAW;QACX,SAAS,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,SAAS,IAAI,CAAC;QACpC,OAAO,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,OAAO,IAAI,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,IAAI,CAAC;QAC1D,aAAa,EAAE,cAAc,CAAC,IAAI,CAAC;QACnC,GAAG,CAAC,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,CAAC;KAChD,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CAAC,MAAqB;IAC3C,OAAO,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,GAAG,cAAc,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;AAChF,CAAC;AAED,SAAS,kBAAkB,CACzB,MAA8B;IAE9B,MAAM,QAAQ,GAA4B,EAAE,CAAC;IAE7C,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,QAAQ,IAAI,EAAE,CAAC,EAAE,CAAC;YAChE,IAAI,QAAQ,CAAC,GAAG,CAAC,KAAK,SAAS,EAAE,CAAC;gBAChC,QAAQ,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC;YACxB,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC;AACnE,CAAC"}
|
|
@@ -0,0 +1,8 @@
|
|
|
1
|
+
export { chunkParsedDocument, defaultChunkingOptions } from "./chunking.js";
|
|
2
|
+
export { estimateTokens, normalizeText } from "./normalize.js";
|
|
3
|
+
export { createPdfParserOptionsHash, PDF2LLM_PACKAGE_VERSION, PDF2LLM_PARSER_NAME, PDF2LLM_PARSER_VERSION, parsePdfFile } from "./pdf2llm.js";
|
|
4
|
+
export { TEXT_PARSER_VERSION, parseSourceFile, parseSourceText, parserNameForKind, parserVersionForKind } from "./parser.js";
|
|
5
|
+
export type { ChunkDraft, ChunkingOptions, ParsedBlock, ParsedDocument, ParserArtifactDraft, ParserWarning } from "./types.js";
|
|
6
|
+
export type { ParsePdfFileOptions, Pdf2LlmConverter } from "./pdf2llm.js";
|
|
7
|
+
export type { ParseSourceFileOptions } from "./parser.js";
|
|
8
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/parsers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,sBAAsB,EAAE,MAAM,eAAe,CAAC;AAC5E,OAAO,EAAE,cAAc,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/D,OAAO,EACL,0BAA0B,EAC1B,uBAAuB,EACvB,mBAAmB,EACnB,sBAAsB,EACtB,YAAY,EACb,MAAM,cAAc,CAAC;AACtB,OAAO,EACL,mBAAmB,EACnB,eAAe,EACf,eAAe,EACf,iBAAiB,EACjB,oBAAoB,EACrB,MAAM,aAAa,CAAC;AACrB,YAAY,EACV,UAAU,EACV,eAAe,EACf,WAAW,EACX,cAAc,EACd,mBAAmB,EACnB,aAAa,EACd,MAAM,YAAY,CAAC;AACpB,YAAY,EAAE,mBAAmB,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAC;AAC1E,YAAY,EAAE,sBAAsB,EAAE,MAAM,aAAa,CAAC"}
|
|
@@ -0,0 +1,5 @@
|
|
|
1
|
+
export { chunkParsedDocument, defaultChunkingOptions } from "./chunking.js";
|
|
2
|
+
export { estimateTokens, normalizeText } from "./normalize.js";
|
|
3
|
+
export { createPdfParserOptionsHash, PDF2LLM_PACKAGE_VERSION, PDF2LLM_PARSER_NAME, PDF2LLM_PARSER_VERSION, parsePdfFile } from "./pdf2llm.js";
|
|
4
|
+
export { TEXT_PARSER_VERSION, parseSourceFile, parseSourceText, parserNameForKind, parserVersionForKind } from "./parser.js";
|
|
5
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/parsers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,sBAAsB,EAAE,MAAM,eAAe,CAAC;AAC5E,OAAO,EAAE,cAAc,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/D,OAAO,EACL,0BAA0B,EAC1B,uBAAuB,EACvB,mBAAmB,EACnB,sBAAsB,EACtB,YAAY,EACb,MAAM,cAAc,CAAC;AACtB,OAAO,EACL,mBAAmB,EACnB,eAAe,EACf,eAAe,EACf,iBAAiB,EACjB,oBAAoB,EACrB,MAAM,aAAa,CAAC"}
|
|
@@ -0,0 +1,13 @@
|
|
|
1
|
+
import type { ParsedBlock, ParserWarning } from "./types.js";
|
|
2
|
+
export interface LineParseResult {
|
|
3
|
+
blocks: ParsedBlock[];
|
|
4
|
+
warnings: ParserWarning[];
|
|
5
|
+
}
|
|
6
|
+
export declare function parseMarkdownLikeLines(sourceText: string, options: {
|
|
7
|
+
mdx: boolean;
|
|
8
|
+
}): LineParseResult;
|
|
9
|
+
export declare function parsePlainTextLines(sourceText: string): LineParseResult;
|
|
10
|
+
export declare function parseRstLines(sourceText: string): LineParseResult;
|
|
11
|
+
export declare function parseAsciiDocLines(sourceText: string): LineParseResult;
|
|
12
|
+
export declare function parseHtmlLines(sourceText: string): LineParseResult;
|
|
13
|
+
//# sourceMappingURL=line-parser.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"line-parser.d.ts","sourceRoot":"","sources":["../../src/parsers/line-parser.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,WAAW,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAE7D,MAAM,WAAW,eAAe;IAC9B,MAAM,EAAE,WAAW,EAAE,CAAC;IACtB,QAAQ,EAAE,aAAa,EAAE,CAAC;CAC3B;AAED,wBAAgB,sBAAsB,CACpC,UAAU,EAAE,MAAM,EAClB,OAAO,EAAE;IAAE,GAAG,EAAE,OAAO,CAAA;CAAE,GACxB,eAAe,CAsFjB;AAED,wBAAgB,mBAAmB,CAAC,UAAU,EAAE,MAAM,GAAG,eAAe,CA4BvE;AAED,wBAAgB,aAAa,CAAC,UAAU,EAAE,MAAM,GAAG,eAAe,CAuDjE;AAED,wBAAgB,kBAAkB,CAAC,UAAU,EAAE,MAAM,GAAG,eAAe,CAmDtE;AAED,wBAAgB,cAAc,CAAC,UAAU,EAAE,MAAM,GAAG,eAAe,CAsDlE"}
|