weiqingy commented on PR #686: URL: https://github.com/apache/flink-agents/pull/686#issuecomment-4471933671
@wenjin272 Thanks for catching this! Confirmed by inspecting both scripts side-by-side: the chat-model sibling at `python/flink_agents/integrations/chat_models/tests/start_ollama_server.sh` only does `ollama pull` and nothing else, while the embedding version has the redundant `ollama run all-minilm:22m` line. As you diagnosed, `ollama run` against an embedding model exits non-zero with `Error: embedding models require input text`, and `subprocess.run(..., check=True)` in `test_ollama_embedding_model.py:37–55` swallows that into the `except` block, leaving `client = None` and silently skipping the test — defeating the gate's purpose. Dropped the line in `3ade7fa`. The preceding `ollama pull all-minilm:22m` is sufficient to make the model discoverable via `client.list()`, which is what the gate actually checks. The next CI run should exercise `test_ollama_embedding_setup` on the Python 3.10 / Ubuntu UT job for real. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
