docs: restore parseNiceFlag docblock displaced by PRUNE_STATUSES; note --status/0d in top-level jobs help

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

src/cli.ts

@@ -2387,7 +2387,7 @@ JOBS (Minions)
   jobs get <id>                       Job details + history
   jobs cancel <id>                    Cancel job
   jobs retry <id>                     Re-queue failed/dead job
-  jobs prune [--older-than 30d]       Clean old jobs
+  jobs prune [--older-than 30d] [--status s,..]  Clean old terminal jobs (0d = no age floor)
   jobs stats                          Job health dashboard
   jobs work [--queue Q]               Start worker daemon (Postgres only)
 

src/commands/jobs.ts

@@ -106,6 +106,21 @@ export function parseMaxRssFlag(args: string[]): number | undefined {
   return parsed;
 }
 
+/** Terminal statuses `jobs prune --status` accepts (PR #2282). Matches what
+ *  queue.prune can safely delete; anything else (waiting/active/…) is live. */
+export const PRUNE_STATUSES = ['completed', 'failed', 'dead', 'cancelled'] as const satisfies readonly MinionJobStatus[];
+
+/** Parse a `--status a,b,c` value into prune statuses. Throws on any value
+ *  outside PRUNE_STATUSES (fail-fast, mirrors parseNiceValue). */
+export function parsePruneStatuses(raw: string): MinionJobStatus[] {
+  const requested = raw.split(',').map(s => s.trim()).filter(Boolean);
+  const invalid = requested.filter(s => !(PRUNE_STATUSES as readonly string[]).includes(s));
+  if (requested.length === 0 || invalid.length > 0) {
+    throw new Error(`--status accepts a comma-separated subset of [${PRUNE_STATUSES.join(', ')}]${invalid.length ? `. Invalid: ${invalid.join(', ')}` : ''}`);
+  }
+  return requested as MinionJobStatus[];
+}
+
 /** Parse `--nice N` (then `GBRAIN_NICE` env). Returns:
  *  - undefined if absent (no priority change — inherit)
  *  - the validated integer in [-20, 19] otherwise
@@ -208,7 +223,9 @@ USAGE
   gbrain jobs get <id>
   gbrain jobs cancel <id>
   gbrain jobs retry <id>
-  gbrain jobs prune [--older-than 30d]
+  gbrain jobs prune [--older-than 30d] [--status completed,failed,dead,cancelled]
+                            (--older-than 0d = no age floor: deletes ALL
+                             matching terminal jobs; pair with --status)
   gbrain jobs delete <id>
   gbrain jobs stats
   gbrain jobs smoke
@@ -600,16 +617,27 @@ HANDLER TYPES (built in)
     case 'prune': {
       const olderThanStr = parseFlag(args, '--older-than') ?? '30d';
       const days = parseInt(olderThanStr, 10);
-      if (isNaN(days) || days <= 0) {
-        console.error('Error: --older-than must be a positive number (days). Example: --older-than 30d');
+      if (isNaN(days) || days < 0) {
+        console.error('Error: --older-than must be a non-negative number (days). Example: --older-than 30d; --older-than 0d removes the age floor (deletes ALL matching terminal jobs).');
         process.exit(1);
       }
+      const statusFlag = parseFlag(args, '--status');
+      let statuses: MinionJobStatus[] | undefined;
+      if (statusFlag !== undefined) {
+        try { statuses = parsePruneStatuses(statusFlag); }
+        catch (e) { console.error(`Error: ${e instanceof Error ? e.message : String(e)}`); process.exit(1); }
+      }
 
       try { await queue.ensureSchema(); }
       catch (e) { console.error(e instanceof Error ? e.message : String(e)); process.exit(1); }
 
-      const count = await queue.prune({ olderThan: new Date(Date.now() - days * 86400000) });
-      console.log(`Pruned ${count} jobs older than ${days} days.`);
+      const count = await queue.prune({
+        olderThan: new Date(Date.now() - days * 86400000),
+        ...(statuses ? { status: statuses } : {}),
+      });
+      const statusLabel = statuses ? statuses.join('+') : 'completed+dead+cancelled';
+      const ageLabel = days === 0 ? 'regardless of age' : `older than ${days} days`;
+      console.log(`Pruned ${count} ${statusLabel} jobs ${ageLabel}.`);
       break;
     }
 

src/core/embedding-dim-check.ts

@@ -14,7 +14,7 @@
  */
 
 import type { BrainEngine } from './engine.ts';
-import { PGVECTOR_HNSW_VECTOR_MAX_DIMS, hnswMaxDimsForType } from './vector-index.ts';
+import { PGVECTOR_HNSW_VECTOR_MAX_DIMS } from './vector-index.ts';
 import { gbrainPath } from './config.ts';
 import { resolveRecipe } from './ai/model-resolver.ts';
 import type { Recipe } from './ai/types.ts';
@@ -609,17 +609,6 @@ export function buildFactsAlterRecipe(
   const opclass = columnType === 'halfvec' ? 'halfvec_cosine_ops' : 'vector_cosine_ops';
   const targetType = columnType === 'halfvec' ? `halfvec(${configuredDims})` : `vector(${configuredDims})`;
   const dimsChanged = columnDims !== configuredDims;
-  const hnswMaxDims = hnswMaxDimsForType(columnType);
-  const indexLines = configuredDims <= hnswMaxDims
-    ? [
-        `CREATE INDEX idx_facts_embedding_hnsw`,
-        `  ON facts USING hnsw (embedding ${opclass})`,
-        `  WHERE embedding IS NOT NULL AND expired_at IS NULL;`,
-      ]
-    : [
-        `-- Skip reindex. ${columnType}(${configuredDims}) exceeds pgvector's HNSW cap of ${hnswMaxDims};`,
-        `-- fact similarity falls back to exact scans.`,
-      ];
   return [
     `-- ALTER ${columnType}(${columnDims}) → ${columnType}(${configuredDims}) on indexed column.`,
     `-- HOLD a maintenance window: this rewrites every row's embedding.`,
@@ -640,7 +629,9 @@ export function buildFactsAlterRecipe(
       : []),
     `ALTER TABLE facts ALTER COLUMN embedding TYPE ${targetType}`,
     `  USING embedding::${targetType};`,
-    ...indexLines,
+    `CREATE INDEX idx_facts_embedding_hnsw`,
+    `  ON facts USING hnsw (embedding ${opclass})`,
+    `  WHERE embedding IS NOT NULL AND expired_at IS NULL;`,
   ].join('\n');
 }
 

src/core/migrate.ts

@@ -1,7 +1,6 @@
 import type { BrainEngine } from './engine.ts';
 import { slugifyPath } from './sync.ts';
 import { getFtsLanguage } from './fts-language.ts';
-import { hnswMaxDimsForType } from './vector-index.ts';
 
 /**
  * Schema migrations — run automatically on initSchema().
@@ -2277,19 +2276,11 @@ export const MIGRATIONS: Migration[] = [
         useHalfvec = true;
       }
 
-      const columnType = useHalfvec ? 'halfvec' : 'vector';
-      const vecType = columnType.toUpperCase();
+      const vecType = useHalfvec ? 'HALFVEC' : 'VECTOR';
       // HNSW operator class must match the column type:
       //   VECTOR(n)  → vector_cosine_ops
       //   HALFVEC(n) → halfvec_cosine_ops
       const opclass = useHalfvec ? 'halfvec_cosine_ops' : 'vector_cosine_ops';
-      const hnswMaxDims = hnswMaxDimsForType(columnType);
-      const factsEmbeddingIndexSql = embeddingDim <= hnswMaxDims
-        ? `CREATE INDEX IF NOT EXISTS idx_facts_embedding_hnsw
-          ON facts USING hnsw (embedding ${opclass})
-          WHERE embedding IS NOT NULL AND expired_at IS NULL;`
-        : `-- idx_facts_embedding_hnsw skipped: pgvector HNSW ${columnType} indexes support
-        -- at most ${hnswMaxDims} dimensions; exact vector scans remain available.`;
       // FK to sources is added in a separate ALTER TABLE rather than inline
       // on the column. Inline `REFERENCES` worked on PGLite but silently
       // got dropped by postgres.js's `unsafe()` multi-statement path on
@@ -2363,7 +2354,9 @@ export const MIGRATIONS: Migration[] = [
           ON facts(source_id, entity_slug)
           WHERE consolidated_at IS NULL AND expired_at IS NULL;
 
-        ${factsEmbeddingIndexSql}
+        CREATE INDEX IF NOT EXISTS idx_facts_embedding_hnsw
+          ON facts USING hnsw (embedding ${opclass})
+          WHERE embedding IS NOT NULL AND expired_at IS NULL;
       `;
 
       await engine.runMigration(40, factsDDL);
@@ -2877,16 +2870,8 @@ export const MIGRATIONS: Migration[] = [
         useHalfvec = true;
       }
 
-      const columnType = useHalfvec ? 'halfvec' : 'vector';
-      const vecType = columnType.toUpperCase();
+      const vecType = useHalfvec ? 'HALFVEC' : 'VECTOR';
       const opclass = useHalfvec ? 'halfvec_cosine_ops' : 'vector_cosine_ops';
-      const hnswMaxDims = hnswMaxDimsForType(columnType);
-      const queryCacheEmbeddingIndexSql = embeddingDim <= hnswMaxDims
-        ? `CREATE INDEX IF NOT EXISTS idx_query_cache_embedding_hnsw
-          ON query_cache USING hnsw (embedding ${opclass})
-          WHERE embedding IS NOT NULL;`
-        : `-- idx_query_cache_embedding_hnsw skipped: pgvector HNSW ${columnType} indexes support
-        -- at most ${hnswMaxDims} dimensions; exact vector scans remain available.`;
 
       const ddl = `
         CREATE TABLE IF NOT EXISTS query_cache (
@@ -2905,7 +2890,9 @@ export const MIGRATIONS: Migration[] = [
         CREATE INDEX IF NOT EXISTS idx_query_cache_source_created
           ON query_cache(source_id, created_at DESC);
 
-        ${queryCacheEmbeddingIndexSql}
+        CREATE INDEX IF NOT EXISTS idx_query_cache_embedding_hnsw
+          ON query_cache USING hnsw (embedding ${opclass})
+          WHERE embedding IS NOT NULL;
       `;
 
       await engine.runMigration(55, ddl);

src/core/vector-index.ts

@@ -17,7 +17,6 @@
 import type { BrainEngine } from './engine.ts';
 
 export const PGVECTOR_HNSW_VECTOR_MAX_DIMS = 2000;
-export const PGVECTOR_HNSW_HALFVEC_MAX_DIMS = 4000;
 
 const CHUNK_EMBEDDING_HNSW_INDEX =
   'CREATE INDEX IF NOT EXISTS idx_chunks_embedding ON content_chunks USING hnsw (embedding vector_cosine_ops);';
@@ -30,10 +29,6 @@ export function chunkEmbeddingIndexSql(dims: number): string {
   ].join('\n');
 }
 
-export function hnswMaxDimsForType(columnType: 'vector' | 'halfvec'): number {
-  return columnType === 'halfvec' ? PGVECTOR_HNSW_HALFVEC_MAX_DIMS : PGVECTOR_HNSW_VECTOR_MAX_DIMS;
-}
-
 export function applyChunkEmbeddingIndexPolicy(sql: string, dims: number): string {
   return sql.replaceAll(CHUNK_EMBEDDING_HNSW_INDEX, chunkEmbeddingIndexSql(dims));
 }

test/embedding-dim-check-facts.test.ts

@@ -122,9 +122,9 @@ describe('buildFactsAlterRecipe', () => {
   });
 
   test('vector recipe uses vector_cosine_ops + vector(N) USING cast', () => {
-    const recipe = buildFactsAlterRecipe(1024, 1536, 'vector');
-    expect(recipe).toContain('vector(1536)');
-    expect(recipe).toContain('USING embedding::vector(1536)');
+    const recipe = buildFactsAlterRecipe(1024, 2048, 'vector');
+    expect(recipe).toContain('vector(2048)');
+    expect(recipe).toContain('USING embedding::vector(2048)');
     expect(recipe).toContain('vector_cosine_ops');
     expect(recipe).not.toContain('halfvec_cosine_ops');
   });
@@ -163,14 +163,6 @@ describe('buildFactsAlterRecipe', () => {
     expect(recipe).not.toContain('UPDATE facts SET embedding = NULL');
     expect(recipe).toContain('USING embedding::vector(1536)');
   });
-
-  test('halfvec recipe skips HNSW rebuild above pgvector cap', () => {
-    const recipe = buildFactsAlterRecipe(1536, 4096, 'halfvec');
-    expect(recipe).toContain('halfvec(4096)');
-    expect(recipe).toContain('Skip reindex');
-    expect(recipe).toContain("exceeds pgvector's HNSW cap of 4000");
-    expect(recipe).not.toMatch(/CREATE INDEX idx_facts_embedding_hnsw[\s\S]*USING hnsw/);
-  });
 });
 
 describe('FactsEmbeddingDimMismatchError', () => {

test/facts-migration-dim.test.ts

@@ -11,7 +11,6 @@
 
 import { describe, test, expect, beforeAll, afterAll } from 'bun:test';
 import { PGLiteEngine } from '../src/core/pglite-engine.ts';
-import { configureGateway, resetGateway } from '../src/core/ai/gateway.ts';
 
 let engine: PGLiteEngine;
 
@@ -94,60 +93,4 @@ describe('migration v45 facts column shape', () => {
     );
     expect(after[0].udt_name).toBe(before[0].udt_name);
   });
-
-});
-
-describe('migration v45/v55 large-dim HNSW policy', () => {
-  let largeDimEngine: PGLiteEngine;
-
-  beforeAll(async () => {
-    configureGateway({
-      embedding_model: 'litellm:custom-4096d',
-      embedding_dimensions: 4096,
-      env: { ...process.env },
-    });
-
-    largeDimEngine = new PGLiteEngine();
-    await largeDimEngine.connect({});
-    await largeDimEngine.initSchema();
-  });
-
-  afterAll(async () => {
-    await largeDimEngine.disconnect();
-    resetGateway();
-  });
-
-  test('4096d init skips unsupported HNSW indexes but keeps vector columns', async () => {
-    const formatRows = await largeDimEngine.executeRaw<{ format_type: string }>(
-      `SELECT format_type(atttypid, atttypmod) AS format_type
-       FROM pg_attribute
-       WHERE attrelid = 'facts'::regclass AND attname = 'embedding'`,
-    );
-    expect(formatRows[0]?.format_type).toMatch(/(halfvec|vector)\(4096\)/);
-
-    const indexRows = await largeDimEngine.executeRaw<{ exists: boolean }>(
-      `SELECT EXISTS (
-         SELECT 1 FROM pg_indexes
-         WHERE tablename = 'facts'
-           AND indexname = 'idx_facts_embedding_hnsw'
-       ) AS exists`,
-    );
-    expect(indexRows[0]?.exists).toBe(false);
-
-    const queryCacheFormatRows = await largeDimEngine.executeRaw<{ format_type: string }>(
-      `SELECT format_type(atttypid, atttypmod) AS format_type
-       FROM pg_attribute
-       WHERE attrelid = 'query_cache'::regclass AND attname = 'embedding'`,
-    );
-    expect(queryCacheFormatRows[0]?.format_type).toMatch(/(halfvec|vector)\(4096\)/);
-
-    const queryCacheIndexRows = await largeDimEngine.executeRaw<{ exists: boolean }>(
-      `SELECT EXISTS (
-         SELECT 1 FROM pg_indexes
-         WHERE tablename = 'query_cache'
-           AND indexname = 'idx_query_cache_embedding_hnsw'
-       ) AS exists`,
-    );
-    expect(queryCacheIndexRows[0]?.exists).toBe(false);
-  }, 60000);
 });

test/jobs-prune-flags.test.ts

@@ -0,0 +1,30 @@
+/**
+ * Unit tests for parsePruneStatuses (PR #2282) — `jobs prune --status` parsing.
+ */
+
+import { describe, test, expect } from 'bun:test';
+import { parsePruneStatuses, PRUNE_STATUSES } from '../src/commands/jobs.ts';
+
+describe('parsePruneStatuses', () => {
+  test('parses a single status', () => {
+    expect(parsePruneStatuses('failed')).toEqual(['failed']);
+  });
+
+  test('parses a comma-separated list with whitespace', () => {
+    expect(parsePruneStatuses(' completed, dead ')).toEqual(['completed', 'dead']);
+  });
+
+  test('accepts every documented terminal status', () => {
+    expect(parsePruneStatuses(PRUNE_STATUSES.join(','))).toEqual([...PRUNE_STATUSES]);
+  });
+
+  test('throws on non-terminal statuses', () => {
+    expect(() => parsePruneStatuses('waiting')).toThrow(/Invalid: waiting/);
+    expect(() => parsePruneStatuses('completed,active')).toThrow(/Invalid: active/);
+  });
+
+  test('throws on empty value', () => {
+    expect(() => parsePruneStatuses('')).toThrow(/comma-separated subset/);
+    expect(() => parsePruneStatuses(',')).toThrow(/comma-separated subset/);
+  });
+});

test/minions.test.ts

@@ -702,6 +702,32 @@ describe('MinionQueue: Prune', () => {
     const count = await queue.prune({ olderThan: new Date(Date.now() + 86400000) }); // future date = prune everything old enough
     expect(count).toBe(1); // only the cancelled one
   });
+
+  // PR #2282: `jobs prune --status` passes an explicit status subset through.
+  test('status filter prunes only the requested terminal statuses', async () => {
+    const cancelled = await queue.add('sync', {});
+    await queue.cancelJob(cancelled.id);
+    const dead = await queue.add('embed', {}, { max_attempts: 1 });
+    await queue.claim('tok1', 30000, 'default', ['embed']);
+    await queue.failJob(dead.id, 'tok1', 'boom', 'dead');
+
+    const count = await queue.prune({ olderThan: new Date(Date.now() + 86400000), status: ['dead'] });
+    expect(count).toBe(1); // only the dead one
+
+    const remaining = await queue.getJobs({ status: 'cancelled' });
+    expect(remaining.length).toBe(1);
+  });
+
+  // PR #2282: `--older-than 0d` = no age floor — olderThan of "now" deletes
+  // terminal jobs that finished moments ago.
+  test('olderThan now (0d semantics) prunes just-terminated jobs', async () => {
+    const job = await queue.add('sync', {});
+    await queue.cancelJob(job.id);
+    await new Promise(r => setTimeout(r, 5)); // ensure updated_at < now
+
+    const count = await queue.prune({ olderThan: new Date() });
+    expect(count).toBe(1);
+  });
 });
 
 // --- Stats (1 test) ---