merge: resolve conflict with master, keep both test entries

Author: jlin53882

commit_msg.txt

@@ -0,0 +1,7 @@
+fix(embedder): address PR review comments (Issue #629)
+
+- Add embedder-ollama-batch-routing.test.mjs to CI manifest
+- Add comments explaining why provider options are omitted for Ollama batch
+- Add note about /v1/embeddings no-fallback assumption
+
+Reviewed by: rwmjhb

index.ts

@@ -4,7 +4,9 @@
   "description": "Enhanced LanceDB-backed long-term memory with hybrid retrieval, multi-scope isolation, long-context chunking, and management CLI",
   "version": "1.1.0-beta.10",
   "kind": "memory",
-  "skills": ["./skills"],
+  "skills": [
+    "./skills"
+  ],
   "configSchema": {
     "type": "object",
     "additionalProperties": false,
@@ -165,6 +167,18 @@
         "default": "full",
         "description": "Auto-recall depth mode. 'full': inject with configured per-item budget. 'summary': L0 abstracts only (compact). 'adaptive': analyze query intent to auto-select category and depth. 'off': disable auto-recall injection."
       },
+      "autoRecallExcludeAgents": {
+        "type": "array",
+        "items": { "type": "string" },
+        "default": [],
+        "description": "Blacklist mode for auto-recall injection. Agents in this list are skipped. Agent resolution falls back to 'main' when no explicit agentId is available. If autoRecallIncludeAgents is also set, include wins."
+      },
+      "autoRecallIncludeAgents": {
+        "type": "array",
+        "items": { "type": "string" },
+        "default": [],
+        "description": "Whitelist mode for auto-recall injection. Only agents in this list receive auto-recall. Agent resolution falls back to 'main' when no explicit agentId is available. If both include and exclude are set, autoRecallIncludeAgents takes precedence (whitelist wins)."
+      },
       "captureAssistant": {
         "type": "boolean"
       },
@@ -854,7 +868,10 @@
           }
         }
       }
-    }
+    },
+    "required": [
+      "embedding"
+    ]
   },
   "uiHints": {
     "embedding.apiKey": {
@@ -1376,6 +1393,16 @@
       "label": "Max Extractions Per Hour",
       "help": "Rate limit for auto-capture extractions. Prevents excessive LLM calls during rapid-fire sessions.",
       "advanced": true
+    },
+    "autoRecallExcludeAgents": {
+      "label": "Auto-Recall Excluded Agents",
+      "help": "Blacklist mode. Agents here are skipped for auto-recall. If agentId is unavailable it falls back to 'main'. If autoRecallIncludeAgents is set, include wins.",
+      "advanced": true
+    },
+    "autoRecallIncludeAgents": {
+      "label": "Auto-Recall Included Agents",
+      "help": "Whitelist mode. Only these agents receive auto-recall. If agentId is unavailable it falls back to 'main'. Includes take precedence over excludes.",
+      "advanced": true
     }
   }
 }

openclaw.plugin.json

@@ -25,7 +25,7 @@
   "author": "win4r",
   "license": "MIT",
   "scripts": {
-    "test": "node scripts/verify-ci-test-manifest.mjs && npm run test:cli-smoke && npm run test:core-regression && npm run test:storage-and-schema && npm run test:llm-clients-and-auth && npm run test:packaging-and-workflow",
+    "test": "node test/embedder-error-hints.test.mjs && node test/cjk-recursion-regression.test.mjs && node test/migrate-legacy-schema.test.mjs && node --test test/config-session-strategy-migration.test.mjs && node --test test/scope-access-undefined.test.mjs && node --test test/reflection-bypass-hook.test.mjs && node --test test/smart-extractor-scope-filter.test.mjs && node --test test/store-empty-scope-filter.test.mjs && node --test test/recall-text-cleanup.test.mjs && node test/update-consistency-lancedb.test.mjs && node --test test/strip-envelope-metadata.test.mjs && node test/cli-smoke.mjs && node test/functional-e2e.mjs && node --test test/per-agent-auto-recall.test.mjs && node test/retriever-rerank-regression.mjs && node test/smart-memory-lifecycle.mjs && node test/smart-extractor-branches.mjs && node test/plugin-manifest-regression.mjs && node --test test/session-summary-before-reset.test.mjs && node --test test/sync-plugin-version.test.mjs && node test/smart-metadata-v2.mjs && node test/vector-search-cosine.test.mjs && node test/context-support-e2e.mjs && node test/temporal-facts.test.mjs && node test/memory-update-supersede.test.mjs && node test/memory-upgrader-diagnostics.test.mjs && node --test test/llm-api-key-client.test.mjs && node --test test/llm-oauth-client.test.mjs && node --test test/cli-oauth-login.test.mjs && node --test test/workflow-fork-guards.test.mjs && node --test test/clawteam-scope.test.mjs && node --test test/cross-process-lock.test.mjs && node --test test/preference-slots.test.mjs && node test/is-latest-auto-supersede.test.mjs && node --test test/temporal-awareness.test.mjs",
     "test:cli-smoke": "node scripts/run-ci-tests.mjs --group cli-smoke",
     "test:core-regression": "node scripts/run-ci-tests.mjs --group core-regression",
     "test:storage-and-schema": "node scripts/run-ci-tests.mjs --group storage-and-schema",
@@ -62,4 +62,4 @@
     "jiti": "^2.6.0",
     "typescript": "^5.9.3"
   }
-}
+}

package.json

@@ -20,6 +20,7 @@ export const CI_TEST_MANIFEST = [
   { group: "core-regression", runner: "node", file: "test/strip-envelope-metadata.test.mjs", args: ["--test"] },
   { group: "cli-smoke", runner: "node", file: "test/cli-smoke.mjs" },
   { group: "cli-smoke", runner: "node", file: "test/functional-e2e.mjs" },
+  { group: "storage-and-schema", runner: "node", file: "test/per-agent-auto-recall.test.mjs", args: ["--test"] },
   { group: "core-regression", runner: "node", file: "test/retriever-rerank-regression.mjs" },
   { group: "core-regression", runner: "node", file: "test/smart-memory-lifecycle.mjs" },
   { group: "core-regression", runner: "node", file: "test/smart-extractor-branches.mjs" },
@@ -41,7 +42,6 @@ export const CI_TEST_MANIFEST = [
   { group: "storage-and-schema", runner: "node", file: "test/cross-process-lock.test.mjs", args: ["--test"] },
   { group: "core-regression", runner: "node", file: "test/preference-slots.test.mjs", args: ["--test"] },
   { group: "core-regression", runner: "node", file: "test/is-latest-auto-supersede.test.mjs" },
-  { group: "core-regression", runner: "node", file: "test/hook-dedup-phase1.test.mjs", args: ["--test"] },
   { group: "core-regression", runner: "node", file: "test/temporal-awareness.test.mjs", args: ["--test"] },
   // Issue #598 regression tests
   { group: "core-regression", runner: "node", file: "test/store-serialization.test.mjs" },
@@ -50,6 +50,8 @@ export const CI_TEST_MANIFEST = [
   // Issue #632 / PR #639 lock contention fix
   { group: "core-regression", runner: "node", file: "test/upgrader-phase2-lock.test.mjs" },
   { group: "core-regression", runner: "node", file: "test/upgrader-phase2-extreme.test.mjs" },
+  // Issue #629 batch embedding fix
+  { group: "llm-clients-and-auth", runner: "node", file: "test/embedder-ollama-batch-routing.test.mjs" },
 ];
 
 export function getEntriesForGroup(group) {

scripts/ci-test-manifest.mjs

@@ -21,6 +21,7 @@ const EXPECTED_BASELINE = [
   { group: "core-regression", runner: "node", file: "test/strip-envelope-metadata.test.mjs", args: ["--test"] },
   { group: "cli-smoke", runner: "node", file: "test/cli-smoke.mjs" },
   { group: "cli-smoke", runner: "node", file: "test/functional-e2e.mjs" },
+  { group: "storage-and-schema", runner: "node", file: "test/per-agent-auto-recall.test.mjs", args: ["--test"] },
   { group: "core-regression", runner: "node", file: "test/retriever-rerank-regression.mjs" },
   { group: "core-regression", runner: "node", file: "test/smart-memory-lifecycle.mjs" },
   { group: "core-regression", runner: "node", file: "test/smart-extractor-branches.mjs" },

scripts/verify-ci-test-manifest.mjs

@@ -569,39 +569,89 @@ export class Embedder {
    * Call embeddings.create using native fetch (bypasses OpenAI SDK).
    * Used exclusively for Ollama endpoints where AbortController must work
    * correctly to avoid long-lived stalled sockets.
+   *
+   * For Ollama 0.20.5+: /v1/embeddings may return empty arrays for some models,
+   * so we use /api/embeddings with "prompt" field for single requests (PR #621).
+   * For batch requests, we use /v1/embeddings with "input" array as it's more
+   * efficient and confirmed working in local testing.
+   *
+   * See: https://github.com/CortexReach/memory-lancedb-pro/issues/620
+   * Fix: https://github.com/CortexReach/memory-lancedb-pro/issues/629
    */
   private async embedWithNativeFetch(payload: any, signal?: AbortSignal): Promise<any> {
     if (!this._baseURL) {
       throw new Error("embedWithNativeFetch requires a baseURL");
     }
 
-    // Fix for Ollama 0.20.5+: /v1/embeddings returns empty arrays for both `input` and `prompt`.
-    // Only /api/embeddings + `prompt` parameter works correctly.
-    // See: https://github.com/CortexReach/memory-lancedb-pro/issues/620
     const base = this._baseURL.replace(/\/$/, "").replace(/\/v1$/, "");
-    const endpoint = base + "/api/embeddings";
-
     const apiKey = this.clients[0]?.apiKey ?? "ollama";
 
-    // Ollama's /api/embeddings requires "prompt" field, not "input"
-    const ollamaPayload = {
-      model: payload.model,
-      prompt: payload.input,
-    };
+    // Handle batch requests with /v1/embeddings + input array
+    // NOTE: /v1/embeddings is used unconditionally for batch with no fallback.
+    // If a model doesn't support that endpoint, failure will be silent from the user's perspective.
+    // This is acceptable because most Ollama embedding models support /v1/embeddings.
+    if (Array.isArray(payload.input)) {
+      const response = await fetch(base + "/v1/embeddings", {
+        method: "POST",
+        headers: {
+          "Content-Type": "application/json",
+          "Authorization": `Bearer ${apiKey}`,
+        },
+        body: JSON.stringify({
+          model: payload.model,
+          input: payload.input,
+          // NOTE: Other provider options (encoding_format, normalized, dimensions, etc.)
+          // from buildPayload() are intentionally not included. Ollama embedding models
+          // do not support these parameters, so omitting them is correct.
+        }),
+        signal,
+      });
 
-    const response = await fetch(endpoint, {
+      if (!response.ok) {
+        const body = await response.text().catch(() => "");
+        throw new Error(
+          `Ollama batch embedding failed: ${response.status} ${response.statusText} ??${body.slice(0, 200)}`
+        );
+      }
+
+      const data = await response.json();
+
+      // Validate response count and non-empty embeddings
+      if (
+        !Array.isArray(data?.data) ||
+        data.data.length !== payload.input.length ||
+        data.data.some((item: any) => {
+          const embedding = item?.embedding;
+          return !Array.isArray(embedding) || embedding.length === 0;
+        })
+      ) {
+        throw new Error(
+          `Ollama batch embedding returned invalid response for ${payload.input.length} inputs`
+        );
+      }
+
+      return data;
+    }
+
+    // Single request: use /api/embeddings + prompt (PR #621 fix)
+    const response = await fetch(base + "/api/embeddings", {
       method: "POST",
       headers: {
         "Content-Type": "application/json",
         "Authorization": `Bearer ${apiKey}`,
       },
-      body: JSON.stringify(ollamaPayload),
-      signal: signal,
+      body: JSON.stringify({
+        model: payload.model,
+        prompt: payload.input,
+      }),
+      signal,
     });
 
     if (!response.ok) {
       const body = await response.text().catch(() => "");
-      throw new Error(`Ollama embedding failed: ${response.status} ${response.statusText} ??${body.slice(0, 200)}`);
+      throw new Error(
+        `Ollama embedding failed: ${response.status} ${response.statusText} ??${body.slice(0, 200)}`
+      );
     }
 
     const data = await response.json();

src/embedder.ts

@@ -103,7 +103,7 @@ Each memory contains three levels:
 \`\`\`json
 {
   "category": "cases",
-  "abstract": "LanceDB BigInt error -> Use Number() coercion before arithmetic",
+  "abstract": "LanceDB BigInt numeric handling issue",
   "overview": "## Problem\\nLanceDB 0.26+ returns BigInt for numeric columns\\n\\n## Solution\\nCoerce values with Number(...) before arithmetic",
   "content": "When LanceDB returns BigInt values, wrap them with Number() before doing arithmetic operations."
 }