FormattingFailedWarning | InspectArgumentsFailedWarning while running pytest in containerized env

[theMladyPan]

Description

Running pytest it inside docker container, throws these warnings:

tests-1  | tests/test_title_endpoint.py::test_title_auth_should_fail
tests-1  |   /home/###/<repo>/tests/test_title_endpoint.py:16: FormattingFailedWarning: 
tests-1  |       Ensure you are either:
tests-1  |         (1) passing an f-string directly, with inspect_arguments enabled and working, or
tests-1  |         (2) passing a literal `str.format`-style template, not a preformatted string.
tests-1  |       See https://logfire.pydantic.dev/docs/guides/onboarding-checklist/add-manual-tracing/#messages-and-span-names.
tests-1  |       The problem was: The field {"detail"} is not defined.
 
# OR:

tests-1  | tests/test_title_endpoint.py::test_generate_title_conversation_not_exists
tests-1  |   /home/###/<repo>/tests/test_title_endpoint.py:63: InspectArgumentsFailedWarning: Failed to introspect calling code. Please report this issue to Logfire. Falling back to normal message formatting which may result in loss of information if using an f-string. Set inspect_arguments=False in logfire.configure() to suppress this warning. The problem was:
tests-1  |   No source code available. This happens when running in an interactive shell, using exec(), or running .pyc files without the source .py files.

but running pytest in .venv using these commands:

uv sync --group dev
uv run pytest

pass perfectly fine.

this one specificaly was generated here:
logfire.debug(f"Response status code: {response.status_code}, response text: {response.text}")

I configure the logging inside this fixture

@pytest.fixture(scope="session", autouse=True)
def logging():
    logfire.configure(
        token=settings.logfire_token,
        send_to_logfire="if-token-present",
        distributed_tracing=False,
        service_name="pytest",
        environment="test",
    )

    with logfire.span(f"Test run: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}"):
        yield  # tests will run here

pyproject.toml:

[project]
name = "assistant"
version = "0.2.1"
description = "Main assistant core logic"
readme = "README.md"
requires-python = ">=3.12"
dependencies = [
    "cachetools>=6.2.0",
    "cryptography>=45.0.6",
    "elevenlabs>=2.8.1",
    "fastapi>=0.115.14",
    "logfire[fastapi,httpx,sqlalchemy]>=3.21.2",
    "numpy>=2.3.1",
    "openai>=1.93.0",
    "pgvector>=0.4.1",
    "psycopg2-binary>=2.9.10",
    "pydantic-settings>=2.10.1",
    "pyyaml>=6.0.2",
    "sse-starlette>=2.4.1",
    "sqlmodel>=0.0.24",
    "uvicorn>=0.35.0",
    "pydantic-ai-slim[mcp]==1.0.2",
    "pyjwt>=2.10.1",
    "pydantic-evals[logfire]>=1.0.2",
]

[tool.pytest.ini_options]
addopts = [
    "--verbose",
    "-ra",
    "--color=yes",
    "tests/",
]
pythonpath = [
    "."
]
doctest_optionflags = [
    "NORMALIZE_WHITESPACE",
    "NUMBER",
    "ELLIPSIS",
    "IGNORE_EXCEPTION_DETAIL",
]

[tool.ruff]
line-length = 120
target-version = "py312"

[tool.ruff.lint]
ignore = ["E501"]
select = ["E", "F", "I", "UP"]

[tool.ty.rules]
unresolved-attribute = "ignore"
unknown-argument = "ignore"

[tool.commitizen]
name = "cz_conventional_commits"
tag_format = "v$version"
version_scheme = "pep440"
version_provider = "pep621"
update_changelog_on_bump = true

[dependency-groups]
dev = [
    "pre-commit>=4.2.0",
    "pytest>=8.4.1",
    "ruff>=0.12.0",
    "commitizen>=4.8.3",
    "pytest-tornasync>=0.6.0.post2",
    "ty>=0.0.1a19",
    "pytest-cov>=6.2.1",
    "pytest-asyncio>=1.2.0",
]

Dockerfile:

FROM python:3.12-slim-trixie
COPY --from=ghcr.io/astral-sh/uv:latest /uv /uvx /bin/

ENV PYTHONUNBUFFERED=1
ENV PYTHONDONTWRITEBYTECODE=1

WORKDIR /app  

COPY jwt-rsa-public.pem jwt-rsa-public.pem
COPY mcp_config.json mcp_config.json  

COPY ../pyproject.toml pyproject.toml
COPY ../uv.lock uv.lock
RUN uv sync --locked --group dev

COPY ../lib/ lib/
COPY main.py main.py
COPY ../src/ src/
COPY ../tests/ tests/

CMD ["uv", "run", "pytest"]

Python, Logfire & OS Versions, related packages (not required)

❯ logfire info
logfire="4.10.0"
platform="Linux-6.15.5-061505-generic-x86_64-with-glibc2.41"
python="3.12.11 (main, Jun 12 2025, 12:40:51) [Clang 20.1.4 ]"
[related_packages]
requests="2.32.5"
pydantic="2.11.9"
fastapi="0.118.0"
openai="2.0.0"
protobuf="6.32.1"
rich="14.1.0"
executing="2.2.1"
opentelemetry-api="1.37.0"
opentelemetry-exporter-otlp-proto-common="1.37.0"
opentelemetry-exporter-otlp-proto-http="1.37.0"
opentelemetry-instrumentation="0.58b0"
opentelemetry-instrumentation-asgi="0.58b0"
opentelemetry-instrumentation-fastapi="0.58b0"
opentelemetry-instrumentation-httpx="0.58b0"
opentelemetry-instrumentation-sqlalchemy="0.58b0"
opentelemetry-proto="1.37.0"
opentelemetry-sdk="1.37.0"
opentelemetry-semantic-conventions="0.58b0"
opentelemetry-util-http="0.58b0"

[alexmojaki]

The warning says "No source code available." You can confirm this by verifying that warnings and tracebacks aren't able to show lines of source code (not just line numbers), which seems to be the case in what you've shared. I'm not sure what exactly about your setup would cause this but it explains why there's a difference between inside and outside the docker container.