Skip to content

fix(integrations): validate pinecone, qdrant, elasticsearch and zep - #10164

Merged
waleedlatif1 merged 2 commits into
stagingfrom
fix/validate-data-stores
Nov 18, 2026
Merged

waleedlatif1 merged 2 commits into
stagingfrom
fix/validate-data-stores

Conversation

@waleedlatif1

@waleedlatif1 waleedlatif1 commented Nov 18, 2026 •

Copy link
Copy Markdown
Collaborator

Summary

Validates the Pinecone, Qdrant, Elasticsearch, Algolia, Convex, Tinybird, Databricks, Mem0, Zep and RabbitMQ integrations against each vendor's official API reference or machine-readable spec (Pinecone 2025-01 OpenAPI, Mem0 and Zep v3 API references, Databricks SDK service definitions, RabbitMQ management plugin source), and fixes what was wrong.

  • Algolia, Convex, Tinybird, Databricks, Mem0 and RabbitMQ were validated and need no change.
  • Canvas requests for saved workflows are unchanged except for the bug fixes listed below. No block, subBlock, operation, tool, param or output id was renamed or removed, and no OAuth scope was added.
  • safeUrlPathSegment gains a preserveWhitespace option so exact-match IDs (Elasticsearch document IDs, Zep IDs, Pinecone namespaces, Qdrant collections) keep surrounding whitespace while dot segments are still rejected.
  • Regression tests cover the Pinecone fixes and the new option; they fail on the pre-fix code.
  • Pinecone's __default__ namespace alias is documented for newer API versions; this integration pins 2025-01, where it could not be verified, so a blank Search With Text namespace is rejected rather than guessed.
  • Left unchanged and unverified: Zep's Get Context still sends mode and minRating, which the current v3 reference no longer lists (the API ignores unknown query params). RabbitMQ's default exchange is addressed with an empty path segment; the broker also accepts the amq.default alias, but the public reference documents neither, so it was not changed.

Release notes

Pinecone

  • Fetch Vectors with more than one ID now returns every vector. Before, the IDs were sent as one comma-joined ID, so a multi-ID fetch found nothing.
  • Search With Text with no namespace now fails with a clear "namespace is required" error. Before, it searched a namespace literally named undefined and silently returned no results.
  • Search With Text with Rerank Options now works when the rerank query is left out or blank: Pinecone reranks against the search query. Before, Sim filled in the query in a shape Pinecone rejects. A rerank query given as {"text": "..."} is now sent as the plain string Pinecone expects.
  • Namespaces in Upsert Text and Search With Text are URL-encoded exactly as entered, so names with spaces or other special characters reach the right namespace. A namespace of . or .. is rejected before the request is sent.

Qdrant

  • Collection names in Upsert Points and Fetch Points are URL-encoded, matching Search Vector, so names with spaces or other special characters work. A collection name of . or .. is rejected on all three operations before the request is sent.

Elasticsearch

  • Index names and document IDs are now validated and encoded. A document ID of . or .. is rejected before the request is sent. Before, Delete Document with ID .. could send a DELETE to the index itself. Document IDs that contain / now reach the right document. Document IDs are still sent exactly as entered, including surrounding spaces.

Zep

  • Thread and user IDs are now validated and encoded on every operation that puts them in the URL. IDs with / or other special characters reach the right thread or user, and . or .. is rejected before the request is sent. IDs are still sent exactly as entered.

Pinecone fetch sends one ids query param per vector, search text targets
__default__ when no namespace is set and sends rerank.query as a string.
Qdrant, Elasticsearch and Zep path segments are encoded and dot segments
rejected.
@vercel

vercel Bot commented Nov 18, 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 Nov 18, 2026 6:22pm UTC

Request Review

@greptile-apps

greptile-apps Bot commented Nov 18, 2026 •

Copy link
Copy Markdown
Contributor

RetriggerView in GreptileConfidence Score: 5/5

[Medium impact] The reviewed changes appear safe to merge.

Summary

This PR fixes request formats and URL paths for Pinecone, Qdrant, Elasticsearch, and Zep. Both earlier findings are addressed, and no new actionable issues were found.

  • Pinecone fetch sends each requested vector ID separately.
  • Pinecone text requests keep namespaces and rerank values in the right shape.
  • Elasticsearch requests keep index names and document IDs inside their intended paths.
  • Qdrant point requests keep collection names inside their intended paths.
  • Zep requests keep thread and user IDs inside their intended paths.

Reviews (2) · Last reviewed commit: "fix(integrations): keep exact ids in enc..." · Reviewed by Greptile

Comment thread apps/sim/tools/pinecone/search_text.ts Resolved

@cubic-dev-ai cubic-dev-ai 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.

All reported issues were addressed across 23 files

Reply with feedback, questions, or to request a fix.

Turn on auto-fix | Re-trigger cubic

Comment thread apps/sim/tools/pinecone/search_text.ts Resolved
Comment thread apps/sim/tools/qdrant/fetch_points.ts Resolved
Comment thread apps/sim/tools/pinecone/search_text.ts Resolved
Comment thread apps/sim/tools/pinecone/search_text.ts Resolved
Comment thread apps/sim/tools/pinecone/upsert_text.ts Resolved
Comment thread apps/sim/tools/pinecone/upsert_text.ts Resolved
@waleedlatif1

Copy link
Copy Markdown
Collaborator Author

@greptile

@waleedlatif1

Copy link
Copy Markdown
Collaborator Author

@cubic-dev-ai review this PR

@cubic-dev-ai

cubic-dev-ai Bot commented Nov 18, 2026

Copy link
Copy Markdown
Contributor

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

@cubic-dev-ai cubic-dev-ai 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.

1 issue found across 27 files

Confidence score: 3/5

  • delete_document.ts trims params.index, so a value like logs can silently target and delete from logs instead. Preserve the supplied value or reject surrounding whitespace across all four operations.
Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. When an issue isn't valid or won't be fixed in this PR, reply in its thread with the reason and then resolve the thread. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="apps/sim/tools/elasticsearch/delete_document.ts">

<violation number="1" location="apps/sim/tools/elasticsearch/delete_document.ts:81">
P2: This trims `params.index` before encoding, so `logs ` silently targets `logs` (including on delete). Preserve the supplied value or reject surrounding whitespace across all four operations.</violation>
</file>

Reply with feedback, questions, or to request a fix.

Turn on auto-fix | Re-trigger cubic

@waleedlatif1
waleedlatif1 merged commit f3668aa into staging Nov 18, 2026
81 of 82 checks passed
@waleedlatif1
waleedlatif1 deleted the fix/validate-data-stores branch November 18, 2026 20:09

This branch was previously deployed

1 inactive deployment
Preview — ef08ed13 Deployed Nov 18, 2026 by vercel[bot]
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