llama-stack-mirror/tests/integration/conftest.py
Ashwin Bharambe 2fe976ed0a
refactor(test): introduce --stack-config and simplify options (#1404)
You now run the integration tests with these options:

```bash
Custom options:
  --stack-config=STACK_CONFIG
                        a 'pointer' to the stack. this can be either be:
                        (a) a template name like `fireworks`, or
                        (b) a path to a run.yaml file, or
                        (c) an adhoc config spec, e.g.
                        `inference=fireworks,safety=llama-guard,agents=meta-
                        reference`
  --env=ENV             Set environment variables, e.g. --env KEY=value
  --text-model=TEXT_MODEL
                        comma-separated list of text models. Fixture name:
                        text_model_id
  --vision-model=VISION_MODEL
                        comma-separated list of vision models. Fixture name:
                        vision_model_id
  --embedding-model=EMBEDDING_MODEL
                        comma-separated list of embedding models. Fixture name:
                        embedding_model_id
  --safety-shield=SAFETY_SHIELD
                        comma-separated list of safety shields. Fixture name:
                        shield_id
  --judge-model=JUDGE_MODEL
                        comma-separated list of judge models. Fixture name:
                        judge_model_id
  --embedding-dimension=EMBEDDING_DIMENSION
                        Output dimensionality of the embedding model to use for
                        testing. Default: 384
  --record-responses    Record new API responses instead of using cached ones.
  --report=REPORT       Path where the test report should be written, e.g.
                        --report=/path/to/report.md

```

Importantly, if you don't specify any of the models (text-model,
vision-model, etc.) the relevant tests will get **skipped!**

This will make running tests somewhat more annoying since all options
will need to be specified. We will make this easier by adding some easy
wrapper yaml configs.

## Test Plan

Example:

```bash
ashwin@ashwin-mbp ~/local/llama-stack/tests/integration (unify_tests) $ 
LLAMA_STACK_CONFIG=fireworks pytest -s -v inference/test_text_inference.py \
   --text-model meta-llama/Llama-3.2-3B-Instruct 
```
2025-03-05 17:02:02 -08:00

160 lines
5.4 KiB
Python

# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
import inspect
import itertools
import os
import textwrap
from dotenv import load_dotenv
from .report import Report
def pytest_configure(config):
config.option.tbstyle = "short"
config.option.disable_warnings = True
load_dotenv()
env_vars = config.getoption("--env") or []
for env_var in env_vars:
key, value = env_var.split("=", 1)
os.environ[key] = value
if config.getoption("--report"):
config.pluginmanager.register(Report(config))
def pytest_addoption(parser):
parser.addoption(
"--stack-config",
help=textwrap.dedent(
"""
a 'pointer' to the stack. this can be either be:
(a) a template name like `fireworks`, or
(b) a path to a run.yaml file, or
(c) an adhoc config spec, e.g. `inference=fireworks,safety=llama-guard,agents=meta-reference`
"""
),
)
parser.addoption("--env", action="append", help="Set environment variables, e.g. --env KEY=value")
parser.addoption(
"--text-model",
help="comma-separated list of text models. Fixture name: text_model_id",
)
parser.addoption(
"--vision-model",
help="comma-separated list of vision models. Fixture name: vision_model_id",
)
parser.addoption(
"--embedding-model",
help="comma-separated list of embedding models. Fixture name: embedding_model_id",
)
parser.addoption(
"--safety-shield",
help="comma-separated list of safety shields. Fixture name: shield_id",
)
parser.addoption(
"--judge-model",
help="comma-separated list of judge models. Fixture name: judge_model_id",
)
parser.addoption(
"--embedding-dimension",
type=int,
help="Output dimensionality of the embedding model to use for testing. Default: 384",
)
parser.addoption(
"--record-responses",
action="store_true",
help="Record new API responses instead of using cached ones.",
)
parser.addoption(
"--report",
help="Path where the test report should be written, e.g. --report=/path/to/report.md",
)
MODEL_SHORT_IDS = {
"meta-llama/Llama-3.2-3B-Instruct": "3B",
"meta-llama/Llama-3.1-8B-Instruct": "8B",
"meta-llama/Llama-3.1-70B-Instruct": "70B",
"meta-llama/Llama-3.1-405B-Instruct": "405B",
"meta-llama/Llama-3.2-11B-Vision-Instruct": "11B",
"meta-llama/Llama-3.2-90B-Vision-Instruct": "90B",
"meta-llama/Llama-3.3-70B-Instruct": "70B",
"meta-llama/Llama-Guard-3-1B": "Guard1B",
"meta-llama/Llama-Guard-3-8B": "Guard8B",
"all-MiniLM-L6-v2": "MiniLM",
}
def get_short_id(value):
return MODEL_SHORT_IDS.get(value, value)
def pytest_generate_tests(metafunc):
"""
This is the main function which processes CLI arguments and generates various combinations of parameters.
It is also responsible for generating test IDs which are succinct enough.
Each option can be comma separated list of values which results in multiple parameter combinations.
"""
params = []
param_values = {}
id_parts = []
# Map of fixture name to its CLI option and ID prefix
fixture_configs = {
"text_model_id": ("--text-model", "txt"),
"vision_model_id": ("--vision-model", "vis"),
"embedding_model_id": ("--embedding-model", "emb"),
"shield_id": ("--safety-shield", "shield"),
"judge_model_id": ("--judge-model", "judge"),
"embedding_dimension": ("--embedding-dimension", "dim"),
}
# Collect all parameters and their values
for fixture_name, (option, id_prefix) in fixture_configs.items():
if fixture_name not in metafunc.fixturenames:
continue
params.append(fixture_name)
val = metafunc.config.getoption(option)
values = [v.strip() for v in str(val).split(",")] if val else [None]
param_values[fixture_name] = values
if val:
id_parts.extend(f"{id_prefix}={get_short_id(v)}" for v in values)
if not params:
return
# Generate all combinations of parameter values
value_combinations = list(itertools.product(*[param_values[p] for p in params]))
# Generate test IDs
test_ids = []
non_empty_params = [(i, values) for i, values in enumerate(param_values.values()) if values[0] is not None]
# Get actual function parameters using inspect
test_func_params = set(inspect.signature(metafunc.function).parameters.keys())
if non_empty_params:
# For each combination, build an ID from the non-None parameters
for combo in value_combinations:
parts = []
for param_name, val in zip(params, combo, strict=True):
# Only include if parameter is in test function signature and value is meaningful
if param_name in test_func_params and val:
prefix = fixture_configs[param_name][1] # Get the ID prefix
parts.append(f"{prefix}={get_short_id(val)}")
if parts:
test_ids.append(":".join(parts))
metafunc.parametrize(params, value_combinations, scope="session", ids=test_ids if test_ids else None)
pytest_plugins = ["tests.integration.fixtures.common"]