test: optimize example test discovery and execution speed #372
Conversation
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
- Add pytest markers to example files for proper categorization - Wrap heavy examples with complex initialization to prevent collection hangs - Fix critical skip logic bug in example conftest (import failure) - Configure qualitative tests to be excluded by default - Update docs Tests now run significantly faster and reliably on constrained systems. Run full suite with 'pytest -m ""'.
Merge ProtectionsYour pull request matches the following merge protections and will not be merged until they are valid. 🟢 Enforce conventional commitWonderful, this rule succeeded.Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/
|
Contributor
|
The PR description has been updated. Please fill out the template for your PR to be reviewed. |
- Modified 58 example files to conditionally import pytest - Allows examples to run standalone without pytest dependency - Maintains pytest marker functionality when run via pytest
Contributor
Author
|
There's a few things to clarify from this PR
|
- USDA website returns 403 Forbidden (likely rate limiting/user-agent filtering) - External network dependency makes test flaky - Marked as qualitative to exclude from default test runs
- Fix vision_openai_examples.py: simplified skip logic, added requirements docstring - Fix langchain_messages.py: enhanced conftest to detect pytest.skip() in subprocess - Fix test_vision_openai.py: added @pytest.mark.qualitative decorator - Configure pytest to run test/ before docs/ for fail-fast behavior All tests now pass/skip/deselect correctly with default configuration.
Contributor
Author
|
Here's an example output run after these fixes, just using the default This was run on a 32GB macbook m1 max - and takes around 4 minutes |
Member
|
Here's the results from a run on my 32GB RAM / 4GB VRAM Thinkpad |
Member
|
I don't think my issues are blocking, so I'm inclined to approve / merge but will wait a bit in case there are other opinions. |
Contributor
Author
-> A parsing error in the check. I'll fix. The URL is set in my environment
We have a number of models used in our tests. I looked at this one for validation as I didn't have it, but actually we'd need to do something on every test to check the model exists. Docs/upfront check can easily get out of date unless we can do it programatically. -> Suggest discussion, then additional issue/pr if action needed and multiple ones like are down to insufficient resources. 4GB vram is low for AI dev (though we'd always want to drive down min requirements for usage). The error at least is clear. A possible mitigation would be more fine grained markers to document memory usage - perhaps resulting in skipping all llm tests? Another alternative might be to allow cpu only (not gpu) -> Similar - let's discuss and plan |
- Remove f-string formatting that converts None to 'None' string - Allows proper None handling by ibm_watsonx_ai library - Fixes WMLClientError when WATSONX_URL env var not set
planetf1
force-pushed
the
feat/optimize-example-test-discovery
branch
from
c43a53b to
1540a3c
Compare
Contributor
Author
|
The issue with watsonx should be fixed now. |
- Add @pytest.mark.watsonx and @pytest.mark.requires_api_key markers - Ensures test is skipped at collection time when WATSONX_* env vars missing - Prevents misleading error messages about 'None' URLs
- Add @pytest.mark.requires_heavy_ram to vision_openai_examples.py - Model requires ~14GB RAM when loaded (too much for 32GB systems) - Updated docstring to document RAM requirement
Member
|
I did not have the qwen model downloaded. Pulled it and trying again just to see what happens |
Contributor
Author
ollama probably intelligently split the layers across gpu and cpu. I have marked the test as large ram, as when I ran it the system was struggling somewhat since the model was ~16GB ram by itself. |
Member
qwen2.5vl:7b ? For me it's only 6GB (we're starting to approach building cross-platform container images levels of fun here 😄 ) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Misc PR
Type of PR
Description
This PR optimizes the test infrastructure to improve test execution speed, reliability, and developer experience. The changes address test collection hangs, improve skip handling, and establish proper test categorization.
Key Changes:
Example Test Discovery Optimization (e1701c1)
@pytest.mark.ollama,@pytest.mark.qualitative, etc.) to 66 example filesdocs/examples/conftest.py(import failure handling)pytest -m ""for full suite)AGENTS.md,README.md,docs/tutorial.md,test/MARKERS_GUIDE.mdStandalone Example Execution (1cd9c7b)
Qualitative Test Marking (c5e36ef)
docs/examples/rag/mellea_pdf.pyas qualitative due to external PDF dependencyTest Failure Fixes (f03581e)
vision_openai_examples.py: Simplified skip logic, added requirements docstringdocs/examples/conftest.py: Detect pytest.skip() exceptions in subprocess stderrtest_vision_openai.py::test_image_block_in_chat: Added@pytest.mark.qualitativedecoratortestpaths = ["test", "docs"]inpyproject.tomlfor fail-fast behaviorImpact:
pytest -m ""Testing