Setup LLM regression framework for simple univariate models

.github/workflows/python-package-cd.yml

@@ -0,0 +1,30 @@
+name: Deploy Python package
+
+on:
+  release:
+    types: [published]
+
+jobs:
+  build-and-deploy:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.12"]
+    steps:
+    - uses: actions/checkout@v3
+    - name: Set up Python ${{ matrix.python-version }}
+      uses: actions/setup-python@v3
+      with:
+        python-version: ${{ matrix.python-version }}
+    - name: Install Nox
+      run: |
+        python -m pip install --upgrade pip
+        python -m pip install --upgrade setuptools
+        python -m pip install "nox==2023.4.22"
+    - name: Build wheel and push to PyPI
+      env:
+        PYPI_USR: ${{ secrets.PYPI_USR }}
+        PYPI_PWD: ${{ secrets.PYPI_PWD }}
+      run: |
+        nox -s build_and_deploy

.github/workflows/python-package-ci.yml

@@ -0,0 +1,33 @@
+name: Python package
+
+on:
+  pull_request:
+    branches: [ "main" ]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.12"]
+    steps:
+    - uses: actions/checkout@v3
+    - name: Set up Python ${{ matrix.python-version }}
+      uses: actions/setup-python@v3
+      with:
+        python-version: ${{ matrix.python-version }}
+    - name: Install Nox
+      run: |
+        python -m pip install --upgrade pip
+        python -m pip install --upgrade setuptools
+        python -m pip install "nox==2023.4.22"
+    - name: Run code formatting checks
+      run: |
+        nox -s check_code_formatting-${{ matrix.python-version }}
+    - name: Run static type checking
+      run: |
+        nox -s check_types-${{ matrix.python-version }}
+    - name: Run tests
+      run: |
+        nox -s run_tests-${{ matrix.python-version }}

.gitignore

@@ -1,3 +1,7 @@
+# Misc
+TODO.md
+.vscode/
+
 # Byte-compiled / optimized / DLL files
 __pycache__/
 *.py[cod]
@@ -49,6 +53,7 @@ coverage.xml
 *.py,cover
 .hypothesis/
 .pytest_cache/
+.ruff_cache
 cover/
 
 # Translations

README.md

@@ -1,2 +1,80 @@
-# llm-regression
-Exploring the classical regression capabilities of LLMs.
+# llm_regression
+
+This is the repository for the llm_regression Python package.
+
+## Developer Setup
+
+Install the package as an [editable dependency](https://setuptools.pypa.io/en/latest/userguide/development_mode.html), together with all the developer tools required to format code, check types and run tests:
+
+```text
+$ pip install -e ".[dev]"
+```
+
+### Developer Task Execution with Nox
+
+We use [Nox](https://nox.thea.codes/en/stable/) for scripting developer tasks, such as formatting code, checking types and running tests. These tasks are defined in `noxfile.py`, a list of which can be returned on the command line,
+
+```text
+$ nox --list
+
+Sessions defined in /Users/.../noxfile.py:
+
+* run_tests-3.10 -> Run unit tests.
+- format_code-3.10 -> Lint code and re-format where necessary.
+* check_code_formatting-3.10 -> Check code for formatting errors.
+* check_types-3.10 -> Run static type checking.
+- build_and_deploy-3.10 -> Build wheel and deploy to PyPI.
+
+sessions marked with * are selected, sessions marked with - are skipped.
+```
+
+Single tasks can be executed easily - e.g.,
+
+```text
+$ nox -s run_tests
+
+nox > Running session run_tests-3.10
+nox > Creating virtual environment (virtualenv) using python3.10 in .nox/run_tests-3-10
+nox > python -m pip install '.[dev]'
+nox > pytest 
+======================================== test session starts ========================================
+platform darwin -- Python 3.10.2, pytest-7.4.2, pluggy-1.3.0
+rootdir: /Users/.../llm_regression
+configfile: pyproject.toml
+testpaths: tests
+collected 1 item                                                                                                                                                         
+
+tests/test_hello_world.py .                                                                                                                                        [100%]
+
+========================================== 1 passed in 0.00s =========================================
+nox > Session run_tests-3.10 was successful.
+```
+
+### Building Packages and Deploying to PyPI
+
+This is automated via the `nox -s build_and_deploy` command. In order to use this, the following environment variables will need to be made available to Python:
+
+```text
+PYPI_USR  # PyPI username
+PYPI_PWD  # PyPI password
+```
+
+These may be specified in a `.env` file from which they will be loaded automatically - e.g.,
+
+```text
+PYPI_USR=XXXX
+PYPI_PWD=XXXX
+```
+
+Note: `.gitignore` will ensure that `.env`is not tracked by Git.
+
+## CI/CD
+
+This repo comes configured to run two [GitHub Actions](https://docs.github.com/en/actions) workflows:
+
+- **Test Python Package (CI)**, defined in `.github/workflows/python-package-ci.yml`
+- **Deploy Python Package (CD)**, defined in `.github/workflows/python-package-cd.yml`
+
+The CI workflow has been configured to run whenever a pull request to the `main` branch is created. The CD workflow has been configured to run whenever a [release](https://docs.github.com/en/repositories/releasing-projects-on-github/managing-releases-in-a-repository) is created on GitHub.
+
+Note, the CD workflow will require `PYPI_USR` and `PYPI_PWD` to be added as [repository secrets](https://docs.github.com/en/actions/security-guides/using-secrets-in-github-actions).

noxfile.py

@@ -0,0 +1,64 @@
+"""Developer task automation."""
+import os
+
+import nox
+
+nox.options.sessions = [
+    "check_code_formatting",
+    "check_types",
+    "run_tests",
+]
+
+PYTHON = ["3.12"]
+
+
+@nox.session(python=PYTHON, reuse_venv=True)
+def run_tests(session: nox.Session):
+    """Run unit tests."""
+    session.install(".[dev]")
+    pytest_args = session.posargs if session.posargs else []
+    session.run("pytest", *pytest_args)
+
+
+@nox.session(python=PYTHON, reuse_venv=True)
+def format_code(session: nox.Session):
+    """Lint code and re-format where necessary."""
+    session.install(".[dev]")
+    session.run("black", "--config=pyproject.toml", ".")
+    session.run("ruff", "check", ".", "--config=pyproject.toml", "--fix")
+
+
+@nox.session(python=PYTHON, reuse_venv=True)
+def check_code_formatting(session: nox.Session):
+    """Check code for formatting errors."""
+    session.install(".[dev]")
+    session.run("black", "--config=pyproject.toml", "--check", ".")
+    session.run("ruff", "check", ".", "--config=pyproject.toml")
+
+
+@nox.session(python=PYTHON, reuse_venv=True)
+def check_types(session: nox.Session):
+    """Run static type checking."""
+    session.install(".[dev]")
+    session.run("mypy")
+
+
+@nox.session(python=PYTHON, reuse_venv=True)
+def build_and_deploy(session: nox.Session):
+    """Build wheel and deploy to PyPI."""
+    try:
+        from dotenv import load_dotenv
+
+        load_dotenv()
+    except ModuleNotFoundError:
+        session.warn("Expecting PYPI_USR and PYPI_PWD in local environment variables.")
+
+    try:
+        PYPI_USR = os.environ["PYPI_USR"]
+        PYPI_PWD = os.environ["PYPI_PWD"]
+    except KeyError as e:
+        session.error(f"{str(e)} not found in local environment variables.")
+    session.install(".[deploy]")
+    session.run("rm", "-rf", "dist")
+    session.run("python", "-m", "build")
+    session.run("twine", "upload", "dist/*", "-u", PYPI_USR, "-p", PYPI_PWD)

pyproject.toml

@@ -0,0 +1,94 @@
+[project]
+name = "llm_regression"
+version = "0.1.0.dev0"
+description = "The llm_regression Python package."
+readme = "README.md"
+authors = [
+  { name="Alex Ioannides", email="[email protected]" },
+]
+dependencies = [
+  "openai==1.30.*",
+  "numpy==1.26.*",
+  "pandas==2.2.*",
+  "python-dotenv==1.0.*",
+  "scikit-learn==1.4.*",
+  "tqdm==4.66.*",
+]
+
+[project.optional-dependencies]
+dev = [
+  "black==23.9.1",
+  "python-dotenv>=1.0.0",
+  "icecream",
+  "ipython",
+  "mypy==1.5.1",
+  "nox==2023.4.22",
+  "pandas-stubs==2.2.2.240514",
+  "pytest==7.4.2",
+  "ruff==0.0.290",
+  "types-pytz==2024.1.0.20240417",
+  "types-tqdm==4.66.0.20240417",
+]
+deploy = [
+  "build>=1.0.0",
+  "pip>=23.2.0",
+  "setuptools>=68.0.0",
+  "twine>=4.0.0",
+  "wheel>=0.41.0",
+]
+
+[project.urls]
+"Homepage" = "https://github.com/AlexIoannides/llm-regression"
+"Bug Tracker" = "https://github.com/AlexIoannides/llm-regression/issues"
+
+[build-system]
+requires = ["setuptools>=68.0"]
+build-backend = "setuptools.build_meta"
+
+[tool.setuptools]
+include-package-data = true
+
+[tool.setuptools.packages.find]
+where = ["src"]
+
+[tool.black]
+line-length = 88
+
+[tool.ruff]
+src = ["src"]
+target-version = "py310"
+line-length = 88
+select = [
+  "D",   # pydocstyle
+  "E",   # pycodestyle errors
+  "F",   # pyflakes
+  "I",   # isort
+  "UP",  # pyupgrade
+  "W",   # pycodestyle warnings
+]
+ignore = [
+  "D203",    # fix pydocstyle warning
+  "D213",    # fix pydocstyle warning
+]
+
+[tool.ruff.per-file-ignores]
+"tests/*" = [
+  "D103",
+]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+
+[tool.mypy]
+python_version = "3.12"
+files = [
+  "src",
+  "tests",
+  "noxfile.py",
+]
+
+[[tool.mypy.overrides]]
+module = [
+  "sklearn.*",
+]
+ignore_missing_imports = true

src/llm_regression/__init__.py

@@ -0,0 +1 @@
+"""The llm_regression package."""