Skip to content

fix(search): shorten direct retrieval and simplify result states - #7809

Merged
icecrasher321 merged 1 commit into
stagingfrom
codex/direct-search-vector-budget
Sep 14, 2026
Merged

icecrasher321 merged 1 commit into
stagingfrom
codex/direct-search-vector-budget

Conversation

@icecrasher321

Copy link
Copy Markdown
Collaborator

Summary

  • Limit direct Search vector retrieval to 3 seconds; keep the 8-second default for Assistant and existing callers. Keyword and tag budgets stay unchanged.
  • Remove partial-result notices and retry controls from returned Search results. Empty searches show “Search found no results.”
  • Keep timeout coverage and the effective vector budget in server diagnostics.

Type of Change

  • Bug fix

Testing

  • 123 targeted unit tests passed, covering route budget selection, retrieval budgets, Search states, and KB block behavior.
  • 15 PostgreSQL integration cases passed, including a delayed vector query timing out in direct Search while completing through Assistant. One optional Go-server case skipped.
  • Typecheck, lint, all 46 CI audits, block-registry and docs-manifest checks passed.
  • Verified populated and empty Search states in the browser.

Checklist

  • Code follows project style guidelines
  • Self-reviewed my changes
  • Tests added/updated and passing
  • No new warnings introduced
  • I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

@vercel

vercel Bot commented Sep 14, 2026

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment
Project Deployment Actions Updated
docs Skipped Skipped Sep 14, 2026 5:33pm UTC

Request Review

@greptile-apps

greptile-apps Bot commented Sep 14, 2026

Copy link
Copy Markdown
Contributor

RetriggerConfidence Score: 5/5

The PR appears safe to merge, with the shortened dashboard vector budget correctly isolated from existing Assistant and workflow behavior.

Summary

  • Propagates the optional vector budget through the application and retrieval layers.
  • Records the effective vector budget in search diagnostics.
  • Simplifies dashboard result states by removing partial-result messaging and retry controls.
  • Adds unit and PostgreSQL integration coverage for budget selection, timeout behavior, diagnostics, and UI states.

Diagram

sequenceDiagram
    participant UI as Knowledge Search UI
    participant Route as Dashboard Search Route
    participant App as Search Application
    participant Retrieval as Hybrid Retrieval
    participant Vector as Vector Leg
    participant Keyword as Keyword/Tag Legs
    participant Diagnostics as Search Diagnostics

    UI->>Route: Search request
    Route->>App: "vectorBudgetMs = 3000"
    App->>Retrieval: Authorized search input
    Retrieval->>Diagnostics: Record effective vector budget
    par Independent retrieval budgets
        Retrieval->>Vector: Search with 3-second deadline
    and
        Retrieval->>Keyword: Search with 8-second deadline
    end
    Vector-->>Retrieval: Results or timeout
    Keyword-->>Retrieval: Results or timeout
    Retrieval-->>App: Results and retrieval status
    App-->>Route: Search response
    Route-->>UI: Documents or empty result state
Loading

Reviews (1) · Last reviewed commit: "fix(search): shorten direct retrieval an..."

@icecrasher321
icecrasher321 merged commit 17e983e into staging Sep 14, 2026
33 checks passed
@icecrasher321
icecrasher321 deleted the codex/direct-search-vector-budget branch September 14, 2026 17:38
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.

1 participant