inspect-test-utils 0.2.0__tar.gz

inspect_test_utils-0.2.0/.gitignore

@@ -0,0 +1,210 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+#poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+#pdm.lock
+#pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+#pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.envrc
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+# Abstra
+# Abstra is an AI-powered process automation framework.
+# Ignore directories containing user credentials, local state, and settings.
+# Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#  Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#  that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#  and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#  you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Cursor
+#  Cursor is an AI-powered code editor. `.cursorignore` specifies files/directories to
+#  exclude from AI features like autocomplete and code analysis. Recommended for sensitive data
+#  refer to https://docs.cursor.com/context/ignore-files
+.cursorignore
+.cursorindexingignore
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/
+
+# Inspect AI
+logs

inspect_test_utils-0.2.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 METR
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

inspect_test_utils-0.2.0/PKG-INFO

@@ -0,0 +1,78 @@
+Metadata-Version: 2.4
+Name: inspect-test-utils
+Version: 0.2.0
+Summary: Simple models and tasks for integration testing
+Project-URL: Repository, https://github.com/METR/inspect-test-utils
+Project-URL: Issues, https://github.com/METR/inspect-test-utils/issues
+Author-email: METR <rasmus.faber-espensen@metr.org>
+License-Expression: MIT
+License-File: LICENSE
+Classifier: Development Status :: 4 - Beta
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Requires-Python: >=3.10
+Requires-Dist: inspect-ai
+Requires-Dist: inspect-scout
+Requires-Dist: pytest>=8.0
+Provides-Extra: dev
+Requires-Dist: basedpyright; extra == 'dev'
+Requires-Dist: pytest; extra == 'dev'
+Requires-Dist: pytest-asyncio; extra == 'dev'
+Requires-Dist: pytest-mock; extra == 'dev'
+Requires-Dist: ruff; extra == 'dev'
+Requires-Dist: types-aiofiles; extra == 'dev'
+Requires-Dist: types-pyyaml; extra == 'dev'
+Requires-Dist: types-requests; extra == 'dev'
+Description-Content-Type: text/markdown
+
+inspect-test-utils
+
+A small collection of tasks, scorers, and a simple model for use with the Inspect AI framework. It is designed to support integration/acceptance tests, demos, and reproductions by providing:
+- Ready-made Tasks that exercise common evaluation patterns (simple generation, numeric closeness, failure injection, and sandbox configuration).
+- Scorers for deterministic or parameterized scoring (including hardcoded outputs and a logarithmic closeness score).
+- A hardcoded ModelAPI implementation that can deterministically emit tool calls and/or final answers, useful for testing tool-calling flows without hitting external APIs.
+
+Passing arguments
+Most tasks and the hardcoded model accept parameters. With the Inspect CLI you can pass them via --task-arg and --model-arg repeatedly:
+- Example: make the task generate 3 samples and set a numeric target for guessing:
+  inspect eval inspect_test_utils/guess_number \
+    --task-arg sample_count=3 \
+    --task-arg target=42.7 \
+    --model hardcoded --model-arg answer=42.6
+
+What’s included
+- Tasks (inspect_test_utils.tasks)
+  - say_hello(sample_count=1): Simple task; expects a response that includes "hello".
+  - guess_number(sample_count=1, target="42.7"): Uses a logarithmic closeness scorer for numeric answers.
+  - hardcoded_score(sample_count=10, hardcoded_score=None, hardcoded_score_by_sample_id_and_epoch=None): Scores are injected from parameters; useful for testing aggregations and edge cases (including NaN).
+  - sometimes_fails_setup(sample_count=10, fail_setup_on_epochs=None, failure_rate=0.2): Randomly raises during setup via a failing solver; useful to test retry/resume behavior.
+  - sometimes_fails_scoring(sample_count=10, fail_score_on_epochs=None, failure_rate=0.2): Randomly raises during scoring; useful to test scorer error handling.
+  - configurable_sandbox(sample_count=1, cpu=0.5, memory="2G", storage="2G", gpu=None, gpu_model=None, allow_internet=False): A task with runtime configurable sandbox. 
+
+- Scorers (inspect_test_utils.scorers)
+  - failing_scorer(fail_on_epochs=None, failure_rate=0.2): Raises errors at a controlled rate for selected epochs.
+  - closeness_log(): Scores 1.0 for exact equality, otherwise 1/(1+log1p(relative_error)) for numeric strings.
+  - hardcoded_scorer(hardcoded_score=None, hardcoded_score_by_sample_id_and_epoch=None): Returns pre-specified Score objects or looks them up by sample id and epoch.
+
+- Model (inspect_test_utils.hardcoded)
+  - hardcoded: A ModelAPI that can emit a sequence of tool calls (e.g., bash) for a number of repetitions and then submit a final answer.
+    Parameters include:
+    - answer: final answer string (default: "done").
+    - repetitions: how many tool-call "turns" before submitting.
+    - tool_calls: list of tool calls or shell strings (e.g., ["echo hi", "ls -la"]) to simulate; defaults to none.
+    - delay: optional delay (seconds) before returning each model output.
+
+## Installation
+
+```bash
+pip install inspect-test-utils
+```
+
+## License
+
+MIT. See [LICENSE](LICENSE).

inspect_test_utils-0.2.0/README.md

@@ -0,0 +1,46 @@
+inspect-test-utils
+
+A small collection of tasks, scorers, and a simple model for use with the Inspect AI framework. It is designed to support integration/acceptance tests, demos, and reproductions by providing:
+- Ready-made Tasks that exercise common evaluation patterns (simple generation, numeric closeness, failure injection, and sandbox configuration).
+- Scorers for deterministic or parameterized scoring (including hardcoded outputs and a logarithmic closeness score).
+- A hardcoded ModelAPI implementation that can deterministically emit tool calls and/or final answers, useful for testing tool-calling flows without hitting external APIs.
+
+Passing arguments
+Most tasks and the hardcoded model accept parameters. With the Inspect CLI you can pass them via --task-arg and --model-arg repeatedly:
+- Example: make the task generate 3 samples and set a numeric target for guessing:
+  inspect eval inspect_test_utils/guess_number \
+    --task-arg sample_count=3 \
+    --task-arg target=42.7 \
+    --model hardcoded --model-arg answer=42.6
+
+What’s included
+- Tasks (inspect_test_utils.tasks)
+  - say_hello(sample_count=1): Simple task; expects a response that includes "hello".
+  - guess_number(sample_count=1, target="42.7"): Uses a logarithmic closeness scorer for numeric answers.
+  - hardcoded_score(sample_count=10, hardcoded_score=None, hardcoded_score_by_sample_id_and_epoch=None): Scores are injected from parameters; useful for testing aggregations and edge cases (including NaN).
+  - sometimes_fails_setup(sample_count=10, fail_setup_on_epochs=None, failure_rate=0.2): Randomly raises during setup via a failing solver; useful to test retry/resume behavior.
+  - sometimes_fails_scoring(sample_count=10, fail_score_on_epochs=None, failure_rate=0.2): Randomly raises during scoring; useful to test scorer error handling.
+  - configurable_sandbox(sample_count=1, cpu=0.5, memory="2G", storage="2G", gpu=None, gpu_model=None, allow_internet=False): A task with runtime configurable sandbox. 
+
+- Scorers (inspect_test_utils.scorers)
+  - failing_scorer(fail_on_epochs=None, failure_rate=0.2): Raises errors at a controlled rate for selected epochs.
+  - closeness_log(): Scores 1.0 for exact equality, otherwise 1/(1+log1p(relative_error)) for numeric strings.
+  - hardcoded_scorer(hardcoded_score=None, hardcoded_score_by_sample_id_and_epoch=None): Returns pre-specified Score objects or looks them up by sample id and epoch.
+
+- Model (inspect_test_utils.hardcoded)
+  - hardcoded: A ModelAPI that can emit a sequence of tool calls (e.g., bash) for a number of repetitions and then submit a final answer.
+    Parameters include:
+    - answer: final answer string (default: "done").
+    - repetitions: how many tool-call "turns" before submitting.
+    - tool_calls: list of tool calls or shell strings (e.g., ["echo hi", "ls -la"]) to simulate; defaults to none.
+    - delay: optional delay (seconds) before returning each model output.
+
+## Installation
+
+```bash
+pip install inspect-test-utils
+```
+
+## License
+
+MIT. See [LICENSE](LICENSE).

inspect_test_utils-0.2.0/inspect_test_utils/__init__.py

@@ -0,0 +1,37 @@
+"""Reusable testing framework for Inspect AI evaluation tasks.
+
+This framework provides:
+- Model-level mocking (HardcodedModelAPI)
+- Solver-level hardcoded execution
+- Test scorers and tasks
+- Pytest fixtures and assertion helpers
+- Eval runner for integration tests
+"""
+
+from inspect_test_utils.assertions import (
+    assert_contains,
+    assert_eval_score,
+    assert_files_exist,
+    assert_score_in_range,
+)
+from inspect_test_utils.eval_runner import (
+    EvalTestResult,
+    run_eval_test,
+)
+from inspect_test_utils.solvers import (
+    hardcoded_bash_solver,
+    hardcoded_python_solver,
+    inspection_solver,
+)
+
+__all__ = [
+    "assert_contains",
+    "assert_eval_score",
+    "assert_files_exist",
+    "assert_score_in_range",
+    "EvalTestResult",
+    "hardcoded_bash_solver",
+    "hardcoded_python_solver",
+    "inspection_solver",
+    "run_eval_test",
+]

inspect_test_utils-0.2.0/inspect_test_utils/_registry.py

@@ -0,0 +1,39 @@
+from inspect_test_utils.hardcoded import hardcoded
+from inspect_test_utils.mockllm import mockllm_wrapper
+from inspect_test_utils.scanners import (
+    model_roles_scanner,
+    suspicious_behaviour,
+    word_counter,
+)
+from inspect_test_utils.tasks import (
+    configurable_sandbox,
+    guess_number,
+    guess_number_keep_guessing,
+    hardcoded_score,
+    network_sandbox,
+    say_hello,
+    say_hello_with_tools,
+    sometimes_fails_scoring,
+    sometimes_fails_setup,
+    timeout,
+    uses_model_roles,
+)
+
+__all__ = [
+    "configurable_sandbox",
+    "guess_number",
+    "guess_number_keep_guessing",
+    "hardcoded",
+    "hardcoded_score",
+    "mockllm_wrapper",
+    "model_roles_scanner",
+    "network_sandbox",
+    "say_hello",
+    "say_hello_with_tools",
+    "sometimes_fails_scoring",
+    "sometimes_fails_setup",
+    "suspicious_behaviour",
+    "timeout",
+    "uses_model_roles",
+    "word_counter",
+]

inspect_test_utils-0.2.0/inspect_test_utils/assertions.py

@@ -0,0 +1,114 @@
+"""Assertion helpers for Inspect AI evaluation tests.
+
+These functions provide convenient assertions for common test patterns.
+"""
+
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from inspect_test_utils.eval_runner import EvalTestResult
+
+
+def assert_eval_score(
+    result: "EvalTestResult",
+    expected: float,
+    tolerance: float = 0.01,
+    *,
+    message: str | None = None,
+) -> None:
+    """Assert that the eval score matches the expected value.
+
+    Args:
+        result: The EvalTestResult from run_eval_test.
+        expected: Expected score value.
+        tolerance: Acceptable deviation from expected score.
+        message: Optional custom failure message.
+
+    Raises:
+        AssertionError: If score doesn't match within tolerance.
+    """
+    actual = result.score
+    if actual is None:
+        raise AssertionError(f"No score returned. Error: {result.error}")
+
+    if abs(actual - expected) > tolerance:
+        msg = message or f"Expected score {expected}, got {actual}"
+        if result.explanation:
+            msg += f"\nExplanation: {result.explanation}"
+        raise AssertionError(msg)
+
+
+def assert_score_in_range(
+    result: "EvalTestResult",
+    min_score: float,
+    max_score: float,
+    *,
+    message: str | None = None,
+) -> None:
+    """Assert that the eval score is within a range.
+
+    Args:
+        result: The EvalTestResult from run_eval_test.
+        min_score: Minimum acceptable score (inclusive).
+        max_score: Maximum acceptable score (inclusive).
+        message: Optional custom failure message.
+
+    Raises:
+        AssertionError: If score is outside the range.
+    """
+    actual = result.score
+    if actual is None:
+        raise AssertionError(f"No score returned. Error: {result.error}")
+
+    if not (min_score <= actual <= max_score):
+        msg = message or f"Expected score in [{min_score}, {max_score}], got {actual}"
+        if result.explanation:
+            msg += f"\nExplanation: {result.explanation}"
+        raise AssertionError(msg)
+
+
+def assert_files_exist(
+    files: list[str],
+    actual_files: list[str],
+    *,
+    message: str | None = None,
+) -> None:
+    """Assert that expected files exist in the actual file list.
+
+    Args:
+        files: List of expected file paths/names.
+        actual_files: List of actual files found.
+        message: Optional custom failure message.
+
+    Raises:
+        AssertionError: If any expected file is missing.
+    """
+    missing = [f for f in files if f not in actual_files]
+    if missing:
+        msg = message or f"Missing files: {missing}"
+        msg += f"\nFound: {actual_files}"
+        raise AssertionError(msg)
+
+
+def assert_contains(
+    needle: str,
+    haystack: str,
+    *,
+    message: str | None = None,
+) -> None:
+    """Assert that a string contains a substring.
+
+    Args:
+        needle: String to search for.
+        haystack: String to search in.
+        message: Optional custom failure message.
+
+    Raises:
+        AssertionError: If needle not found in haystack.
+    """
+    if needle not in haystack:
+        msg = message or f"Expected to find '{needle}' in output"
+        # Show a preview of the haystack
+        preview = haystack[:500] + "..." if len(haystack) > 500 else haystack
+        msg += f"\nActual output:\n{preview}"
+        raise AssertionError(msg)