@inerrata-corporation/errata 2.0.2-dev.302 → 2.0.2-dev.307

package/errata.mjs

@@ -28119,6 +28119,96 @@ var init_src10 = __esm({
   }
 });
 
+// src/llm-provider.ts
+function chatProvider() {
+  return (process.env["EXTRACTION_PROVIDER"] ?? "anthropic").toLowerCase() === "azure" ? "azure" : "anthropic";
+}
+function deploymentFor(model) {
+  const key = `MODEL_${model.toUpperCase().replace(/[^A-Z0-9]/g, "_")}_OPENAI`;
+  return process.env[key] ?? process.env["AZURE_OPENAI_DEPLOYMENT"];
+}
+function isChatConfigured(model) {
+  if (chatProvider() === "azure") {
+    return Boolean(
+      process.env["AZURE_OPENAI_API_KEY"] && process.env["AZURE_OPENAI_ENDPOINT"] && deploymentFor(model)
+    );
+  }
+  return Boolean(process.env["ANTHROPIC_API_KEY"]);
+}
+async function chat(req) {
+  const provider = chatProvider();
+  try {
+    if (provider === "azure") {
+      const endpoint = (process.env["AZURE_OPENAI_ENDPOINT"] ?? "").replace(/\/+$/, "");
+      const version2 = process.env["AZURE_OPENAI_API_VERSION"] ?? "2024-10-21";
+      const deployment = deploymentFor(req.model);
+      if (!endpoint || !deployment) {
+        console.warn("[errata] llm: azure selected but endpoint/deployment missing \u2014 skipping");
+        return null;
+      }
+      const resp2 = await fetch(
+        `${endpoint}/openai/deployments/${deployment}/chat/completions?api-version=${version2}`,
+        {
+          method: "POST",
+          headers: {
+            "api-key": process.env["AZURE_OPENAI_API_KEY"] ?? "",
+            "content-type": "application/json"
+          },
+          body: JSON.stringify({
+            messages: [
+              ...req.system ? [{ role: "system", content: req.system }] : [],
+              { role: "user", content: req.user }
+            ],
+            // `max_completion_tokens`: the newer deployments reject `max_tokens`.
+            max_completion_tokens: req.maxTokens
+          })
+        }
+      );
+      if (!resp2.ok) {
+        console.warn(
+          `[errata] llm: azure ${resp2.status} \u2014 ${(await resp2.text().catch(() => "")).slice(0, 200)}`
+        );
+        return null;
+      }
+      const json3 = await resp2.json();
+      return json3.choices?.[0]?.message?.content ?? null;
+    }
+    const resp = await fetch("https://api.anthropic.com/v1/messages", {
+      method: "POST",
+      headers: {
+        "x-api-key": process.env["ANTHROPIC_API_KEY"] ?? "",
+        "anthropic-version": "2023-06-01",
+        "content-type": "application/json"
+      },
+      body: JSON.stringify({
+        model: req.model,
+        max_tokens: req.maxTokens,
+        ...req.system ? { system: req.system } : {},
+        messages: [{ role: "user", content: req.user }]
+      })
+    });
+    if (!resp.ok) {
+      console.warn(
+        `[errata] llm: anthropic ${resp.status} \u2014 ${(await resp.text().catch(() => "")).slice(0, 200)}`
+      );
+      return null;
+    }
+    const json2 = await resp.json();
+    return json2.content?.find((c) => c.type === "text")?.text ?? null;
+  } catch (err2) {
+    console.warn(
+      `[errata] llm: ${provider} transport failed \u2014`,
+      err2 instanceof Error ? err2.message : err2
+    );
+    return null;
+  }
+}
+var init_llm_provider = __esm({
+  "src/llm-provider.ts"() {
+    "use strict";
+  }
+});
+
 // src/symbol-summaries.ts
 import { existsSync as existsSync8, readFileSync as readFileSync6, writeFileSync as writeFileSync7 } from "node:fs";
 import { join as join7 } from "node:path";
@@ -28148,6 +28238,19 @@ function summariesByBodyHash(cache) {
   }
   return m;
 }
+function buildDocIndex(store) {
+  const byName = /* @__PURE__ */ new Map();
+  for (const c of store.findNodesByLabel("Comment")) {
+    const name2 = c.attrs["attachedTo"];
+    const kind = c.attrs["kind"];
+    if (typeof name2 !== "string" || !name2) continue;
+    if (typeof kind !== "string" || !DOC_COMMENT_KINDS.has(kind)) continue;
+    const text = typeof c.attrs["text"] === "string" ? c.attrs["text"] : c.description;
+    if (!text?.trim()) continue;
+    if (!byName.has(name2)) byName.set(name2, text.trim().slice(0, MAX_DOC_CHARS));
+  }
+  return byName;
+}
 function parseSummaryJson(text, expected) {
   const start2 = text.indexOf("[");
   const end = text.lastIndexOf("]");
@@ -28166,44 +28269,32 @@ function parseSummaryJson(text, expected) {
   }
   return out2;
 }
-function haikuIntentSummarizer(apiKey) {
-  if (!apiKey) return void 0;
+function haikuIntentSummarizer(_apiKey) {
+  if (!isChatConfigured(SUMMARY_MODEL)) return void 0;
   return async (symbols) => {
-    try {
-      const listing = symbols.map(
-        (s, i2) => `### ${i2 + 1}. kind: ${s.kind}
+    const listing = symbols.map(
+      (s, i2) => `### ${i2 + 1}. kind: ${s.kind}
 name: ${s.name}
-${s.snippet ? `\`\`\`
+` + (s.doc ? `doc: ${s.doc}
+` : "") + `${s.snippet ? `\`\`\`
 ${s.snippet}
 \`\`\`` : "(no source available \u2014 describe from the name)"}`
-      ).join("\n\n");
-      const resp = await fetch("https://api.anthropic.com/v1/messages", {
-        method: "POST",
-        headers: {
-          "x-api-key": apiKey,
-          "anthropic-version": "2023-06-01",
-          "content-type": "application/json"
-        },
-        body: JSON.stringify({
-          model: SUMMARY_MODEL,
-          max_tokens: 2e3,
-          messages: [{ role: "user", content: `${SUMMARY_PROMPT}
+    ).join("\n\n");
+    const text = await chat({
+      model: SUMMARY_MODEL,
+      user: `${SUMMARY_PROMPT}
 
 SYMBOLS:
-${listing}` }]
-        })
-      });
-      if (!resp.ok) return symbols.map(() => null);
-      const json2 = await resp.json();
-      return parseSummaryJson(json2.content?.find((c) => c.type === "text")?.text ?? "", symbols.length);
-    } catch {
-      return symbols.map(() => null);
-    }
+${listing}`,
+      maxTokens: 2e3
+    });
+    if (text === null) return symbols.map(() => null);
+    return parseSummaryJson(text, symbols.length);
   };
 }
 function envIntentSummarizer() {
   if (process.env["ERRATA_SYMBOL_SUMMARIES"] !== "1") return void 0;
-  return haikuIntentSummarizer(process.env["ANTHROPIC_API_KEY"] ?? null);
+  return haikuIntentSummarizer();
 }
 function readSnippet(workspaceRoot, attrs) {
   const relPath = attrs["relPath"];
@@ -28227,6 +28318,7 @@ async function runSymbolSummarySweep(store, workspaceRoot, configDir, summarizer
   const cache = loadSymbolSummaryCache(configDir);
   const candidates = [];
   const seen = /* @__PURE__ */ new Set();
+  const docs = buildDocIndex(store);
   for (const [label, kind] of SUMMARY_LABELS) {
     for (const n of store.findNodesByLabel(label)) {
       const bodyHash = n.attrs["bodyHash"];
@@ -28236,7 +28328,15 @@ async function runSymbolSummarySweep(store, workspaceRoot, configDir, summarizer
       if (!name2 || !isDistinctiveIdentifier(name2)) continue;
       seen.add(bodyHash);
       const snippet = readSnippet(workspaceRoot, n.attrs);
-      candidates.push({ bodyHash, name: name2, kind, ...snippet ? { snippet } : {} });
+      const qname = typeof n.attrs["qname"] === "string" ? n.attrs["qname"] : void 0;
+      const doc = docs.get(name2) ?? (qname ? docs.get(qname) : void 0);
+      candidates.push({
+        bodyHash,
+        name: name2,
+        kind,
+        ...snippet ? { snippet } : {},
+        ...doc ? { doc } : {}
+      });
     }
   }
   const todo = candidates.slice(0, maxSymbols);
@@ -28266,12 +28366,13 @@ async function runSymbolSummarySweep(store, workspaceRoot, configDir, summarizer
   }
   return { generated, rejected, remaining };
 }
-var SUMMARY_LABELS, MAX_SNIPPET_CHARS, DEFAULT_MAX_SYMBOLS_PER_SWEEP, DEFAULT_BATCH_SIZE, SUMMARY_MODEL, SUMMARY_PROMPT;
+var SUMMARY_LABELS, MAX_SNIPPET_CHARS, DEFAULT_MAX_SYMBOLS_PER_SWEEP, DEFAULT_BATCH_SIZE, DOC_COMMENT_KINDS, MAX_DOC_CHARS, SUMMARY_MODEL, SUMMARY_PROMPT;
 var init_symbol_summaries = __esm({
   "src/symbol-summaries.ts"() {
     "use strict";
     init_src();
     init_generalize_graph();
+    init_llm_provider();
     SUMMARY_LABELS = [
       ["Function", "function"],
       ["Method", "method"],
@@ -28280,10 +28381,13 @@ var init_symbol_summaries = __esm({
     MAX_SNIPPET_CHARS = 1500;
     DEFAULT_MAX_SYMBOLS_PER_SWEEP = 64;
     DEFAULT_BATCH_SIZE = 16;
+    DOC_COMMENT_KINDS = /* @__PURE__ */ new Set(["docstring", "explanation"]);
+    MAX_DOC_CHARS = 600;
     SUMMARY_MODEL = "claude-haiku-4-5-20251001";
     SUMMARY_PROMPT = `For each code symbol below, write ONE short English phrase stating its intent/role, shaped like "a <kind> that <does what>" (e.g. "a function that binds the server's listen port").
+Some symbols include a "doc:" line \u2014 the author's own description of what it is for. When present TRUST IT as the primary source of intent and translate it into the required shape; the code is secondary evidence. Carry the SPECIFIC job the symbol does, not the category it belongs to: "a function that releases a pooled connection on every exit path including cancellation" is useful; "a function that handles cleanup" is not.
 HARD RULES \u2014 a violating phrase is discarded:
-- NEVER include any identifier, symbol name, variable, type name, file path, or module name from the code \u2014 describe meaning in plain English words only.
+- NEVER include any identifier, symbol name, variable, type name, file path, or module name from the code or the doc \u2014 describe meaning in plain English words only.
 - NEVER include quoted strings, literals, numbers copied from the code, or code syntax.
 - One phrase per symbol, under 200 characters, lowercase start, no trailing period.
 Output a JSON array of strings, one per symbol in the SAME ORDER (use null for a symbol you cannot describe). Output JSON only, no prose.`;
@@ -52902,7 +53006,7 @@ function startLoopLagMonitor(thresholdMs = 1e3) {
 }
 
 // src/engine.ts
-var DAEMON_VERSION = true ? "2.0.2-dev.302" : "2.0.0-alpha.0";
+var DAEMON_VERSION = true ? "2.0.2-dev.307" : "2.0.0-alpha.0";
 var IGNORED_PATH = /[\\/](?:\.git|node_modules|\.errata|\.claude|\.codex|\.cursor|\.turbo|\.next|dist|coverage|test-results|playwright-report|__pycache__)(?:[\\/]|$)/;
 var IGNORED_NOISE = /castalia\.db|eventlog\.sqlite|turn-cursors/;
 var GIT_OP_MUTE_MS = 4e3;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@inerrata-corporation/errata",
-  "version": "2.0.2-dev.302",
+  "version": "2.0.2-dev.307",
   "description": "errata - local-first observation engine for AI coding agents",
   "type": "module",
   "bin": {