fix(deps): pin openai below 3 so a fresh install can call the LLM - #468
Merged
Merged
Conversation
openai 3.19.2 (released 2026-09-24) resolves with an httpx 1.0 pre-release and its client raises 'module httpx has no attribute Timeout' on every request, so a fresh 'pip install everos' today fails every LLM and embedding call (memorize and hybrid/vector search 500). Reproduced on macOS and on a stock Windows 11 machine with plain pip; environments from uv.lock (openai 2.36.0) are unaffected, which is why CI never saw it. Pin openai<3 and release 1.4.1 (version bump + CHANGELOG section). Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
openaireleased its 3.x line today (3.19.2, 2026-09-24 00:06 UTC).pip install everos(any version: the pin wasopenai>=1.0.0) now resolves it together with an httpx 1.0 pre-release, and the client raisesAttributeError: module 'httpx' has no attribute 'Timeout'on every request — memorize and hybrid / vector search return 500. Reproduced on macOS and on the stock Windows 11 box with plain pip; the rc/1.4.0 smoke tests only exercised keyword search and the CLI, which is why it slipped. Environments built fromuv.lock(openai 2.36.0) are unaffected, which is why CI never saw it.Fix:
openai>=1.0.0,<3(the 2.x line the lock tests), version 1.4.1, CHANGELOG section for the release page.Area
Verification
everos==1.4.0→ openai 3.19.2 / httpx 1.0.dev6 →AttributeErroronembeddings.create; same venv afterpip install "openai<3"→ openai 2.x / httpx 0.28.1 → the call reaches the network (APIConnectionErroragainst a dead port), i.e. the client works.make lintgreen;uv lockregenerated with the constraint (resolved version unchanged: 2.36.0).🤖 Generated with Claude Code