Was running uv run pytest -m 'not ollama and not qualitative' on an environment without ollama and have several tests fail as a result.
I believe these were the tests (with link to where it calls ollama):
- test/backends/test_tool_calls.py (
|
return MelleaSession(backend=OllamaModelBackend(), ctx=ChatContext()) |
)
- test/core/test_model_output_thunk.py (
|
m = start_session(model_options={ModelOption.MAX_NEW_TOKENS: 5}) |
-- no backend == ollama)
- test/stdlib/test_functional.py (
|
m = start_session(model_options={ModelOption.MAX_NEW_TOKENS: 5}) |
-- no backend == ollama)
- test/stdlib/requirements/test_requirement.py (
|
m = start_session(ctx=ctx) |
-- no backend == ollama, all the tests in this file create their own session so there are multiple instances)
Was running
uv run pytest -m 'not ollama and not qualitative'on an environment without ollama and have several tests fail as a result.I believe these were the tests (with link to where it calls ollama):
mellea/test/backends/test_tool_calls.py
Line 19 in 9cafe05
mellea/test/core/test_model_output_thunk.py
Line 14 in 9cafe05
mellea/test/stdlib/test_functional.py
Line 13 in 9cafe05
mellea/test/stdlib/requirements/test_requirement.py
Line 13 in 9cafe05