Skip to content

fix(memory): harden embedding reindex recovery - #1966

Merged
yyhhyyyyyy merged 3 commits into
devfrom
fix/embedding-batch-limit-reindex-feedback
Jul 14, 2026
Merged

fix(memory): harden embedding reindex recovery#1966
yyhhyyyyyy merged 3 commits into
devfrom
fix/embedding-batch-limit-reindex-feedback

Conversation

@yyhhyyyyyy

@yyhhyyyyyy yyhhyyyyyy commented Jul 14, 2026

Copy link
Copy Markdown
Collaborator
  • enforce provider embedding batch limits with bounded fallback
  • propagate abort signals through embedding and dimension requests
  • eliminate reindex drain and warmup result races
  • redact sensitive errors and localize rebuild failure feedback

Summary by CodeRabbit

  • New Features
    • Added enhanced memory reindexing status tracking (completed/blocked/aborted) with user-visible failure details.
    • Added a Memory Diagnostics banner for failed reindex runs, including one-click retry when the failure is retryable.
    • Added localized messaging for incomplete vector rebuilding and internal reindex failure reasons.
  • Bug Fixes
    • Improved embedding batching for large embedding requests by adapting to provider limits.
    • Added request cancellation support for embedding and dimension operations across the embedding pipeline.
    • Preserve memory content and fall back to keyword search when vector rebuilding is incomplete.
    • Clear stale reindex status when memories are removed.

@coderabbitai

coderabbitai Bot commented Jul 14, 2026

Copy link
Copy Markdown
Contributor

Review Change Stack

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 6f2cea96-7aa8-4e62-95b4-40bf3d3bfed7

📥 Commits

Reviewing files that changed from the base of the PR and between 85389f2 and c410225.

📒 Files selected for processing (6)
  • src/main/presenter/llmProviderPresenter/providers/aiSdkProvider.ts
  • src/main/presenter/memoryPresenter/infra/embeddingPipeline.ts
  • src/renderer/settings/components/MemoryDiagnosticsPanel.vue
  • test/main/presenter/llmProviderPresenter/openAICompatibleProvider.test.ts
  • test/main/presenter/memory/embeddingPipeline.test.ts
  • test/renderer/components/MemoryDiagnosticsPanel.test.ts
🚧 Files skipped from review as they are similar to previous changes (5)
  • src/renderer/settings/components/MemoryDiagnosticsPanel.vue
  • test/renderer/components/MemoryDiagnosticsPanel.test.ts
  • src/main/presenter/llmProviderPresenter/providers/aiSdkProvider.ts
  • test/main/presenter/memory/embeddingPipeline.test.ts
  • src/main/presenter/memoryPresenter/infra/embeddingPipeline.ts

📝 Walkthrough

Walkthrough

Embedding operations now support cancellation and adaptive batch sizing. Memory reindexing records structured outcomes and sanitized errors, exposes them through status APIs, and displays localized retry diagnostics in the settings panel.

Changes

Embedding runtime and cancellation

Layer / File(s) Summary
Adaptive embedding batching and provider cancellation
src/main/presenter/llmProviderPresenter/..., src/main/presenter/index.ts, src/shared/types/presenters/*
Embedding calls propagate AbortSignal, use bounded batching, learn provider/model-specific limits, retry batch-size failures, and support cancellation across provider routes.
Runtime and batch-limit validation
test/main/presenter/llmProviderPresenter/*
Tests cover learned limits, TTL refresh, concurrency, batching, retry probing, error classification, and abort handling.

Memory reindex lifecycle

Layer / File(s) Summary
Structured reindex state and pipeline outcomes
src/main/presenter/memoryPresenter/domain/types.ts, src/main/presenter/memoryPresenter/infra/embeddingPipeline.ts, src/shared/contracts/routes/memory.routes.ts
Reindex drains and warmups return structured completed, blocked, or aborted results, persist per-agent status, and clear state during cleanup.
Management and error normalization
src/main/presenter/memoryPresenter/services/managementService.ts, src/main/presenter/memoryPresenter/infra/reindexResult.ts, test/main/presenter/memory/*
Management status includes the latest reindex result; failures are sanitized, classified for retryability, and covered across rebuild and cleanup scenarios.

Diagnostics UI

Layer / File(s) Summary
Failure banner and localization
src/renderer/settings/components/MemoryDiagnosticsPanel.vue, src/renderer/src/i18n/*/settings.json, test/renderer/components/MemoryDiagnosticsPanel.test.ts
The diagnostics panel displays localized reindex failure reasons and a retry action, hiding the banner during active or completed rebuilds.

Estimated code review effort: 5 (Critical) | ~120 minutes

Sequence Diagram(s)

sequenceDiagram
  participant MemoryPresenter
  participant LLMProviderPresenter
  participant AiSdkProvider
  participant EmbeddingPipeline
  participant MemoryDiagnosticsPanel
  MemoryPresenter->>LLMProviderPresenter: request embeddings with AbortSignal
  LLMProviderPresenter->>AiSdkProvider: forward cancellation signal
  AiSdkProvider->>EmbeddingPipeline: provide embedding results or errors
  EmbeddingPipeline->>EmbeddingPipeline: record reindex outcome
  EmbeddingPipeline->>MemoryDiagnosticsPanel: expose lastReindex through status
  MemoryDiagnosticsPanel->>MemoryPresenter: request retry
Loading

Possibly related PRs

Suggested reviewers: zerob13

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name Status Explanation Resolution
Docstring Coverage ⚠️ Warning Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Write docstrings for the functions missing them to satisfy the coverage threshold.
✅ Passed checks (4 passed)
Check name Status Explanation
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check ✅ Passed The title clearly matches the main memory embedding reindex recovery changes and is concise enough for a commit history scan.
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.
✨ Finishing Touches
📝 Generate docstrings
  • Create stacked PR
  • Commit on current branch
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch fix/embedding-batch-limit-reindex-feedback

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@coderabbitai coderabbitai Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actionable comments posted: 2

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)
src/main/presenter/llmProviderPresenter/providers/aiSdkProvider.ts (1)

1196-1209: 🩺 Stability & Availability | 🟠 Major | ⚡ Quick win

Honor cancellation before returning hard-coded dimensions.

These early returns never observe an already-aborted signal, so a cancelled reindex can continue past dimension resolution. Call signal?.throwIfAborted() before the model-specific branches.

Proposed fix
   public async runDimensions(modelId: string, signal?: AbortSignal): Promise<LLM_EMBEDDING_ATTRS> {
+    signal?.throwIfAborted()
+
     if (modelId === 'text-embedding-3-small' || modelId === 'text-embedding-ada-002') {
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/main/presenter/llmProviderPresenter/providers/aiSdkProvider.ts` around
lines 1196 - 1209, Update runDimensions to call signal?.throwIfAborted() before
the model-specific hard-coded dimension branches, ensuring an already-aborted
operation stops before either early return.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@src/main/presenter/memoryPresenter/infra/embeddingPipeline.ts`:
- Around line 949-952: Sanitize result.error before logging it in the failed
branch of the warmup flow. Reuse the existing toMemoryReindexError conversion to
redact provider credentials, then log the sanitized error instead of
String(result.error), while preserving clear agent context and the vector-store
readiness reset.
- Around line 1234-1241: Update the invalid-dimension error construction in the
embedding pipeline to preserve the stable `embedding-invalid` code even when
attrs.errorMsg is provided. Use the existing MemoryReindexFailure path with the
provider message as its diagnostic detail, while retaining the current fallback
message and invalid-dimension validation behavior.

---

Outside diff comments:
In `@src/main/presenter/llmProviderPresenter/providers/aiSdkProvider.ts`:
- Around line 1196-1209: Update runDimensions to call signal?.throwIfAborted()
before the model-specific hard-coded dimension branches, ensuring an
already-aborted operation stops before either early return.
🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

  • Push a commit to this branch (recommended)
  • Create a new PR with the fixes

ℹ️ Review info
⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 23f28468-8bd9-4476-a909-f0aee42a9705

📥 Commits

Reviewing files that changed from the base of the PR and between e412217 and 5fccf25.

📒 Files selected for processing (44)
  • src/main/presenter/configPresenter/modelCapabilities.ts
  • src/main/presenter/index.ts
  • src/main/presenter/llmProviderPresenter/aiSdk/embeddingBatchLimits.ts
  • src/main/presenter/llmProviderPresenter/aiSdk/runtime.ts
  • src/main/presenter/llmProviderPresenter/baseProvider.ts
  • src/main/presenter/llmProviderPresenter/index.ts
  • src/main/presenter/llmProviderPresenter/managers/embeddingManager.ts
  • src/main/presenter/llmProviderPresenter/providers/aiSdkProvider.ts
  • src/main/presenter/llmProviderPresenter/providers/ollamaProvider.ts
  • src/main/presenter/memoryPresenter/domain/types.ts
  • src/main/presenter/memoryPresenter/index.ts
  • src/main/presenter/memoryPresenter/infra/embeddingPipeline.ts
  • src/main/presenter/memoryPresenter/infra/reindexResult.ts
  • src/main/presenter/memoryPresenter/services/managementService.ts
  • src/renderer/settings/components/MemoryDiagnosticsPanel.vue
  • src/renderer/src/i18n/da-DK/settings.json
  • src/renderer/src/i18n/de-DE/settings.json
  • src/renderer/src/i18n/en-US/settings.json
  • src/renderer/src/i18n/es-ES/settings.json
  • src/renderer/src/i18n/fa-IR/settings.json
  • src/renderer/src/i18n/fr-FR/settings.json
  • src/renderer/src/i18n/he-IL/settings.json
  • src/renderer/src/i18n/id-ID/settings.json
  • src/renderer/src/i18n/it-IT/settings.json
  • src/renderer/src/i18n/ja-JP/settings.json
  • src/renderer/src/i18n/ko-KR/settings.json
  • src/renderer/src/i18n/ms-MY/settings.json
  • src/renderer/src/i18n/pl-PL/settings.json
  • src/renderer/src/i18n/pt-BR/settings.json
  • src/renderer/src/i18n/ru-RU/settings.json
  • src/renderer/src/i18n/tr-TR/settings.json
  • src/renderer/src/i18n/vi-VN/settings.json
  • src/renderer/src/i18n/zh-CN/settings.json
  • src/renderer/src/i18n/zh-HK/settings.json
  • src/renderer/src/i18n/zh-TW/settings.json
  • src/shared/contracts/routes/memory.routes.ts
  • src/shared/types/presenters/core.presenter.d.ts
  • src/shared/types/presenters/llmprovider.presenter.d.ts
  • test/main/presenter/llmProviderPresenter/aiSdkRuntime.test.ts
  • test/main/presenter/llmProviderPresenter/embeddingBatchLimits.test.ts
  • test/main/presenter/memory/embeddingPipeline.test.ts
  • test/main/presenter/memory/managementService.test.ts
  • test/main/presenter/memory/reindexResult.test.ts
  • test/renderer/components/MemoryDiagnosticsPanel.test.ts

Comment thread src/main/presenter/memoryPresenter/infra/embeddingPipeline.ts
Comment thread src/main/presenter/memoryPresenter/infra/embeddingPipeline.ts Outdated
Comment thread src/renderer/settings/components/MemoryDiagnosticsPanel.vue
@yyhhyyyyyy
yyhhyyyyyy merged commit d1879c0 into dev Jul 14, 2026
4 checks passed
@yyhhyyyyyy
yyhhyyyyyy deleted the fix/embedding-batch-limit-reindex-feedback branch July 14, 2026 05:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants