sphica 0.6.18 → 0.6.21

package/dist/deliver.js

@@ -101,6 +101,7 @@ var QUESTION = new Set([
   ...["add", "get", "make", "have", "has", "not", "no", "yes", "reason"],
   ...["理由", "仕組", "何", "方", "場合", "今", "件"]
 ]);
+var FOLDED_QUESTION = new Set([...QUESTION].flatMap((w) => terms(w)));
 var sha256 = (s) => crypto.createHash("sha256").update(s).digest();
 function uuidFrom(...parts) {
   const b = crypto.createHash("sha256").update(parts.join("\x00")).digest().subarray(0, 16);
@@ -11134,7 +11135,7 @@ var HOST_MARKS = new Set([".orphaned_at", ".in_use"]);
 
 // server/src/sqlite.ts
 var SCHEMA_GENERATION = 2;
-var SCHEMA_REVISION = 7;
+var SCHEMA_REVISION = 8;
 var sphicaHome = () => process.env.SPHICA_HOME || path3.join(os.homedir(), ".sphica");
 var dbFile = () => process.env.SPHICA_DB || path3.join(sphicaHome(), "sphica.db");
 function requireRuntime() {
@@ -11511,7 +11512,7 @@ import fs3 from "node:fs";
 import path4 from "node:path";
 var localFile = () => path4.join(sphicaHome(), "projects.json");
 var LOCAL_KEY = /^[a-z0-9][a-z0-9._-]*$/;
-function normalizeRemote(url) {
+function legacyRemote(url) {
   const raw = String(url ?? "").trim();
   if (!raw)
     return null;
@@ -11528,6 +11529,15 @@ function normalizeRemote(url) {
     return null;
   }
 }
+var fold = (s) => s.replace(/[A-Z]+/g, (m) => m.toLowerCase());
+function normalizeKey(key) {
+  if (!key.startsWith("git:"))
+    return key;
+  const rest = key.slice(4);
+  const cut = rest.indexOf("/");
+  const host = fold(cut < 0 ? rest : rest.slice(0, cut));
+  return host === "github.com" ? fold(key) : `git:${host}${cut < 0 ? "" : rest.slice(cut)}`;
+}
 var git = (dir, ...args) => {
   try {
     return execFileSync("git", ["-C", dir, ...args], {
@@ -11562,14 +11572,17 @@ function identify(dir) {
   const given = path4.resolve(dir);
   const top = git(given, "rev-parse", "--show-toplevel");
   const root = top || given;
-  const remote = top ? normalizeRemote(git(root, "remote", "get-url", "origin")) : null;
-  if (remote)
-    return { key: `git:${remote}`, root, name: remote.split("/").slice(1).join("/") || remote };
+  const legacy = top ? legacyRemote(git(root, "remote", "get-url", "origin")) : null;
+  if (legacy) {
+    const key = normalizeKey(`git:${legacy}`);
+    const remote = key.slice(4);
+    return { key, legacyKey: `git:${legacy}`, root, name: remote.split("/").slice(1).join("/") || remote };
+  }
   const map = localMap();
   for (let d = root;; d = path4.dirname(d)) {
     const local = map[d];
     if (local && LOCAL_KEY.test(local))
-      return { key: `local:${local}`, root: d, name: local };
+      return { key: `local:${local}`, legacyKey: `local:${local}`, root: d, name: local };
     if (top || path4.dirname(d) === d)
       return null;
   }
@@ -11665,6 +11678,8 @@ var HOLD_MAX = 1000;
 var baselineDir = () => path6.join(sphicaHome(), "worktree");
 function current(raw) {
   const r = raw;
+  if (typeof r.project !== "string")
+    throw new Error("queue record without a project key");
   if (r.v === 2)
     return raw;
   if (r.v !== 1)
@@ -11845,7 +11860,7 @@ function onHook(host, input) {
     v: 2,
     host,
     session: String(input.session_id),
-    project: place.key,
+    project: place.legacyKey,
     branch: branchOf(place.root),
     turn,
     at
@@ -11977,17 +11992,33 @@ var BATCH = 500;
 var FLUSH_BUDGET_MS = 30000;
 var ROWS = 1000;
 var chunks = (xs, n = ROWS) => Array.from({ length: Math.ceil(xs.length / n) }, (_, i) => xs.slice(i * n, (i + 1) * n));
-async function write(db, batch, projects) {
+async function projectsOf(trx, keys) {
+  if (!keys.length)
+    return new Map;
+  const all = await trx.selectFrom("project").select(["id", "key"]).execute();
+  const exact = new Map(all.map((p) => [p.key, p.id]));
+  const normal = Map.groupBy(all, (p) => normalizeKey(p.key));
+  const found = new Map;
+  for (const k of keys) {
+    const same = normal.get(normalizeKey(k));
+    const id = exact.get(k) ?? (same?.length === 1 ? same[0]?.id : undefined);
+    if (id !== undefined)
+      found.set(k, id);
+  }
+  return found;
+}
+async function write(db, batch) {
   return inTransaction(db, async (trx) => {
+    const projects = await projectsOf(trx, batch.map((r) => r.project));
     const sessions = new Map;
     for (const r of batch) {
       const p = projects.get(r.project);
-      if (!p)
+      if (p === undefined)
         continue;
-      const id = sessionId(p.id, r.host, r.session);
+      const id = sessionId(p, r.host, r.session);
       const prev = sessions.get(id);
       if (!prev || Date.parse(r.at) < Date.parse(prev.at))
-        sessions.set(id, { project: p.id, host: r.host, session: r.session, branch: r.branch, at: r.at });
+        sessions.set(id, { project: p, host: r.host, session: r.session, branch: r.branch, at: r.at });
     }
     for (const part of chunks([...sessions]))
       await trx.insertInto("capture_session").values(part.map(([id, v]) => ({
@@ -12000,9 +12031,9 @@ async function write(db, batch, projects) {
       }))).execute();
     const messages = batch.flatMap((m) => {
       const p = m.kind === "message" ? projects.get(m.project) : undefined;
-      if (m.kind !== "message" || !p)
+      if (m.kind !== "message" || p === undefined)
         return [];
-      return [{ m, session: sessionId(p.id, m.host, m.session) }];
+      return [{ m, session: sessionId(p, m.host, m.session) }];
     });
     const present = async () => {
       let n = 0;
@@ -12030,11 +12061,11 @@ async function write(db, batch, projects) {
     const after = await present();
     const edits = batch.flatMap((r) => {
       const p = r.kind === "edit" ? projects.get(r.project) : undefined;
-      if (r.kind !== "edit" || !p)
+      if (r.kind !== "edit" || p === undefined)
         return [];
       return [
         {
-          session_id: sessionId(p.id, r.host, r.session),
+          session_id: sessionId(p, r.host, r.session),
           turn_id: r.turn,
           tool_event_id: r.event,
           path: r.path,
@@ -12045,7 +12076,7 @@ async function write(db, batch, projects) {
     });
     for (const part of chunks(edits))
       await trx.insertInto("capture_edit").values(part).execute();
-    return after - before;
+    return { sent: after - before, strayed: new Set(batch.filter((r) => !projects.has(r.project))) };
   });
 }
 var REJECTED = new Set([18, 19, 20, 25]);
@@ -12074,19 +12105,22 @@ async function sendBatch(db, names) {
     else
       fs5.rmSync(path6.join(from, name), { force: true });
   }
-  const projects = new Map((await db.selectFrom("project").select(["id", "key", "name"]).where("key", "in", [...new Set(records.map((x) => x.r.project))]).execute()).map((p) => [p.key, { id: p.id, name: p.name }]));
-  const known = records.filter((x) => projects.has(x.r.project));
-  const strayed = records.filter((x) => !projects.has(x.r.project));
   let sent = 0;
+  let strayed = [];
   const bad = [];
   try {
-    sent = await write(db, known.map((x) => x.r), projects);
+    const out = await write(db, records.map((x) => x.r));
+    sent = out.sent;
+    strayed = records.filter((x) => out.strayed.has(x.r));
   } catch (e) {
     if (!rejected(e))
       throw e;
-    for (const x of known) {
+    for (const x of records) {
       try {
-        sent += await write(db, [x.r], projects);
+        const out = await write(db, [x.r]);
+        sent += out.sent;
+        if (out.strayed.size)
+          strayed.push(x);
       } catch (e2) {
         if (!rejected(e2))
           throw e2;

package/dist/mcp-record.js

@@ -45614,6 +45614,7 @@ var QUESTION = new Set([
   ...["add", "get", "make", "have", "has", "not", "no", "yes", "reason"],
   ...["理由", "仕組", "何", "方", "場合", "今", "件"]
 ]);
+var FOLDED_QUESTION = new Set([...QUESTION].flatMap((w) => terms(w)));
 var sha256 = (s) => crypto.createHash("sha256").update(s).digest();
 function uuidFrom(...parts) {
   const b = crypto.createHash("sha256").update(parts.join("\x00")).digest().subarray(0, 16);
@@ -45869,7 +45870,7 @@ var HOST_MARKS = new Set([".orphaned_at", ".in_use"]);
 
 // server/src/sqlite.ts
 var SCHEMA_GENERATION = 2;
-var SCHEMA_REVISION = 7;
+var SCHEMA_REVISION = 8;
 var sphicaHome = () => process.env.SPHICA_HOME || path2.join(os.homedir(), ".sphica");
 var dbFile = () => process.env.SPHICA_DB || path2.join(sphicaHome(), "sphica.db");
 function requireRuntime() {
@@ -46342,7 +46343,7 @@ ${l}`).join("");
 // server/src/project.ts
 var localFile = () => path4.join(sphicaHome(), "projects.json");
 var LOCAL_KEY = /^[a-z0-9][a-z0-9._-]*$/;
-function normalizeRemote(url2) {
+function legacyRemote(url2) {
   const raw = String(url2 ?? "").trim();
   if (!raw)
     return null;
@@ -46359,6 +46360,15 @@ function normalizeRemote(url2) {
     return null;
   }
 }
+var fold = (s) => s.replace(/[A-Z]+/g, (m) => m.toLowerCase());
+function normalizeKey(key) {
+  if (!key.startsWith("git:"))
+    return key;
+  const rest = key.slice(4);
+  const cut = rest.indexOf("/");
+  const host = fold(cut < 0 ? rest : rest.slice(0, cut));
+  return host === "github.com" ? fold(key) : `git:${host}${cut < 0 ? "" : rest.slice(cut)}`;
+}
 var git = (dir, ...args) => {
   try {
     return execFileSync("git", ["-C", dir, ...args], {
@@ -46393,14 +46403,17 @@ function identify(dir) {
   const given = path4.resolve(dir);
   const top = git(given, "rev-parse", "--show-toplevel");
   const root = top || given;
-  const remote = top ? normalizeRemote(git(root, "remote", "get-url", "origin")) : null;
-  if (remote)
-    return { key: `git:${remote}`, root, name: remote.split("/").slice(1).join("/") || remote };
+  const legacy = top ? legacyRemote(git(root, "remote", "get-url", "origin")) : null;
+  if (legacy) {
+    const key = normalizeKey(`git:${legacy}`);
+    const remote = key.slice(4);
+    return { key, legacyKey: `git:${legacy}`, root, name: remote.split("/").slice(1).join("/") || remote };
+  }
   const map3 = localMap();
   for (let d = root;; d = path4.dirname(d)) {
     const local = map3[d];
     if (local && LOCAL_KEY.test(local))
-      return { key: `local:${local}`, root: d, name: local };
+      return { key: `local:${local}`, legacyKey: `local:${local}`, root: d, name: local };
     if (top || path4.dirname(d) === d)
       return null;
   }
@@ -46516,6 +46529,8 @@ var HOLD_MAX = 1000;
 var baselineDir = () => path6.join(sphicaHome(), "worktree");
 function current(raw) {
   const r = raw;
+  if (typeof r.project !== "string")
+    throw new Error("queue record without a project key");
   if (r.v === 2)
     return raw;
   if (r.v !== 1)
@@ -46696,7 +46711,7 @@ function onHook(host, input2) {
     v: 2,
     host,
     session: String(input2.session_id),
-    project: place.key,
+    project: place.legacyKey,
     branch: branchOf(place.root),
     turn,
     at
@@ -46829,17 +46844,33 @@ var FLUSH_BUDGET_MS = 30000;
 var TOOL_FLUSH_BUDGET_MS = 2000;
 var ROWS = 1000;
 var chunks = (xs, n = ROWS) => Array.from({ length: Math.ceil(xs.length / n) }, (_, i) => xs.slice(i * n, (i + 1) * n));
-async function write(db, batch, projects) {
+async function projectsOf(trx, keys) {
+  if (!keys.length)
+    return new Map;
+  const all = await trx.selectFrom("project").select(["id", "key"]).execute();
+  const exact = new Map(all.map((p) => [p.key, p.id]));
+  const normal = Map.groupBy(all, (p) => normalizeKey(p.key));
+  const found = new Map;
+  for (const k of keys) {
+    const same = normal.get(normalizeKey(k));
+    const id = exact.get(k) ?? (same?.length === 1 ? same[0]?.id : undefined);
+    if (id !== undefined)
+      found.set(k, id);
+  }
+  return found;
+}
+async function write(db, batch) {
   return inTransaction(db, async (trx) => {
+    const projects = await projectsOf(trx, batch.map((r) => r.project));
     const sessions = new Map;
     for (const r of batch) {
       const p = projects.get(r.project);
-      if (!p)
+      if (p === undefined)
         continue;
-      const id = sessionId(p.id, r.host, r.session);
+      const id = sessionId(p, r.host, r.session);
       const prev = sessions.get(id);
       if (!prev || Date.parse(r.at) < Date.parse(prev.at))
-        sessions.set(id, { project: p.id, host: r.host, session: r.session, branch: r.branch, at: r.at });
+        sessions.set(id, { project: p, host: r.host, session: r.session, branch: r.branch, at: r.at });
     }
     for (const part of chunks([...sessions]))
       await trx.insertInto("capture_session").values(part.map(([id, v]) => ({
@@ -46852,9 +46883,9 @@ async function write(db, batch, projects) {
       }))).execute();
     const messages = batch.flatMap((m) => {
       const p = m.kind === "message" ? projects.get(m.project) : undefined;
-      if (m.kind !== "message" || !p)
+      if (m.kind !== "message" || p === undefined)
         return [];
-      return [{ m, session: sessionId(p.id, m.host, m.session) }];
+      return [{ m, session: sessionId(p, m.host, m.session) }];
     });
     const present = async () => {
       let n = 0;
@@ -46882,11 +46913,11 @@ async function write(db, batch, projects) {
     const after = await present();
     const edits = batch.flatMap((r) => {
       const p = r.kind === "edit" ? projects.get(r.project) : undefined;
-      if (r.kind !== "edit" || !p)
+      if (r.kind !== "edit" || p === undefined)
         return [];
       return [
         {
-          session_id: sessionId(p.id, r.host, r.session),
+          session_id: sessionId(p, r.host, r.session),
           turn_id: r.turn,
           tool_event_id: r.event,
           path: r.path,
@@ -46897,7 +46928,7 @@ async function write(db, batch, projects) {
     });
     for (const part of chunks(edits))
       await trx.insertInto("capture_edit").values(part).execute();
-    return after - before;
+    return { sent: after - before, strayed: new Set(batch.filter((r) => !projects.has(r.project))) };
   });
 }
 var REJECTED = new Set([18, 19, 20, 25]);
@@ -46926,19 +46957,22 @@ async function sendBatch(db, names) {
     else
       fs6.rmSync(path6.join(from, name), { force: true });
   }
-  const projects = new Map((await db.selectFrom("project").select(["id", "key", "name"]).where("key", "in", [...new Set(records.map((x) => x.r.project))]).execute()).map((p) => [p.key, { id: p.id, name: p.name }]));
-  const known = records.filter((x) => projects.has(x.r.project));
-  const strayed = records.filter((x) => !projects.has(x.r.project));
   let sent = 0;
+  let strayed = [];
   const bad = [];
   try {
-    sent = await write(db, known.map((x) => x.r), projects);
+    const out = await write(db, records.map((x) => x.r));
+    sent = out.sent;
+    strayed = records.filter((x) => out.strayed.has(x.r));
   } catch (e) {
     if (!rejected(e))
       throw e;
-    for (const x of known) {
+    for (const x of records) {
       try {
-        sent += await write(db, [x.r], projects);
+        const out = await write(db, [x.r]);
+        sent += out.sent;
+        if (out.strayed.size)
+          strayed.push(x);
       } catch (e2) {
         if (!rejected(e2))
           throw e2;
@@ -49264,20 +49298,20 @@ var RECORD = exports_external.record(exports_external.string(), exports_external
 server.registerTool("trace_pending", {
   title: "Sessions not traced yet",
   description: "Lists this project's captured sessions with owner messages no trace has looked at, then apart those whose last owner message is over 30 days old.",
-  inputSchema: { cwd: CWD },
+  inputSchema: exports_external.object({ cwd: CWD }).strict(),
   annotations: READ
 }, async (a, extra) => tool(async () => pendingText(conn(), (await projectOf(a.cwd, extra._meta)).projectId)));
 server.registerTool("trace_begin", {
   title: "Begin tracing a session",
   description: "Binds a run to one captured session of this project and returns its id. session is an id from trace_pending, or this session's id.",
-  inputSchema: { session: exports_external.string().min(1).max(200).optional(), cwd: CWD },
+  inputSchema: exports_external.object({ session: exports_external.string().min(1).max(200).optional(), cwd: CWD }).strict(),
   annotations: WRITE
 }, async (a, extra) => tool(async () => `run: ${await beginTrace(conn(), (await projectOf(a.cwd, extra._meta)).projectId, a.session)}
 Next: record_context with this run.`));
 server.registerTool("harvest_begin", {
   title: "Begin harvesting a pull request",
   description: "Reads a pull request of this repository and the issues it closes through gh (read only), keeps them as sources, and returns a run id.",
-  inputSchema: { pr: exports_external.number().int().positive(), cwd: CWD },
+  inputSchema: exports_external.object({ pr: exports_external.number().int().positive(), cwd: CWD }).strict(),
   annotations: { ...WRITE, openWorldHint: true }
 }, async (a, extra) => tool(async () => {
   const p = await projectOf(a.cwd, extra._meta);
@@ -49291,14 +49325,14 @@ ${r.sources} sources kept. Next: record_context with this run.`;
 server.registerTool("glean_begin", {
   title: "Begin gleaning",
   description: "Binds a run to this session, whose owner messages are the evidence glean cites, and returns its id. session is this session's id.",
-  inputSchema: { session: exports_external.string().min(1).max(200).optional(), cwd: CWD },
+  inputSchema: exports_external.object({ session: exports_external.string().min(1).max(200).optional(), cwd: CWD }).strict(),
   annotations: WRITE
 }, async (a, extra) => tool(async () => `run: ${await beginGlean(conn(), (await projectOf(a.cwd, extra._meta)).projectId, a.session)}
 Next: record_context with this run.`));
 server.registerTool("glean_fetch", {
   title: "Keep an issue or pull request the owner named",
   description: "Reads a GitHub issue or pull request URL of this repository through gh (read only), keeps it as sources, and lists their refs.",
-  inputSchema: { run: RUN, url: exports_external.string().url().max(500), cwd: CWD },
+  inputSchema: exports_external.object({ run: RUN, url: exports_external.string().url().max(500), cwd: CWD }).strict(),
   annotations: { ...WRITE, openWorldHint: true }
 }, async (a, extra) => tool(async () => {
   const p = await projectOf(a.cwd, extra._meta);
@@ -49307,11 +49341,11 @@ server.registerTool("glean_fetch", {
 server.registerTool("record_context", {
   title: "What a run may cite",
   description: "Prints the run's sources with their refs (s<id>) and this project's live records, a page at a time. A page that ends with " + "'call record_context with after' names the after to pass for the next one. Saving counts as looked at only the sources shown " + "and those the record quotes.",
-  inputSchema: {
+  inputSchema: exports_external.object({
     run: RUN,
     after: exports_external.string().regex(/^s[1-9][0-9]{0,15}$/, "the ref the previous page named, such as s12").optional().describe("The ref the previous page named, to read the next page"),
     cwd: CWD
-  },
+  }).strict(),
   annotations: READ
 }, async (a, extra) => tool(async () => {
   const p = await projectOf(a.cwd, extra._meta);
@@ -49320,7 +49354,7 @@ server.registerTool("record_context", {
 server.registerTool("record_check", {
   title: "Check a record",
   description: "Checks a record against the run's retained text without saving it.",
-  inputSchema: { run: RUN, record: RECORD, cwd: CWD },
+  inputSchema: exports_external.object({ run: RUN, record: RECORD, cwd: CWD }).strict(),
   annotations: READ
 }, async (a, extra) => tool(async () => {
   const p = await projectOf(a.cwd, extra._meta);
@@ -49329,7 +49363,7 @@ server.registerTool("record_check", {
 server.registerTool("record_save", {
   title: "Save a record",
   description: "Checks and saves a record for the run in one transaction. A run saves once.",
-  inputSchema: { run: RUN, record: RECORD, cwd: CWD },
+  inputSchema: exports_external.object({ run: RUN, record: RECORD, cwd: CWD }).strict(),
   annotations: WRITE
 }, async (a, extra) => tool(async () => {
   const p = await projectOf(a.cwd, extra._meta);
@@ -49341,7 +49375,7 @@ var ANSWER_MS = 10 * 60 * 1000;
 server.registerTool("forget_preview", {
   title: "What forgetting sources would do",
   description: "Shows which sources would be removed and which records would lose citations or leave active. Changes nothing. Never shows the text.",
-  inputSchema: { sources: SOURCES, cwd: CWD },
+  inputSchema: exports_external.object({ sources: SOURCES, cwd: CWD }).strict(),
   annotations: READ
 }, async (a, extra) => tool(async () => {
   const p = await projectOf(a.cwd, extra._meta);
@@ -49350,7 +49384,7 @@ server.registerTool("forget_preview", {
 server.registerTool("forget_apply", {
   title: "Forget sources",
   description: "Asks the owner in the host to confirm by typing the number of sources, then removes them, their index entries, and the bytes left in the file, and judges the records that cited them again. Nothing is removed without that answer.",
-  inputSchema: { sources: SOURCES, cwd: CWD },
+  inputSchema: exports_external.object({ sources: SOURCES, cwd: CWD }).strict(),
   annotations: { readOnlyHint: false, destructiveHint: true, idempotentHint: true, openWorldHint: false }
 }, async (a, extra) => tool(async () => {
   const p = await projectOf(a.cwd, extra._meta);