@chatman-media/kb 1.3.0 → 1.3.1

package/dist/index.js

@@ -44647,7 +44647,7 @@ var NO_CONTEXT_MARKER = "__NO_CONTEXT__";
 var replaceEmDash = {
   name: "replace-em-dash",
   description: "U+2014 \xAB\u2014\xBB \u2192 \xAB-\xBB (\u0441 \u043D\u043E\u0440\u043C\u0430\u043B\u0438\u0437\u0430\u0446\u0438\u0435\u0439 \u043F\u0440\u043E\u0431\u0435\u043B\u043E\u0432)",
-  apply: (s) => s.replace(/\s*\u2014\s*/g, " - ").replace(/ {2,}/g, " ")
+  apply: (s) => s.replace(/\s{0,200}\u2014\s{0,200}/g, " - ").replace(/ {2,}/g, " ")
 };
 var replaceEnDash = {
   name: "replace-en-dash",
@@ -59649,7 +59649,7 @@ Schema:
 ${schemaStr}`;
 }
 function parseStructuredOutput(raw, schema) {
-  const cleaned = raw.trim().replace(/^```(?:json)?\s*/i, "").replace(/\s*```$/, "").trim();
+  const cleaned = raw.trim().replace(/^```(?:json)?\s*/i, "").replace(/\s{0,1000}```$/, "").trim();
   let parsed;
   try {
     parsed = JSON.parse(cleaned);
@@ -60296,7 +60296,7 @@ class OllamaChatClient {
   disableThinking;
   timeoutMs;
   constructor(opts) {
-    this.host = opts.host.replace(/\/+$/, "");
+    this.host = opts.host.replace(/\/{1,512}$/, "");
     this.model = opts.model;
     this.fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
     this.disableThinking = opts.disableThinking ?? true;
@@ -60427,7 +60427,7 @@ class OllamaEmbeddingClient {
   fetchImpl;
   timeoutMs;
   constructor(opts) {
-    this.host = opts.host.replace(/\/+$/, "");
+    this.host = opts.host.replace(/\/{1,512}$/, "");
     this.model = opts.model;
     this.dim = opts.dim;
     this.fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
@@ -60478,7 +60478,7 @@ class OpenAIChatClient {
     if (!opts.apiKey)
       throw new Error("OpenAIChatClient: apiKey required");
     this.apiKey = opts.apiKey;
-    this.baseUrl = opts.baseUrl.replace(/\/+$/, "");
+    this.baseUrl = opts.baseUrl.replace(/\/{1,512}$/, "");
     this.model = opts.model;
     this.timeoutMs = opts.timeoutMs ?? 60000;
     this.fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
@@ -60624,7 +60624,7 @@ class OpenAIEmbeddingClient {
     if (!opts.apiKey)
       throw new Error("OpenAIEmbeddingClient: apiKey required");
     this.apiKey = opts.apiKey;
-    this.baseUrl = opts.baseUrl.replace(/\/+$/, "");
+    this.baseUrl = opts.baseUrl.replace(/\/{1,512}$/, "");
     this.model = opts.model;
     this.dim = opts.dim;
     this.timeoutMs = opts.timeoutMs ?? 60000;
@@ -60701,7 +60701,7 @@ class OpenRouterChatClient {
       throw new Error("OpenRouterChatClient: model required");
     }
     this.apiKey = opts.apiKey;
-    this.baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL).replace(/\/+$/, "");
+    this.baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL).replace(/\/{1,512}$/, "");
     this.model = opts.model;
     this.siteUrl = opts.siteUrl;
     this.appName = opts.appName;
@@ -61111,7 +61111,7 @@ function chunkBySections(text, opts = {}) {
     bodyLines = [];
   };
   for (const line of lines) {
-    const m = line.match(/^(#{1,6})\s+(.+)$/);
+    const m = line.match(/^(#{1,6})\s+(\S.*)$/);
     if (m) {
       flush();
       currentHeading = m[2] ?? null;
@@ -61719,7 +61719,10 @@ function* walk(dir) {
 function stripNonContent(raw) {
   let s = raw;
   s = s.replace(/^---\r?\n[\s\S]*?\r?\n---\r?\n?/, "");
-  s = s.replace(/<!--[\s\S]*?-->/g, "");
+  for (let prev = "";prev !== s; ) {
+    prev = s;
+    s = s.replace(/<!--[\s\S]{0,10000}?-->/g, "");
+  }
   s = s.replace(/\n{3,}/g, `
 
 `).trim();
@@ -61806,7 +61809,7 @@ class CohereReranker {
       throw new Error("CohereReranker: apiKey required");
     this.apiKey = opts.apiKey;
     this.model = opts.model ?? "rerank-v3.5";
-    this.baseUrl = (opts.baseUrl ?? "https://api.cohere.com/v2").replace(/\/+$/, "");
+    this.baseUrl = (opts.baseUrl ?? "https://api.cohere.com/v2").replace(/\/{1,512}$/, "");
     this.timeoutMs = opts.timeoutMs ?? 30000;
     this.fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
   }
@@ -61852,7 +61855,7 @@ class JinaReranker {
       throw new Error("JinaReranker: apiKey required");
     this.apiKey = opts.apiKey;
     this.model = opts.model ?? "jina-reranker-v2-base-multilingual";
-    this.baseUrl = (opts.baseUrl ?? "https://api.jina.ai/v1").replace(/\/+$/, "");
+    this.baseUrl = (opts.baseUrl ?? "https://api.jina.ai/v1").replace(/\/{1,512}$/, "");
     this.timeoutMs = opts.timeoutMs ?? 30000;
     this.fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
   }
@@ -62424,7 +62427,7 @@ async function classifyPhoto(opts) {
   if (!opts.apiKey || opts.apiKey.trim().length === 0) {
     throw new Error("classifyPhoto: apiKey required");
   }
-  const baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL2).replace(/\/+$/, "");
+  const baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL2).replace(/\/{1,512}$/, "");
   const fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
   const mime = opts.mimeType?.trim() ? opts.mimeType : "image/jpeg";
   const base643 = Buffer.from(opts.bytes).toString("base64");
@@ -62519,7 +62522,7 @@ async function extractPassportIdentity(opts) {
   if (!opts.apiKey || opts.apiKey.trim().length === 0) {
     throw new Error("extractPassportIdentity: apiKey required");
   }
-  const baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL2).replace(/\/+$/, "");
+  const baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL2).replace(/\/{1,512}$/, "");
   const fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
   const mime = opts.mimeType?.trim() ? opts.mimeType : "image/jpeg";
   const base643 = Buffer.from(opts.bytes).toString("base64");

package/dist/ingest.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"ingest.d.ts","sourceRoot":"","sources":["../src/ingest.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,KAAK,YAAY,EAAa,MAAM,YAAY,CAAC;AAC1D,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,2BAA2B,CAAC;AAEjE,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,YAAY,CAAC;AAI3C,MAAM,WAAW,UAAU;IACzB,EAAE,EAAE,QAAQ,CAAC;IACb,QAAQ,EAAE,eAAe,CAAC;IAC1B,KAAK,CAAC,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC;IAC9B,KAAK,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB;;sEAEkE;IAClE,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB;iFAC6E;IAC7E,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,gBAAgB;IAC/B,MAAM,EAAE,MAAM,CAAC;IACf,UAAU,EAAE,MAAM,CAAC;IACnB,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,OAAO,CAAC;CAClB;AAED,wBAAsB,UAAU,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,UAAU,GAAG,OAAO,CAAC,gBAAgB,CAAC,CA+C1F;AAED;;;;GAIG;AACH,wBAAsB,UAAU,CAC9B,KAAK,EAAE;IAAE,KAAK,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,EACtC,IAAI,EAAE,UAAU,GACf,OAAO,CAAC,gBAAgB,CAAC,CA4C3B;AAED,MAAM,WAAW,sBAAsB;IACrC,SAAS,EAAE,MAAM,CAAC;IAClB,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,wBAAsB,eAAe,CACnC,GAAG,EAAE,MAAM,EACX,IAAI,EAAE,UAAU,GACf,OAAO,CAAC,sBAAsB,CAAC,CAejC;AAED;;;GAGG;AACH,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAK7E;AAaD;;;GAGG;AACH,wBAAgB,eAAe,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,CAMnD"}
+{"version":3,"file":"ingest.d.ts","sourceRoot":"","sources":["../src/ingest.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,KAAK,YAAY,EAAa,MAAM,YAAY,CAAC;AAC1D,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,2BAA2B,CAAC;AAEjE,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,YAAY,CAAC;AAI3C,MAAM,WAAW,UAAU;IACzB,EAAE,EAAE,QAAQ,CAAC;IACb,QAAQ,EAAE,eAAe,CAAC;IAC1B,KAAK,CAAC,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC;IAC9B,KAAK,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB;;sEAEkE;IAClE,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB;iFAC6E;IAC7E,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,gBAAgB;IAC/B,MAAM,EAAE,MAAM,CAAC;IACf,UAAU,EAAE,MAAM,CAAC;IACnB,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,OAAO,CAAC;CAClB;AAED,wBAAsB,UAAU,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,UAAU,GAAG,OAAO,CAAC,gBAAgB,CAAC,CA+C1F;AAED;;;;GAIG;AACH,wBAAsB,UAAU,CAC9B,KAAK,EAAE;IAAE,KAAK,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,EACtC,IAAI,EAAE,UAAU,GACf,OAAO,CAAC,gBAAgB,CAAC,CA4C3B;AAED,MAAM,WAAW,sBAAsB;IACrC,SAAS,EAAE,MAAM,CAAC;IAClB,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,wBAAsB,eAAe,CACnC,GAAG,EAAE,MAAM,EACX,IAAI,EAAE,UAAU,GACf,OAAO,CAAC,sBAAsB,CAAC,CAejC;AAED;;;GAGG;AACH,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAK7E;AAaD;;;GAGG;AACH,wBAAgB,eAAe,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,CAWnD"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@chatman-media/kb",
-  "version": "1.3.0",
+  "version": "1.3.1",
   "description": "Tenant-scoped Knowledge Base: hybrid retrieval (pgvector + BM25), ingest, answer pipeline, persona/skill composition. LLM I/O живёт в @chatman-media/llm-router.",
   "type": "module",
   "main": "./dist/index.js",
@@ -56,7 +56,7 @@
     "url": "https://github.com/chatman-media/lead-engine/issues"
   },
   "dependencies": {
-    "@chatman-media/llm-router": "1.0.0",
+    "@chatman-media/llm-router": "1.1.0",
     "unpdf": "^1.6.2",
     "zod": "^4.4.1"
   },

package/src/chunk.ts

@@ -137,7 +137,7 @@ export function chunkBySections(text: string, opts: Partial<ChunkOptions> = {}):
   };
 
   for (const line of lines) {
-    const m = line.match(/^(#{1,6})\s+(.+)$/);
+    const m = line.match(/^(#{1,6})\s+(\S.*)$/);
     if (m) {
       flush();
       currentHeading = m[2] ?? null;

package/src/ingest.ts

@@ -187,7 +187,12 @@ function* walk(dir: string): Generator<string> {
 export function stripNonContent(raw: string): string {
   let s = raw;
   s = s.replace(/^---\r?\n[\s\S]*?\r?\n---\r?\n?/, "");
-  s = s.replace(/<!--[\s\S]*?-->/g, "");
+  // Strip HTML comments. Bounded quantifier avoids polynomial backtracking;
+  // loop until stable so a leftover `<!--` can't survive a single pass.
+  for (let prev = ""; prev !== s; ) {
+    prev = s;
+    s = s.replace(/<!--[\s\S]{0,10000}?-->/g, "");
+  }
   s = s.replace(/\n{3,}/g, "\n\n").trim();
   return `${s}\n`;
 }

package/src/reranker.ts

@@ -49,7 +49,7 @@ export class CohereReranker implements Reranker {
     if (!opts.apiKey) throw new Error("CohereReranker: apiKey required");
     this.apiKey = opts.apiKey;
     this.model = opts.model ?? "rerank-v3.5";
-    this.baseUrl = (opts.baseUrl ?? "https://api.cohere.com/v2").replace(/\/+$/, "");
+    this.baseUrl = (opts.baseUrl ?? "https://api.cohere.com/v2").replace(/\/{1,512}$/, "");
     this.timeoutMs = opts.timeoutMs ?? 30_000;
     this.fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
   }
@@ -128,7 +128,7 @@ export class JinaReranker implements Reranker {
     if (!opts.apiKey) throw new Error("JinaReranker: apiKey required");
     this.apiKey = opts.apiKey;
     this.model = opts.model ?? "jina-reranker-v2-base-multilingual";
-    this.baseUrl = (opts.baseUrl ?? "https://api.jina.ai/v1").replace(/\/+$/, "");
+    this.baseUrl = (opts.baseUrl ?? "https://api.jina.ai/v1").replace(/\/{1,512}$/, "");
     this.timeoutMs = opts.timeoutMs ?? 30_000;
     this.fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
   }

package/src/structured-output.ts

@@ -24,7 +24,7 @@ export function parseStructuredOutput<T extends z.ZodTypeAny>(
   const cleaned = raw
     .trim()
     .replace(/^```(?:json)?\s*/i, "")
-    .replace(/\s*```$/, "")
+    .replace(/\s{0,1000}```$/, "")
     .trim();
 
   let parsed: unknown;

package/src/text-style-rules.ts

@@ -37,7 +37,7 @@ export interface TextStyleRule {
 export const replaceEmDash: TextStyleRule = {
   name: "replace-em-dash",
   description: "U+2014 «—» → «-» (с нормализацией пробелов)",
-  apply: (s) => s.replace(/\s*—\s*/g, " - ").replace(/ {2,}/g, " "),
+  apply: (s) => s.replace(/\s{0,200}—\s{0,200}/g, " - ").replace(/ {2,}/g, " "),
 };
 
 /**

package/src/vision.ts

@@ -74,7 +74,7 @@ export async function classifyPhoto(opts: ClassifyPhotoOptions): Promise<PhotoCl
   if (!opts.apiKey || opts.apiKey.trim().length === 0) {
     throw new Error("classifyPhoto: apiKey required");
   }
-  const baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL).replace(/\/+$/, "");
+  const baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL).replace(/\/{1,512}$/, "");
   const fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
   const mime = opts.mimeType?.trim() ? opts.mimeType : "image/jpeg";
   const base64 = Buffer.from(opts.bytes).toString("base64");
@@ -210,7 +210,7 @@ export async function extractPassportIdentity(
   if (!opts.apiKey || opts.apiKey.trim().length === 0) {
     throw new Error("extractPassportIdentity: apiKey required");
   }
-  const baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL).replace(/\/+$/, "");
+  const baseUrl = (opts.baseUrl ?? DEFAULT_BASE_URL).replace(/\/{1,512}$/, "");
   const fetchImpl = opts.fetch ?? globalThis.fetch.bind(globalThis);
   const mime = opts.mimeType?.trim() ? opts.mimeType : "image/jpeg";
   const base64 = Buffer.from(opts.bytes).toString("base64");