@yurtsever/capsa 0.1.0-alpha.0 → 0.1.0-alpha.2

package/dist/index.js

@@ -10053,6 +10053,7 @@ function loadSqlite() {
 }
 var DB_DIR = ".capsa";
 var DB_FILE = "index.db";
+var SCHEMA_VERSION = 2;
 function dbPath(projectRoot) {
   return join6(projectRoot, DB_DIR, DB_FILE);
 }
@@ -10069,6 +10070,16 @@ var Store = class {
     this.db.exec("PRAGMA journal_mode = WAL");
     this.migrate();
   }
+  /** True when this index was written by an older schema and must be rebuilt. */
+  needsRebuild() {
+    const row = this.db.prepare("SELECT value FROM meta WHERE key = 'schema'").get();
+    return row?.value !== String(SCHEMA_VERSION);
+  }
+  /** Drop all indexed content (not the log, not decisions) and mark the schema current. */
+  reset() {
+    this.db.exec("DELETE FROM chunks_vec; DELETE FROM chunks_fts; DELETE FROM chunks; DELETE FROM items;");
+    this.db.prepare("INSERT OR REPLACE INTO meta(key, value) VALUES ('schema', ?)").run(String(SCHEMA_VERSION));
+  }
   migrate() {
     this.db.exec(`
       CREATE TABLE IF NOT EXISTS meta (key TEXT PRIMARY KEY, value TEXT NOT NULL);
@@ -10097,7 +10108,10 @@ var Store = class {
       );
       CREATE INDEX IF NOT EXISTS chunks_item ON chunks(item_id);
 
-      CREATE VIRTUAL TABLE IF NOT EXISTS chunks_fts USING fts5(text, tokenize = 'unicode61');
+      -- porter: "enforcement" must find "Enforce". The indexed text is the
+      -- chunk plus its provenance header (title, heading trail) so a section
+      -- called "Acceptance criteria" still knows which ticket it belongs to.
+      CREATE VIRTUAL TABLE IF NOT EXISTS chunks_fts USING fts5(text, tokenize = 'porter unicode61');
 
       CREATE VIRTUAL TABLE IF NOT EXISTS chunks_vec USING vec0(
         chunk_id INTEGER PRIMARY KEY,
@@ -10151,10 +10165,14 @@ var Store = class {
            status = excluded.status, updated_at = excluded.updated_at, tags = excluded.tags,
            indexed_at = strftime('%Y-%m-%dT%H:%M:%fZ','now')`).run(row.id, row.format, row.kind, row.name, row.path, row.rel_path, row.content_hash, row.status, row.updated_at, row.tags);
   }
-  insertChunk(itemId, idx, headingPath, text, embedding) {
+  /**
+   * `text` is what the user sees; `searchText` (chunk + provenance header)
+   * is what FTS and the embedder see. Falls back to `text` when omitted.
+   */
+  insertChunk(itemId, idx, headingPath, text, embedding, searchText = text) {
     const res = this.db.prepare("INSERT INTO chunks (item_id, idx, heading_path, text) VALUES (?, ?, ?, ?)").run(itemId, idx, JSON.stringify(headingPath), text);
     const id = Number(res.lastInsertRowid);
-    this.db.prepare("INSERT INTO chunks_fts (rowid, text) VALUES (?, ?)").run(id, text);
+    this.db.prepare("INSERT INTO chunks_fts (rowid, text) VALUES (?, ?)").run(id, searchText);
     this.db.prepare("INSERT INTO chunks_vec (chunk_id, embedding) VALUES (?, ?)").run(BigInt(id), new Uint8Array(embedding.buffer, embedding.byteOffset, embedding.byteLength));
     return id;
   }
@@ -10221,7 +10239,7 @@ import { relative as relative5 } from "path";
 import { readFile as readFile5, stat as stat3 } from "fs/promises";
 import { basename as basename4, dirname as dirname7, relative as relative3, sep as sep4 } from "path";
 var DOC_DIRS = /* @__PURE__ */ new Set(["docs", "doc", "runbooks", "runbook", "wiki", "adr", "decisions"]);
-var EXCLUDED_DIRS = /* @__PURE__ */ new Set(["tickets", "issues", "plans", "node_modules"]);
+var EXCLUDED_DIRS = /* @__PURE__ */ new Set(["tickets", "issues", "todo", "backlog", "plans", "plan", "roadmap", "milestones", "milestone", "node_modules"]);
 function segments(relPath) {
   return relPath.split(/[\\/]/).filter(Boolean);
 }
@@ -10294,7 +10312,7 @@ function firstHeading(raw2) {
 import { readFile as readFile6, stat as stat4 } from "fs/promises";
 import { basename as basename5, dirname as dirname8, relative as relative4, sep as sep5 } from "path";
 var TICKET_DIRS = /* @__PURE__ */ new Set(["tickets", "issues", "todo", "backlog"]);
-var PLAN_DIRS = /* @__PURE__ */ new Set(["plans", "plan", "roadmap"]);
+var PLAN_DIRS = /* @__PURE__ */ new Set(["plans", "plan", "roadmap", "milestones", "milestone"]);
 var CHECKBOX = /^\s*[-*+]\s+\[( |x|X)\]\s+/gm;
 function segments2(relPath) {
   return relPath.split(/[\\/]/).filter(Boolean);
@@ -10414,6 +10432,10 @@ async function indexProject(projectRoot, embedder, onProgress) {
   const started = Date.now();
   const store = openStore(projectRoot, embedder);
   try {
+    if (store.needsRebuild()) {
+      onProgress?.({ phase: "rebuild" });
+      store.reset();
+    }
     onProgress?.({ phase: "scan" });
     const result = await scan(knowledgeAdapters(), { roots: [projectRoot] });
     const items = result.items.filter((i) => i.scope === "project" && isInside(projectRoot, i.path));
@@ -10431,8 +10453,8 @@ async function indexProject(projectRoot, embedder, onProgress) {
       if (existing)
         store.deleteItem(existing.id);
       const chunks = chunkMarkdown(item.content);
-      const texts = chunks.map((c) => withContext(item, c.headingPath, c.text));
-      const vectors = texts.length ? await embedder.embed(texts) : [];
+      const searchTexts = chunks.map((c) => withContext(item, c.headingPath, c.text));
+      const vectors = searchTexts.length ? await embedder.embed(searchTexts) : [];
       const k = item.knowledge;
       store.upsertItem({
         id: item.id,
@@ -10446,7 +10468,7 @@ async function indexProject(projectRoot, embedder, onProgress) {
         updated_at: k?.updatedAt ?? null,
         tags: k?.tags ? JSON.stringify(k.tags) : null
       });
-      chunks.forEach((c, i) => store.insertChunk(item.id, c.index, c.headingPath, c.text, vectors[i]));
+      chunks.forEach((c, i) => store.insertChunk(item.id, c.index, c.headingPath, c.text, vectors[i], searchTexts[i]));
       indexed++;
       onProgress?.({ phase: "embed", scanned: items.length, indexed, skipped });
     }
@@ -10503,6 +10525,7 @@ var DEFAULT_WEIGHTS = {
     plan: 1.05,
     ticket: 1,
     runbook: 1,
+    doc: 0.95,
     wiki: 0.95,
     commit: 0.85
   }
@@ -10518,7 +10541,9 @@ function recencyFactor(updatedAt, w, now = Date.now()) {
   return w.recencyFloor + (1 - w.recencyFloor) * decay;
 }
 function signalFactor(s, w = DEFAULT_WEIGHTS, now = Date.now()) {
-  return recencyFactor(s.updatedAt, w, now) * w.status[s.status ?? "unknown"] * w.kind[s.kind];
+  const status = w.status[s.status ?? "unknown"] ?? 1;
+  const kind = w.kind[s.kind] ?? 1;
+  return recencyFactor(s.updatedAt, w, now) * status * kind;
 }
 function rrf(lists, k = DEFAULT_WEIGHTS.rrfK) {
   const scores = /* @__PURE__ */ new Map();
@@ -10536,6 +10561,7 @@ async function retrieve(store, embedder, query, options = {}) {
   const maxTokens = options.maxTokens ?? 1500;
   const candidates = options.candidates ?? 30;
   const weights = options.weights ?? DEFAULT_WEIGHTS;
+  const minRelative = options.minRelativeScore ?? 0.5;
   const [qv] = await embedder.embed([query]);
   const vec = qv ? store.vectorSearch(qv, candidates).map((r) => r.chunkId) : [];
   const fts = store.textSearch(query, candidates).map((r) => r.chunkId);
@@ -10549,7 +10575,10 @@ async function retrieve(store, embedder, query, options = {}) {
   const perItem = /* @__PURE__ */ new Map();
   const chunks = [];
   let tokens = 0;
+  const best = scored[0]?.score ?? 0;
   for (const { row, score } of scored) {
+    if (chunks.length > 0 && score < best * minRelative)
+      break;
     const used = perItem.get(row.item_id) ?? 0;
     if (used >= 2)
       continue;
@@ -39984,6 +40013,7 @@ async function runIndexCommand(args) {
   const embedder = ollamaEmbedder({ model: flag(args, "--model") });
   console.error(`capsa: indexing ${root} with ${embedder.model} \u2026`);
   const res = await indexProject(root, embedder, (p) => {
+    if (p.phase === "rebuild") console.error("capsa: index schema changed, rebuilding from scratch \u2026");
     if (p.phase === "embed") process.stderr.write(`\r  embedded ${p.indexed} / ${p.scanned} (skipped ${p.skipped})   `);
   });
   process.stderr.write("\n");

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@yurtsever/capsa",
-  "version": "0.1.0-alpha.0",
+  "version": "0.1.0-alpha.2",
   "description": "Your project's knowledge in one capsule. Local-first project memory and cockpit for AI coding agents: indexes runbooks, tickets, plans and instruction files, serves the smallest relevant context over MCP.",
   "type": "module",
   "license": "MIT",