АДРЕСНЫЙ РЕЖИМ - локальная подель на декомпозе

llm_normalizer/backend/src/services/addressMcpClient.ts

@@ -3,6 +3,7 @@
   ASSISTANT_MCP_PROXY_URL,
   ASSISTANT_MCP_TIMEOUT_MS
 } from "../config";
+import iconv from "iconv-lite";
 
 interface McpExecuteQueryResponse {
   success?: unknown;
@@ -36,8 +37,81 @@ function parseFiniteNumber(value: unknown): number | null {
   return null;
 }
 
+function textMojibakeScore(value: string): number {
+  const source = String(value ?? "");
+  const cyrillic = (source.match(/[А-Яа-яЁё]/g) ?? []).length;
+  const latin = (source.match(/[A-Za-z]/g) ?? []).length;
+  const hardMarkers = (source.match(/[Ѓѓ‚„…†‡€‰‹ЉЊЌЋЏ‘’“”•–—™љ›њќћџ]/g) ?? []).length;
+  const pairMarkers = (source.match(/(?:Р.|С.|Ð.|Ñ.)/g) ?? []).length;
+  return cyrillic + latin - hardMarkers * 3 - pairMarkers * 2;
+}
+
+function looksLikeMojibake(value: string): boolean {
+  const source = String(value ?? "");
+  if (!source.trim()) {
+    return false;
+  }
+  if (/[Ѓѓ‚„…†‡€‰‹ЉЊЌЋЏ‘’“”•–—™љ›њќћџ]/.test(source)) {
+    return true;
+  }
+  return (source.match(/(?:Р.|С.|Ð.|Ñ.)/g) ?? []).length >= 2;
+}
+
+function decodeUtf8FromWin1251Mojibake(value: string): string {
+  if (!looksLikeMojibake(value)) {
+    return value;
+  }
+  try {
+    const bytes = iconv.encode(value, "win1251");
+    const decoded = bytes.toString("utf8");
+    return textMojibakeScore(decoded) > textMojibakeScore(value) ? decoded : value;
+  } catch {
+    return value;
+  }
+}
+
+function decodeUtf8FromLatin1Mojibake(value: string): string {
+  if (!looksLikeMojibake(value)) {
+    return value;
+  }
+  try {
+    const decoded = Buffer.from(value, "latin1").toString("utf8");
+    return textMojibakeScore(decoded) > textMojibakeScore(value) ? decoded : value;
+  } catch {
+    return value;
+  }
+}
+
+function normalizeMojibakeString(value: string): string {
+  const fromWin1251 = decodeUtf8FromWin1251Mojibake(value);
+  return decodeUtf8FromLatin1Mojibake(fromWin1251);
+}
+
+function normalizeMojibakeValue(value: unknown): unknown {
+  if (typeof value === "string") {
+    return normalizeMojibakeString(value);
+  }
+  if (Array.isArray(value)) {
+    return value.map((item) => normalizeMojibakeValue(item));
+  }
+  if (value && typeof value === "object") {
+    const source = value as Record<string, unknown>;
+    const normalized: Record<string, unknown> = {};
+    for (const [key, raw] of Object.entries(source)) {
+      const repairedKey = normalizeMojibakeString(key);
+      normalized[repairedKey] = normalizeMojibakeValue(raw);
+    }
+    return normalized;
+  }
+  return value;
+}
+
+function normalizeMojibakeRows(rows: Array<Record<string, unknown>>): Array<Record<string, unknown>> {
+  return rows.map((row) => normalizeMojibakeValue(row) as Record<string, unknown>);
+}
+
 function parseRowsFromTextTable(source: string): Array<Record<string, unknown>> {
-  const normalized = String(source ?? "").replace(/\r/g, "").trim();
+  const normalized = normalizeMojibakeString(String(source ?? "")).replace(/\r/g, "").trim();
   if (!normalized) {
     return [];
   }
@@ -111,7 +185,7 @@ function parseRowsFromTextTable(source: string): Array<Record<string, unknown>>
     rows.push(row);
   }
 
-  return rows;
+  return normalizeMojibakeRows(rows);
 }
 
 function parseExecutePayload(payload: unknown): AddressMcpQueryResult {
@@ -133,9 +207,11 @@ function parseExecutePayload(payload: unknown): AddressMcpQueryResult {
   }
 
   if (Array.isArray(source.data)) {
-    const rows = source.data
-      .map((item) => (item && typeof item === "object" ? (item as Record<string, unknown>) : null))
-      .filter((item): item is Record<string, unknown> => item !== null);
+    const rows = normalizeMojibakeRows(
+      source.data
+        .map((item) => (item && typeof item === "object" ? (item as Record<string, unknown>) : null))
+        .filter((item): item is Record<string, unknown> => item !== null)
+    );
     return {
       ok: true,
       rows,
@@ -152,9 +228,11 @@ function parseExecutePayload(payload: unknown): AddressMcpQueryResult {
   }
 
   if (source.data && typeof source.data === "object" && Array.isArray((source.data as { rows?: unknown }).rows)) {
-    const rows = ((source.data as { rows: unknown[] }).rows ?? [])
-      .map((item) => (item && typeof item === "object" ? (item as Record<string, unknown>) : null))
-      .filter((item): item is Record<string, unknown> => item !== null);
+    const rows = normalizeMojibakeRows(
+      ((source.data as { rows: unknown[] }).rows ?? [])
+        .map((item) => (item && typeof item === "object" ? (item as Record<string, unknown>) : null))
+        .filter((item): item is Record<string, unknown> => item !== null)
+    );
     return {
       ok: true,
       rows,