@polycode-projects/the-mechanical-code-talker 2.0.3 → 2.2.0

package/README.md

@@ -180,7 +180,7 @@ resolves to a real graph traversal or declines honestly:
   (*because/although/while*), conditionals, and false-premise flags ("why
   does X still import Y" when it no longer does).
 
-The full catalog with measured coverage lives in `CAPABILITIES_1.7.3.md` and
+The full catalog with measured coverage lives in `CAPABILITIES_2.0.3.md` and
 the `BENCHMARK_*.md` reports.
 
 **Response finishing.** Before an answer prints, it is segmented into typed

package/ROADMAP.md

@@ -100,7 +100,30 @@ blending web-sourced facts with graph/operator facts.
 
 ## Design docs
 
-Every substantial design lives in its own `PLAN_*.md` at the repo root (active) or `archive/`
-(shipped and closed) — this file points to them, it doesn't repeat their content. `SKILL_*.md` docs
-specify the repeatable measurement/build cycles (benchmarks, the background strategy advisor,
-plain-prose writing). `HANDOVER.md` is the single current-open-items list.
+Every substantial design lives in its own `PLAN_*.md` at the repo root; `archive/` holds the shipped
+and closed ones. This file points to them, it doesn't repeat their content. Each plan states its own
+status in its opening lines — read it there, because a status quoted here would rot.
+
+| Plan | What it's for |
+| --- | --- |
+| [PLAN_ADVENTURE.md](PLAN_ADVENTURE.md) | a text adventure as an architectural stretch: imperative command grammar, NPC turn scheduler, room-look digest |
+| [PLAN_AGENTS.md](PLAN_AGENTS.md) | the governing plan for the multi-repo arc (marginalia, seonix, a pluggable LLM rung), with its own phase sequencing |
+| [PLAN_CHILD_CORPUS.md](PLAN_CHILD_CORPUS.md) | a wider default seed corpus, chosen by age of acquisition |
+| [PLAN_CLASS_QUERY.md](PLAN_CLASS_QUERY.md) | "list/count all X of class Y", reconciled against what already shipped |
+| [PLAN_CODE.md](PLAN_CODE.md) | program synthesis over tmct's closed DSLs, plus JS/HTML/CSS fragments and goal-directed program repair |
+| [PLAN_CONSISTENCY_CHECK.md](PLAN_CONSISTENCY_CHECK.md) | tmct as a consistency service for an LLM tool loop |
+| [PLAN_EMBEDDINGS.md](PLAN_EMBEDDINGS.md) | the semantic-similarity axis, and the way back to it |
+| [PLAN_GRAPH_SCAN.md](PLAN_GRAPH_SCAN.md) | seed and query cost at `init:xl`/`init:xxl` corpus scale |
+| [PLAN_GUESS_NUMBER.md](PLAN_GUESS_NUMBER.md) | closed-loop planning over hidden state, via belief-interval bisection |
+| [PLAN_MUD.md](PLAN_MUD.md) | persistent, shared tmct worlds over a `server:` memory backend |
+| [PLAN_NLU_BENCHMARKS.md](PLAN_NLU_BENCHMARKS.md) | scoring tmct on the CLINC150 and HWU64 intent sets |
+| [PLAN_OPEN_ITEMS.md](PLAN_OPEN_ITEMS.md) | the build order closing the backlog `HANDOVER.md` carries |
+| [PLAN_PARAPHRASE_VERIFICATION.md](PLAN_PARAPHRASE_VERIFICATION.md) | checking a paraphrase against the graph before it prints |
+| [PLAN_PURGE.md](PLAN_PURGE.md) | promote the load-bearing code, delete the dead weight |
+| [PLAN_REPO_INDEX.md](PLAN_REPO_INDEX.md) | tmct grows its own code parsers, ported from seonix |
+| [PLAN_SYLLOGIST.md](PLAN_SYLLOGIST.md) | the reasoning engine's incrementality and retraction horizon |
+| [PLAN_SYLLOGIST_EL_DL.md](PLAN_SYLLOGIST_EL_DL.md) | beyond OWL 2 RL: an EL classifier, then a DL tableau prover |
+
+`SKILL_*.md` docs specify the repeatable measurement and build cycles (the benchmarks, the capability
+audit, the background strategy advisor, plain-prose writing). `HANDOVER.md` is the single
+current-open-items list.

package/corpus/generated/ace-surface-variants.jsonl

@@ -1,3 +1,4 @@
+{"kind":"rescue","sentence":"Code search is a retrieval task","rescued":"Code hunt is a retrieval task","from":"search","to":"hunt","pos":"noun","synsetId":"00947217-n","sourceFile":"PLAN_EMBEDDINGS.md","provenance":"wordnet:00947217-n"}
 {"kind":"variant","seed":"the ace of trumps is a sure winner","generated":"the one of trumps is a sure winner","from":"ace","to":"one","pos":"noun","synsetId":"13764713-n","sourceCorpus":"corpus/tier2/human-examples-large.jsonl","provenance":"wordnet:13764713-n"}
 {"kind":"variant","seed":"the ace of trumps is a sure winner","generated":"the single of trumps is a sure winner","from":"ace","to":"single","pos":"noun","synsetId":"13764713-n","sourceCorpus":"corpus/tier2/human-examples-large.jsonl","provenance":"wordnet:13764713-n"}
 {"kind":"variant","seed":"the ace of trumps is a sure winner","generated":"the unity of trumps is a sure winner","from":"ace","to":"unity","pos":"noun","synsetId":"13764713-n","sourceCorpus":"corpus/tier2/human-examples-large.jsonl","provenance":"wordnet:13764713-n"}

package/corpus/generated/manifest.json

@@ -2,8 +2,8 @@
   "version": 1,
   "generated": "by scripts/generate-template-variants.mjs",
   "file": "ace-surface-variants.jsonl",
-  "rows": 17,
-  "bytes": 4774,
-  "sha256": "5364415f73f7a366371f1fdcc072b5059f81871ef48b106088f05b1441def569",
+  "rows": 18,
+  "bytes": 5012,
+  "sha256": "28fae5a3353f5dd1df365b4407e0ac2a5c66b98b232d8bf454fe3573f2c16760",
   "license": "CC-BY-4.0 (WordNet-derived synonym substitutions of Open English WordNet / SemCor example sentences and this repo's own MPL-2.0 docs prose — see corpus/generated/README.md)"
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@polycode-projects/the-mechanical-code-talker",
-  "version": "2.0.3",
+  "version": "2.2.0",
   "private": false,
   "type": "module",
   "description": "The Mechanical Code Talker (tmct) — a tolerant, offline, $0 chat surface that guides you toward precision queries about a software repository. ELIZA/PARRY-style but domain-obsessed with code. No model calls; no codebase index of its own.",

package/src/adapters/toml-config.mjs

@@ -139,7 +139,6 @@ export async function normalizeConfig(raw, { configDir } = {}) {
   if (t.impl_of_interface !== undefined) tune.implOfInterface = t.impl_of_interface;
   if (t.beam_search !== undefined) tune.beamSearch = t.beam_search;
   if (t.beam_width !== undefined) tune.beamWidth = t.beam_width;
-  if (t.embed_rank !== undefined) tune.embedRank = t.embed_rank;
   if (t.prose_layers !== undefined) tune.proseLayers = t.prose_layers;
   const exp = t.expansion || {};
   const expansion = {};

package/src/domain/codegraph.mjs

@@ -1,5 +1,4 @@
 import { lookupByProseTokens, proseLayerHits, splitIdentifierWords } from "./prose.mjs";
-import { cosine } from "./vector.mjs";
 import { CREATED_AT_PROP, UPDATED_AT_PROP } from "./memory/trust.mjs";
 
 // Pure (no-network, no-fs) query logic over the typed `entities` payload that the
@@ -511,15 +510,6 @@ const LIT_COMP_CAP = 4;
 const LIT_FRAC = 1.0;
 const LIT_CAP_FRAC = 0.9;
 
-// opt-in via embedRank + an injected embedder: static-embedding re-rank over
-// path/symbol/doc text read from the graph (never source), cached per-process in EMB_CACHE
-const EMB_FRAC = 0.2;
-const EMB_CAP_FRAC = 0.35;
-const EMB_TEXT_SYMBOL_CAP = 64;
-const EMB_TEXT_DOC_CAP = 12;
-const EMB_CACHE = new WeakMap(); // graph -> { embedder, texts, vecs: Map<moduleId, Float32Array> }
-let embedWarned = false;
-
 // opt-in via beamSearch: multi-ply adaptive expansion of the proximity nudge above.
 // Beam width is a margin relative to each ply's best score (not a fixed count), so a
 // weak-then-strong candidate isn't prematurely discarded. Successors are generated
@@ -544,30 +534,6 @@ const SPIRAL_HOP_DECAY = 0.6;
 const SPIRAL_PROX_FRAC = 0.2;
 const SPIRAL_PROX_CAP_FRAC = 0.35;
 
-/** embedRank: per-module embeddable text from path components + defined symbol
- *  names + doc first-lines, cached alongside the vectors in EMB_CACHE. */
-function moduleEmbedTexts(graph) {
-  const texts = new Map(); // moduleId -> text
-  const defIdx = definesIndex(graph);
-  const docs = new Map();  // moduleId -> [doc first-lines]
-  for (const ind of graph.individuals) {
-    const doc = (ind.attributes || []).find((a) => a.key === "doc")?.value;
-    if (!doc) continue;
-    const modId = (ind.class || "") === "Module" ? ind.id : moduleIdOf(graph, ind);
-    if (!modId) continue;
-    let arr = docs.get(modId);
-    if (!arr) docs.set(modId, (arr = []));
-    if (arr.length < EMB_TEXT_DOC_CAP) arr.push(String(doc).split("\n")[0]);
-  }
-  for (const ind of graph.individuals) {
-    if ((ind.class || "") !== "Module") continue;
-    const parts = String(ind.label).split(/[^a-zA-Z0-9_]+/).filter(Boolean);
-    const syms = (defIdx.get(ind.id) || []).slice(0, EMB_TEXT_SYMBOL_CAP);
-    texts.set(ind.id, [...parts, ...syms, ...(docs.get(ind.id) || [])].join(" "));
-  }
-  return texts;
-}
-
 /** Split a lowercased path label into boundary components: django/utils/text.py →
  *  {django,utils,text,py}. Component equality (not substring) stops "text" matching "ci<text>". */
 function pathComponents(labelLc) {
@@ -773,7 +739,7 @@ export function spiralExpand(graph, scored = [], {
  *  IDF-weights each query token, scores path/symbol/exact-symbol matches, and
  *  re-ranks with a bounded import-proximity bonus. Pure; deterministic. */
 function scoreModules(graph, tokens, opts = {}) {
-  const { demoteNonProd = false, callAdjacency = false, implOfInterface = false, beamSearch = false, spiral = false, proseBoost = false, proseLayers = false, literalMention = false, embedRank = false, rawQuery = "" } = opts;
+  const { demoteNonProd = false, callAdjacency = false, implOfInterface = false, beamSearch = false, spiral = false, proseBoost = false, proseLayers = false, literalMention = false, rawQuery = "" } = opts;
   const beamWidth = Number.isFinite(opts.beamWidth) && opts.beamWidth > 0 ? opts.beamWidth : 8;
   const defIdx = definesIndex(graph);
   // Precompute each module's path components + defined-symbol exact/component sets, once.
@@ -967,36 +933,6 @@ function scoreModules(graph, tokens, opts = {}) {
       s.score += Math.min(signal * PROSE_LAYER_FRAC, s.score * PROSE_LAYER_CAP_FRAC);
     }
   }
-  // embedRank: the embedder is injected so this module stays fs-free; absent -> a one-time
-  // stderr note, never a failure
-  if (embedRank) {
-    if (!opts.embedder) {
-      if (!embedWarned) {
-        embedWarned = true;
-        process.stderr.write("tmct: embedRank requested but no embedder available (weights not fetched? see `npm run refs:embeddings`) — flag is a no-op\n");
-      }
-    } else if (scored.length) {
-      const embedder = opts.embedder;
-      let cache = EMB_CACHE.get(graph);
-      if (!cache || cache.embedder !== embedder) {
-        cache = { embedder, texts: moduleEmbedTexts(graph), vecs: new Map() };
-        EMB_CACHE.set(graph, cache);
-      }
-      const qv = embedder.embed(rawQuery || tokens.join(" "));
-      let maxBase = 0;
-      for (const s of scored) maxBase = Math.max(maxBase, s.score);
-      for (const s of scored) {
-        let v = cache.vecs.get(s.ind.id);
-        if (!v) {
-          v = embedder.embed(cache.texts.get(s.ind.id) || String(s.ind.label));
-          cache.vecs.set(s.ind.id, v);
-        }
-        const sim = Math.max(0, cosine(qv, v)); // negative similarity never penalises
-        if (!sim) continue;
-        s.score += Math.min(sim * maxBase * EMB_FRAC, s.score * EMB_CAP_FRAC);
-      }
-    }
-  }
   // beamSearch (opt-in): multi-ply generalization of the single-hop families above.
   if (beamSearch && scored.length > 1) beamExpand(graph, scored, beamWidth);
   // SPIRAL (opt-in): bounded-radius ego walk that may introduce lexically-invisible modules — runs
@@ -1024,9 +960,9 @@ export function searchModulesRanked(graph, query, opts = {}) {
   const tokens = raw.toLowerCase().split(/[^a-z0-9_]+/).filter(Boolean);
   if (!tokens.length) return [];
   // literalMention needs the query BEFORE tokenization (the tokenizer destroys the dotted refs
-  // it matches on) and embedRank embeds the raw phrasing; threaded only when a flag that
-  // consumes it is on, so the OFF path is provably unchanged.
-  const effOpts = (opts.literalMention || opts.embedRank) ? { ...opts, rawQuery: raw } : opts;
+  // it matches on); threaded only when a flag that consumes it is on, so the OFF path is
+  // provably unchanged.
+  const effOpts = opts.literalMention ? { ...opts, rawQuery: raw } : opts;
   return scoreModules(graph, tokens, effOpts).map((s) => ({ path: String(s.ind.label), score: s.score }));
 }
 

package/src/domain/markdown-links.mjs

@@ -0,0 +1,55 @@
+// markdown-links.mjs — pull the relative link targets out of a markdown string.
+// Pure: a string in, `[{ target, line }]` out, no filesystem and no imports, so
+// the CI jobs that run without `npm ci` can reach it.
+//
+// Code is not prose. A doc that writes `[text](target)` inside backticks is
+// showing you what a link looks like, not making one, and a checker that cannot
+// tell the difference reports the example as a broken link to a file named
+// "target". So the spans are blanked before the link patterns run — blanked
+// rather than cut, because every offset behind them still has to name the right
+// line number.
+
+const blank = (line) => line.replace(/[^\n]/g, " ");
+const FENCE = /^[ \t]{0,3}(`{3,}|~{3,})/;
+
+/** Replace every fenced block and inline code span with spaces, keeping the
+ *  string's length and its newlines so later offsets still map to their line.
+ *  Fences are walked line by line: a lazy multiline regex stops at the end of
+ *  the opening fence's own line and blanks only the markers. */
+export function blankCodeSpans(markdown) {
+  let fence = null;
+  const lines = markdown.split("\n").map((line) => {
+    const marker = line.match(FENCE)?.[1];
+    if (fence) {
+      if (marker && marker[0] === fence[0] && marker.length >= fence.length) fence = null;
+      return blank(line);
+    }
+    if (marker) fence = marker;
+    return marker ? blank(line) : line;
+  });
+  return lines.join("\n").replace(/(`+)[\s\S]*?\1/g, blank);
+}
+
+// Inline links/images: [text](target "title") — target ends at the first
+// whitespace or closing paren. Reference definitions: [label]: target.
+const INLINE_LINK = /!?\[[^\]]*\]\(([^)\s]+)(?:\s+"[^"]*")?\)/g;
+const REFERENCE_DEF = /^\s{0,3}\[[^\]]+\]:\s+(\S+)/gm;
+
+/** Every repo-relative link target in `markdown`, with the 1-based line it sits
+ *  on. External URLs, bare #anchors and absolute paths are out of scope. */
+export function relativeTargets(markdown) {
+  const prose = blankCodeSpans(markdown);
+  const targets = [];
+  for (const regex of [INLINE_LINK, REFERENCE_DEF]) {
+    for (const match of prose.matchAll(regex)) {
+      let target = match[1];
+      if (/^(https?|mailto|ftp):/i.test(target)) continue;
+      if (target.startsWith("#") || target.startsWith("/") || target.startsWith("<")) continue;
+      target = decodeURIComponent(target.split("#")[0].split("?")[0]);
+      if (!target) continue;
+      const line = prose.slice(0, match.index).split("\n").length;
+      targets.push({ target, line });
+    }
+  }
+  return targets;
+}

package/src/surfaces/web/memory-ask-browser.bundle.js

@@ -936,23 +936,6 @@
     }
   });
 
-  // src/domain/vector.mjs
-  function cosine(a, b) {
-    let dot = 0, na = 0, nb = 0;
-    const n = Math.min(a.length, b.length);
-    for (let i = 0; i < n; i++) {
-      dot += a[i] * b[i];
-      na += a[i] * a[i];
-      nb += b[i] * b[i];
-    }
-    if (na === 0 || nb === 0) return 0;
-    return dot / (Math.sqrt(na) * Math.sqrt(nb));
-  }
-  var init_vector = __esm({
-    "src/domain/vector.mjs"() {
-    }
-  });
-
   // src/domain/memory/trust.mjs
   function parseChatTagRest(rest) {
     const at = rest.indexOf("@");
@@ -1307,27 +1290,6 @@
     }
     return idx;
   }
-  function moduleEmbedTexts(graph) {
-    const texts = /* @__PURE__ */ new Map();
-    const defIdx = definesIndex(graph);
-    const docs = /* @__PURE__ */ new Map();
-    for (const ind of graph.individuals) {
-      const doc = (ind.attributes || []).find((a) => a.key === "doc")?.value;
-      if (!doc) continue;
-      const modId = (ind.class || "") === "Module" ? ind.id : moduleIdOf(graph, ind);
-      if (!modId) continue;
-      let arr = docs.get(modId);
-      if (!arr) docs.set(modId, arr = []);
-      if (arr.length < EMB_TEXT_DOC_CAP) arr.push(String(doc).split("\n")[0]);
-    }
-    for (const ind of graph.individuals) {
-      if ((ind.class || "") !== "Module") continue;
-      const parts = String(ind.label).split(/[^a-zA-Z0-9_]+/).filter(Boolean);
-      const syms = (defIdx.get(ind.id) || []).slice(0, EMB_TEXT_SYMBOL_CAP);
-      texts.set(ind.id, [...parts, ...syms, ...docs.get(ind.id) || []].join(" "));
-    }
-    return texts;
-  }
   function identComponents(name) {
     return new Set(String(name).replace(/([a-z0-9])([A-Z])/g, "$1_$2").toLowerCase().split(/[^a-z0-9]+/).filter(Boolean));
   }
@@ -1499,7 +1461,7 @@
     return results;
   }
   function scoreModules(graph, tokens, opts = {}) {
-    const { demoteNonProd = false, callAdjacency = false, implOfInterface = false, beamSearch = false, spiral = false, proseBoost = false, proseLayers = false, literalMention = false, embedRank = false, rawQuery = "" } = opts;
+    const { demoteNonProd = false, callAdjacency = false, implOfInterface = false, beamSearch = false, spiral = false, proseBoost = false, proseLayers = false, literalMention = false, rawQuery = "" } = opts;
     const beamWidth = Number.isFinite(opts.beamWidth) && opts.beamWidth > 0 ? opts.beamWidth : 8;
     const defIdx = definesIndex(graph);
     const modules = [];
@@ -1691,34 +1653,6 @@
         s.score += Math.min(signal * PROSE_LAYER_FRAC, s.score * PROSE_LAYER_CAP_FRAC);
       }
     }
-    if (embedRank) {
-      if (!opts.embedder) {
-        if (!embedWarned) {
-          embedWarned = true;
-          process.stderr.write("tmct: embedRank requested but no embedder available (weights not fetched? see `npm run refs:embeddings`) \u2014 flag is a no-op\n");
-        }
-      } else if (scored.length) {
-        const embedder = opts.embedder;
-        let cache2 = EMB_CACHE.get(graph);
-        if (!cache2 || cache2.embedder !== embedder) {
-          cache2 = { embedder, texts: moduleEmbedTexts(graph), vecs: /* @__PURE__ */ new Map() };
-          EMB_CACHE.set(graph, cache2);
-        }
-        const qv = embedder.embed(rawQuery || tokens.join(" "));
-        let maxBase = 0;
-        for (const s of scored) maxBase = Math.max(maxBase, s.score);
-        for (const s of scored) {
-          let v = cache2.vecs.get(s.ind.id);
-          if (!v) {
-            v = embedder.embed(cache2.texts.get(s.ind.id) || String(s.ind.label));
-            cache2.vecs.set(s.ind.id, v);
-          }
-          const sim = Math.max(0, cosine(qv, v));
-          if (!sim) continue;
-          s.score += Math.min(sim * maxBase * EMB_FRAC, s.score * EMB_CAP_FRAC);
-        }
-      }
-    }
     if (beamSearch && scored.length > 1) beamExpand(graph, scored, beamWidth);
     if (spiral && scored.length) spiralExpand(graph, scored, {
       depth: Number.isFinite(opts.spiralDepth) && opts.spiralDepth > 0 ? opts.spiralDepth : SPIRAL_DEPTH_DEFAULT,
@@ -1732,7 +1666,7 @@
     const raw = String(query || "");
     const tokens = raw.toLowerCase().split(/[^a-z0-9_]+/).filter(Boolean);
     if (!tokens.length) return [];
-    const effOpts = opts.literalMention || opts.embedRank ? { ...opts, rawQuery: raw } : opts;
+    const effOpts = opts.literalMention ? { ...opts, rawQuery: raw } : opts;
     return scoreModules(graph, tokens, effOpts).map((s) => ({ path: String(s.ind.label), score: s.score }));
   }
   function renderSearch(graph, query, { limit = SEARCH_LIMIT, kind = "", decorator = "", name = "" } = {}) {
@@ -2414,11 +2348,10 @@ ${shown.join("\n")}${tail}`;
   ${list.join("\n  ")}` + (edges.length > EXPORTS_CAP ? `
   \u2026+${edges.length - EXPORTS_CAP} more` : "");
   }
-  var PROP_KIND, isProvRef, DESCRIBE_EDGE_CAP, PROV_CAP, IMPACT_DEPTHS_LISTED, IMPACT_PER_DEPTH, IMPACT_TESTS_PER_DEP, SEARCH_LIMIT, SEARCH_SYMBOLS_SHOWN, PATH_W, SYM_W, EXACT_W, SYM_MATCH_CAP, PROX_FRAC, PROX_CAP_FRAC, isTestLabel, NONPROD_DEMOTE, isNonProdLabel, CALL_PROX_FRAC, CALL_PROX_CAP_FRAC, IMPL_PROX_FRAC, IMPL_PROX_CAP_FRAC, isCsModuleLabel, looksLikeCsInterface, PROSE_PROX_FRAC, PROSE_PROX_CAP_FRAC, PROSE_LOOKUP_LIMIT, PROSE_LAYER_FRAC, PROSE_LAYER_CAP_FRAC, PROSE_LAYER_DISCOUNT, LIT_W, LIT_MIN_COMPONENTS, LIT_COMP_CAP, LIT_FRAC, LIT_CAP_FRAC, EMB_FRAC, EMB_CAP_FRAC, EMB_TEXT_SYMBOL_CAP, EMB_TEXT_DOC_CAP, EMB_CACHE, embedWarned, BEAM_MARGIN_FRAC, BEAM_PROX_FRAC, BEAM_PROX_CAP_FRAC, BEAM_OVERFLOW_CAP, BEAM_PLIES, BEAM_EDGE_GROUPS, SPIRAL_DEPTH_DEFAULT, SPIRAL_NODE_LIMIT_DEFAULT, SPIRAL_Q_DEFAULT, SPIRAL_EXPAND_KINDS, SPIRAL_EMIT_FRAC, SPIRAL_HOP_DECAY, SPIRAL_PROX_FRAC, SPIRAL_PROX_CAP_FRAC, edgesOfKindCache, MEMBERS_CAP, SUBCLASS_CAP, CALL_CAP, attrVal, ARCH_PKG_CAP, ARCH_HUB_CAP, COVERAGE_CAP, TESTS_GRAIN_NOTE, HISTORY_CAP, CALL_SYMBOL_CLASSES, CALL_HINT_CAP, SYMBOL_CLASSES, CONTEXT_SIBLING_CAP, CLASS_MEMBER_CAP, COCHANGE_MID_CAP, CONTEXT_TESTS_CAP, TINY_MAX_LOC, TINY_MAX_ARITY, LARGE_CLASS_MEMBERS, INLINE_CALLEE_CAP, splitDecs, tokenize, countParams, modeOf, COCHANGE_CAP, EXPORTS_CAP;
+  var PROP_KIND, isProvRef, DESCRIBE_EDGE_CAP, PROV_CAP, IMPACT_DEPTHS_LISTED, IMPACT_PER_DEPTH, IMPACT_TESTS_PER_DEP, SEARCH_LIMIT, SEARCH_SYMBOLS_SHOWN, PATH_W, SYM_W, EXACT_W, SYM_MATCH_CAP, PROX_FRAC, PROX_CAP_FRAC, isTestLabel, NONPROD_DEMOTE, isNonProdLabel, CALL_PROX_FRAC, CALL_PROX_CAP_FRAC, IMPL_PROX_FRAC, IMPL_PROX_CAP_FRAC, isCsModuleLabel, looksLikeCsInterface, PROSE_PROX_FRAC, PROSE_PROX_CAP_FRAC, PROSE_LOOKUP_LIMIT, PROSE_LAYER_FRAC, PROSE_LAYER_CAP_FRAC, PROSE_LAYER_DISCOUNT, LIT_W, LIT_MIN_COMPONENTS, LIT_COMP_CAP, LIT_FRAC, LIT_CAP_FRAC, BEAM_MARGIN_FRAC, BEAM_PROX_FRAC, BEAM_PROX_CAP_FRAC, BEAM_OVERFLOW_CAP, BEAM_PLIES, BEAM_EDGE_GROUPS, SPIRAL_DEPTH_DEFAULT, SPIRAL_NODE_LIMIT_DEFAULT, SPIRAL_Q_DEFAULT, SPIRAL_EXPAND_KINDS, SPIRAL_EMIT_FRAC, SPIRAL_HOP_DECAY, SPIRAL_PROX_FRAC, SPIRAL_PROX_CAP_FRAC, edgesOfKindCache, MEMBERS_CAP, SUBCLASS_CAP, CALL_CAP, attrVal, ARCH_PKG_CAP, ARCH_HUB_CAP, COVERAGE_CAP, TESTS_GRAIN_NOTE, HISTORY_CAP, CALL_SYMBOL_CLASSES, CALL_HINT_CAP, SYMBOL_CLASSES, CONTEXT_SIBLING_CAP, CLASS_MEMBER_CAP, COCHANGE_MID_CAP, CONTEXT_TESTS_CAP, TINY_MAX_LOC, TINY_MAX_ARITY, LARGE_CLASS_MEMBERS, INLINE_CALLEE_CAP, splitDecs, tokenize, countParams, modeOf, COCHANGE_CAP, EXPORTS_CAP;
   var init_codegraph = __esm({
     "src/domain/codegraph.mjs"() {
       init_prose();
-      init_vector();
       init_trust();
       PROP_KIND = {
         // v2.0 faithful tokens (SEON-faithful realign)
@@ -2484,12 +2417,6 @@ ${shown.join("\n")}${tail}`;
       LIT_COMP_CAP = 4;
       LIT_FRAC = 1;
       LIT_CAP_FRAC = 0.9;
-      EMB_FRAC = 0.2;
-      EMB_CAP_FRAC = 0.35;
-      EMB_TEXT_SYMBOL_CAP = 64;
-      EMB_TEXT_DOC_CAP = 12;
-      EMB_CACHE = /* @__PURE__ */ new WeakMap();
-      embedWarned = false;
       BEAM_MARGIN_FRAC = 0.5;
       BEAM_PROX_FRAC = 0.2;
       BEAM_PROX_CAP_FRAC = 0.35;

package/src/adapters/embed.mjs

@@ -1,169 +0,0 @@
-// embed.mjs — deterministic static-embedding lookup (model2vec potion-base-8M,
-// 29,528 WordPiece subwords × 256 fp32 dims). Pure table lookup + float
-// arithmetic, no ONNX runtime, no network after the one-time fetch — the same
-// text always embeds to the same vector.
-//
-// The safetensors reader and WordPiece tokenizer below are hand-rolled: both
-// formats are simple enough to parse directly, avoiding an ONNX/HF tokenizer
-// dependency for it.
-//
-// Weights are gitignored (vendor/, fetched by scripts/fetch-embeddings.mjs)
-// and never in the npm package; loadEmbedder() returns null when absent so
-// CI/tests never require the download.
-
-import { readFileSync, existsSync } from "node:fs";
-import { join, dirname } from "node:path";
-import { fileURLToPath } from "node:url";
-
-const MODEL_FILE = "model.safetensors";
-const TOKENIZER_FILE = "tokenizer.json";
-const CONFIG_FILE = "config.json";
-
-/** Default artifact dir: $TMCT_EMBED_DIR, else <repo>/vendor/embeddings/potion-base-8M
- *  (gitignored via vendor/ — the location scripts/fetch-embeddings.mjs writes). */
-export function defaultEmbeddingsDir() {
-  if (process.env.TMCT_EMBED_DIR) return process.env.TMCT_EMBED_DIR;
-  const here = dirname(fileURLToPath(import.meta.url)); // packages/tmct/src
-  return join(here, "..", "..", "..", "..", "vendor", "embeddings", "potion-base-8M");
-}
-
-// ---- safetensors (hand-rolled: u64le header length + JSON header + raw tensors) -------
-
-/** Read the single 2-D F32 embedding tensor from a safetensors file →
- *  { matrix: Float32Array (row-major), rows, dim }. model2vec exports exactly one
- *  tensor named "embeddings"; any lone 2-D F32 tensor is accepted for test fixtures. */
-function readSafetensors(file) {
-  const buf = readFileSync(file);
-  const headerLen = Number(buf.readBigUInt64LE(0));
-  const header = JSON.parse(buf.subarray(8, 8 + headerLen).toString("utf8"));
-  const name = header.embeddings
-    ? "embeddings"
-    : Object.keys(header).find((k) => k !== "__metadata__" && header[k]?.shape?.length === 2);
-  const t = name && header[name];
-  if (!t) throw new Error(`no 2-D tensor in ${file}`);
-  if (t.dtype !== "F32") throw new Error(`unsupported dtype ${t.dtype} in ${file} (only F32)`);
-  const [rows, dim] = t.shape;
-  const [start, end] = t.data_offsets;
-  const bytes = buf.subarray(8 + headerLen + start, 8 + headerLen + end);
-  if (bytes.byteLength !== rows * dim * 4) throw new Error(`tensor size mismatch in ${file}`);
-  // Copy into a fresh ArrayBuffer: the slice's byteOffset inside the file buffer is not
-  // guaranteed 4-byte aligned, and Float32Array requires alignment.
-  const matrix = new Float32Array(rows * dim);
-  new Uint8Array(matrix.buffer).set(bytes);
-  return { matrix, rows, dim };
-}
-
-// ---- Bert-style WordPiece tokenizer (from tokenizer.json) -----------------------------
-
-const isPunct = (ch) => {
-  const c = ch.codePointAt(0);
-  // ASCII punctuation ranges (Bert treats these as standalone tokens) + general unicode P/S.
-  return (c >= 33 && c <= 47) || (c >= 58 && c <= 64) || (c >= 91 && c <= 96) || (c >= 123 && c <= 126) ||
-    /[\p{P}\p{S}]/u.test(ch);
-};
-const isCjk = (ch) => {
-  const c = ch.codePointAt(0);
-  return (c >= 0x4e00 && c <= 0x9fff) || (c >= 0x3400 && c <= 0x4dbf) ||
-    (c >= 0xf900 && c <= 0xfaff) || (c >= 0x20000 && c <= 0x2ffff);
-};
-
-function makeTokenizer(spec) {
-  if (spec?.model?.type !== "WordPiece") {
-    throw new Error(`unsupported tokenizer model "${spec?.model?.type}" (only WordPiece)`);
-  }
-  const vocab = spec.model.vocab; // token -> id
-  const contPrefix = spec.model.continuing_subword_prefix ?? "##";
-  const maxChars = spec.model.max_input_chars_per_word ?? 100;
-  const unkId = vocab[spec.model.unk_token] ?? null;
-  const lowercase = spec.normalizer?.lowercase !== false;
-
-  // BertNormalizer: clean control chars, pad CJK, lowercase (+ strip accents when lowercasing).
-  const normalize = (text) => {
-    let s = String(text).replace(/[\u0000\ufffd]/g, "").replace(/[\u0001-\u0008\u000b\u000c\u000e-\u001f\u007f]/g, " ");
-    s = [...s].map((ch) => (isCjk(ch) ? ` ${ch} ` : ch)).join("");
-    if (lowercase) s = s.toLowerCase().normalize("NFD").replace(/\p{Mn}/gu, "");
-    return s;
-  };
-  // BertPreTokenizer: split on whitespace; every punctuation char is its own word.
-  const preTokenize = (s) => {
-    const words = [];
-    for (const chunk of s.split(/\s+/)) {
-      if (!chunk) continue;
-      let cur = "";
-      for (const ch of chunk) {
-        if (isPunct(ch)) {
-          if (cur) { words.push(cur); cur = ""; }
-          words.push(ch);
-        } else cur += ch;
-      }
-      if (cur) words.push(cur);
-    }
-    return words;
-  };
-  // Greedy longest-match WordPiece per word; a word with no valid segmentation → [UNK].
-  const wordPiece = (word) => {
-    if (word.length > maxChars) return unkId == null ? [] : [unkId];
-    const ids = [];
-    let start = 0;
-    while (start < word.length) {
-      let end = word.length;
-      let id = null;
-      while (end > start) {
-        const piece = (start > 0 ? contPrefix : "") + word.slice(start, end);
-        if (vocab[piece] !== undefined) { id = vocab[piece]; break; }
-        end--;
-      }
-      if (id == null) return unkId == null ? [] : [unkId];
-      ids.push(id);
-      start = end;
-    }
-    return ids;
-  };
-  // No [CLS]/[SEP] template — model2vec pools content subwords only.
-  return (text) => preTokenize(normalize(text)).flatMap(wordPiece);
-}
-
-// ---- embedder ---------------------------------------------------------------------------
-
-const LOADED = new Map(); // dir -> embedder | null (per-process cache; weights load once)
-
-/** Load tokenizer + embedding matrix from `dir` (default: defaultEmbeddingsDir()).
- *  Returns null — silently — when the artifacts are absent (the one-time
- *  `npm run refs:embeddings` fetch has not run): callers no-op, never fail.
- *  The returned embedder is { dim, dir, embed(text) → L2-normalised Float32Array }. */
-export function loadEmbedder({ dir = defaultEmbeddingsDir() } = {}) {
-  if (LOADED.has(dir)) return LOADED.get(dir);
-  const modelFile = join(dir, MODEL_FILE);
-  const tokFile = join(dir, TOKENIZER_FILE);
-  if (!existsSync(modelFile) || !existsSync(tokFile)) {
-    LOADED.set(dir, null);
-    return null;
-  }
-  const { matrix, rows, dim } = readSafetensors(modelFile);
-  const tokenize = makeTokenizer(JSON.parse(readFileSync(tokFile, "utf8")));
-  let cfg = {};
-  try { cfg = JSON.parse(readFileSync(join(dir, CONFIG_FILE), "utf8")); } catch { /* optional */ }
-  const doNormalize = cfg.normalize !== false;
-
-  const embed = (text) => {
-    const ids = tokenize(text);
-    const v = new Float32Array(dim);
-    if (!ids.length) return v; // zero vector: cosine 0 against everything
-    for (const id of ids) {
-      if (id < 0 || id >= rows) continue;
-      const off = id * dim;
-      for (let j = 0; j < dim; j++) v[j] += matrix[off + j];
-    }
-    for (let j = 0; j < dim; j++) v[j] /= ids.length; // mean pool
-    if (doNormalize) {
-      let norm = 0;
-      for (let j = 0; j < dim; j++) norm += v[j] * v[j];
-      norm = Math.sqrt(norm);
-      if (norm > 0) for (let j = 0; j < dim; j++) v[j] /= norm;
-    }
-    return v;
-  };
-  const embedder = { dim, dir, embed };
-  LOADED.set(dir, embedder);
-  return embedder;
-}

package/src/domain/vector.mjs

@@ -1,12 +0,0 @@
-// vector.mjs — pure vector arithmetic over embedding vectors. No model, no fs:
-// the loader that reads weights off disk lives in src/adapters/embed.mjs.
-
-/** Cosine similarity. Over L2-normalised vectors this is just the dot product, but the
- *  full form is kept so unnormalised test fixtures behave. 0 when either vector is zero. */
-export function cosine(a, b) {
-  let dot = 0, na = 0, nb = 0;
-  const n = Math.min(a.length, b.length);
-  for (let i = 0; i < n; i++) { dot += a[i] * b[i]; na += a[i] * a[i]; nb += b[i] * b[i]; }
-  if (na === 0 || nb === 0) return 0;
-  return dot / (Math.sqrt(na) * Math.sqrt(nb));
-}