Merge pull request #57 from allenai/rust

Rust + CI etc

Author: sergeyf

.github/workflows/main.yaml

@@ -47,6 +47,20 @@ jobs:
       - name: Setup uv
         uses: astral-sh/setup-uv@v3
 
+      - name: Setup Rust
+        uses: dtolnay/rust-toolchain@stable
+
+      - name: Cache cargo build
+        uses: actions/cache@v4
+        with:
+          path: |
+            ~/.cargo/registry
+            ~/.cargo/git
+            s2and_rust/target
+          key: cargo-${{ runner.os }}-${{ hashFiles('s2and_rust/Cargo.lock') }}
+          restore-keys: |
+            cargo-${{ runner.os }}-
+
       # Optional: ensure a specific Python (uv can also manage this on its own)
       - name: Setup Python
         uses: actions/setup-python@v5
@@ -70,24 +84,28 @@ jobs:
         shell: bash
         run: |
           if [[ -f uv.lock ]]; then
-            uv sync --all-extras --dev --frozen
+            uv sync --extra dev --frozen
           else
             # No lock present; resolve once, then install
-            uv sync --all-extras --dev
+            uv sync --extra dev
           fi
 
+      # Build/install Rust extension for parity tests
+      - name: Build Rust extension
+        run: uvx --from maturin maturin develop -m s2and_rust/Cargo.toml
+
       # Type checking (run mypy commands directly)
       - name: mypy (s2and)
-        run: uv run mypy s2and --ignore-missing-imports
+        run: uv run --no-project mypy s2and --ignore-missing-imports
       - name: mypy (scripts)
-        run: uv run mypy scripts/*.py --ignore-missing-imports
+        run: uv run --no-project mypy scripts/*.py --ignore-missing-imports
 
       # Single pytest run with coverage (replaces the two docker pytest calls)
       - name: pytest (coverage)
         env:
           # keep startup lean; avoid user-level plugins on hosted runners
           PYTHONPATH: .
         run: |
-          uv run pytest tests/ \
+          uv run --no-project pytest tests/ \
             --cov=s2and --cov-report=term-missing --cov-fail-under=40
 

.github/workflows/release-rust.yml

@@ -0,0 +1,336 @@
+name: build-and-publish
+
+on:
+  pull_request:
+    types: [opened, synchronize, reopened, labeled, unlabeled]
+  push:
+    branches: [main]
+  workflow_dispatch:
+    inputs:
+      force_build:
+        description: "Force build jobs (manual/PR testing)"
+        required: false
+        type: boolean
+      publish_s2and:
+        description: "Publish s2and to PyPI (manual run)"
+        required: false
+        type: boolean
+      publish_rust:
+        description: "Publish s2and-rust to PyPI (manual run)"
+        required: false
+        type: boolean
+
+jobs:
+  detect-versions:
+    name: detect version changes
+    runs-on: ubuntu-latest
+    outputs:
+      s2and_changed: ${{ steps.detect.outputs.s2and_changed }}
+      rust_changed: ${{ steps.detect.outputs.rust_changed }}
+      publish_any: ${{ steps.detect.outputs.publish_any }}
+      publish_s2and: ${{ steps.detect.outputs.publish_s2and }}
+      publish_rust: ${{ steps.detect.outputs.publish_rust }}
+      force_build: ${{ steps.detect.outputs.force_build }}
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Detect version changes
+        id: detect
+        env:
+          BEFORE_SHA: ${{ github.event.before }}
+        run: |
+          python - <<'PY'
+          import json
+          import os
+          import subprocess
+          import sys
+          import tomllib
+
+          event_name = os.environ.get("GITHUB_EVENT_NAME") or ""
+          event_path = os.environ.get("GITHUB_EVENT_PATH") or ""
+          event = None
+          if event_path and os.path.exists(event_path):
+            with open(event_path, "r", encoding="utf-8") as f:
+              event = json.load(f)
+
+          force_build = False
+          publish_s2and = False
+          publish_rust = False
+          if event_name == "workflow_dispatch":
+            inputs = (event or {}).get("inputs", {})
+            force_build = str(inputs.get("force_build", "")).lower() in {"1", "true", "yes", "on"}
+            publish_s2and = str(inputs.get("publish_s2and", "")).lower() in {"1", "true", "yes", "on"}
+            publish_rust = str(inputs.get("publish_rust", "")).lower() in {"1", "true", "yes", "on"}
+          elif event_name == "pull_request":
+            labels = (event or {}).get("pull_request", {}).get("labels", [])
+            force_build = any((label or {}).get("name", "").lower() == "force-build" for label in labels)
+
+          before = os.environ.get("BEFORE_SHA") or ""
+          if event_name == "pull_request" and event:
+            before = (event.get("pull_request", {}).get("base", {}).get("sha", "")) or ""
+          if before.startswith("0000000"):
+            before = ""
+
+          def read_toml_at(path, rev=None):
+            if rev:
+              try:
+                data = subprocess.check_output(["git", "show", f"{rev}:{path}"], text=True)
+              except subprocess.CalledProcessError:
+                return None
+              return tomllib.loads(data)
+            with open(path, "rb") as f:
+              return tomllib.load(f)
+
+          def get_project_version(toml_obj, path):
+            if toml_obj is None:
+              return None
+            try:
+              return toml_obj["project"]["version"]
+            except KeyError:
+              raise SystemExit(f"Missing [project].version in {path}")
+
+          def get_cargo_version(toml_obj, path):
+            if toml_obj is None:
+              return None
+            try:
+              return toml_obj["package"]["version"]
+            except KeyError:
+              raise SystemExit(f"Missing [package].version in {path}")
+
+          cur_py = read_toml_at("pyproject.toml")
+          cur_rust_py = read_toml_at("s2and_rust/pyproject.toml")
+          cur_rust_cargo = read_toml_at("s2and_rust/Cargo.toml")
+
+          cur_py_ver = get_project_version(cur_py, "pyproject.toml")
+          cur_rust_py_ver = get_project_version(cur_rust_py, "s2and_rust/pyproject.toml")
+          cur_rust_cargo_ver = get_cargo_version(cur_rust_cargo, "s2and_rust/Cargo.toml")
+
+          if cur_rust_py_ver != cur_rust_cargo_ver:
+            raise SystemExit(
+              f"Rust version mismatch: pyproject={cur_rust_py_ver} cargo={cur_rust_cargo_ver}"
+            )
+
+          before_py_ver = get_project_version(read_toml_at("pyproject.toml", before), "pyproject.toml") if before else None
+          before_rust_py_ver = (
+            get_project_version(read_toml_at("s2and_rust/pyproject.toml", before), "s2and_rust/pyproject.toml")
+            if before
+            else None
+          )
+
+          s2and_changed = before_py_ver != cur_py_ver
+          rust_changed = before_rust_py_ver != cur_rust_py_ver
+          publish_any = s2and_changed or rust_changed or publish_s2and or publish_rust
+
+          out_path = os.environ["GITHUB_OUTPUT"]
+          with open(out_path, "a", encoding="utf-8") as f:
+            f.write(f"s2and_changed={'true' if s2and_changed else 'false'}\n")
+            f.write(f"rust_changed={'true' if rust_changed else 'false'}\n")
+            f.write(f"publish_any={'true' if publish_any else 'false'}\n")
+            f.write(f"publish_s2and={'true' if publish_s2and else 'false'}\n")
+            f.write(f"publish_rust={'true' if publish_rust else 'false'}\n")
+            f.write(f"force_build={'true' if force_build else 'false'}\n")
+
+          print(f"s2and: {before_py_ver} -> {cur_py_ver} changed={s2and_changed}")
+          print(f"rust:  {before_rust_py_ver} -> {cur_rust_py_ver} changed={rust_changed}")
+          print(f"force_build: {force_build}")
+          print(f"publish_s2and: {publish_s2and}")
+          print(f"publish_rust: {publish_rust}")
+          PY
+
+  s2and-dist:
+    name: s2and sdist+wheel
+    runs-on: ubuntu-latest
+    needs: [detect-versions]
+    if: needs.detect-versions.outputs.s2and_changed == 'true' || needs.detect-versions.outputs.force_build == 'true' || needs.detect-versions.outputs.publish_s2and == 'true'
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+
+      - name: Build s2and distributions
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install build
+          python -m build --sdist --wheel --outdir dist
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: dist-s2and
+          path: dist/*
+
+  wheels-windows:
+    name: wheels (windows, py${{ matrix.py }})
+    runs-on: windows-latest
+    needs: [detect-versions]
+    if: needs.detect-versions.outputs.rust_changed == 'true' || needs.detect-versions.outputs.force_build == 'true' || needs.detect-versions.outputs.publish_rust == 'true'
+    strategy:
+      fail-fast: false
+      matrix:
+        py: ["3.10", "3.11", "3.12"]
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.py }}
+
+      - name: Build wheels
+        uses: PyO3/maturin-action@86b9d133d34bc1b40018696f782949dac11bd380
+        with:
+          maturin-version: v1.11.5
+          command: build
+          args: --release --locked --compatibility pypi --out dist
+          working-directory: s2and_rust
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: dist-s2and-rust-windows-py${{ matrix.py }}
+          path: s2and_rust/dist/*
+
+  wheels-macos:
+    name: wheels (macos universal2, py${{ matrix.py }})
+    runs-on: macos-latest
+    needs: [detect-versions]
+    if: needs.detect-versions.outputs.rust_changed == 'true' || needs.detect-versions.outputs.force_build == 'true' || needs.detect-versions.outputs.publish_rust == 'true'
+    strategy:
+      fail-fast: false
+      matrix:
+        py: ["3.10", "3.11", "3.12"]
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.py }}
+
+      - name: Install Rust targets (universal2)
+        run: rustup target add x86_64-apple-darwin aarch64-apple-darwin
+
+      - name: Build wheels (universal2)
+        uses: PyO3/maturin-action@86b9d133d34bc1b40018696f782949dac11bd380
+        env:
+          ARCHFLAGS: "-arch x86_64 -arch arm64"
+        with:
+          maturin-version: v1.11.5
+          command: build
+          args: --release --locked --compatibility pypi --out dist
+          working-directory: s2and_rust
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: dist-s2and-rust-macos-universal2-py${{ matrix.py }}
+          path: s2and_rust/dist/*
+
+  wheels-linux:
+    name: wheels (linux ${{ matrix.platform.name }})
+    runs-on: ubuntu-latest
+    needs: [detect-versions]
+    if: needs.detect-versions.outputs.rust_changed == 'true' || needs.detect-versions.outputs.force_build == 'true' || needs.detect-versions.outputs.publish_rust == 'true'
+    strategy:
+      fail-fast: false
+      matrix:
+        platform:
+          - name: manylinux2014-x86_64
+            target: x86_64-unknown-linux-gnu
+            manylinux: "2_17"
+          # Optional: enable when you want aarch64 wheels.
+          - name: manylinux2014-aarch64
+            target: aarch64-unknown-linux-gnu
+            manylinux: "2_17"
+          - name: musllinux-x86_64
+            target: x86_64-unknown-linux-musl
+            manylinux: "musllinux_1_2"
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Build wheels (manylinux/musllinux)
+        uses: PyO3/maturin-action@86b9d133d34bc1b40018696f782949dac11bd380
+        with:
+          maturin-version: v1.11.5
+          command: build
+          target: ${{ matrix.platform.target }}
+          manylinux: ${{ matrix.platform.manylinux }}
+          args: >
+            --release --locked --compatibility pypi --out dist
+            -i python3.10 -i python3.11 -i python3.12
+          working-directory: s2and_rust
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: dist-s2and-rust-linux-${{ matrix.platform.name }}
+          path: s2and_rust/dist/*
+
+  sdist:
+    name: sdist
+    runs-on: ubuntu-latest
+    needs: [detect-versions]
+    if: needs.detect-versions.outputs.rust_changed == 'true' || needs.detect-versions.outputs.force_build == 'true' || needs.detect-versions.outputs.publish_rust == 'true'
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Build sdist
+        uses: PyO3/maturin-action@86b9d133d34bc1b40018696f782949dac11bd380
+        with:
+          maturin-version: v1.11.5
+          command: sdist
+          args: --out dist
+          working-directory: s2and_rust
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: dist-s2and-rust-sdist
+          path: s2and_rust/dist/*
+
+  # Publish is split so skipped build jobs don't block unrelated publishes.
+  # Manual runs can publish by setting workflow_dispatch inputs.
+  publish-s2and:
+    name: publish s2and to PyPI
+    if: (github.event_name == 'push' && github.ref == 'refs/heads/main' && needs.detect-versions.outputs.s2and_changed == 'true') || (github.event_name == 'workflow_dispatch' && github.ref == 'refs/heads/main' && needs.detect-versions.outputs.publish_s2and == 'true')
+    needs: [detect-versions, s2and-dist]
+    runs-on: ubuntu-latest
+    environment:
+      name: pypi
+    permissions:
+      id-token: write
+
+    steps:
+      - name: Download s2and dists
+        uses: actions/download-artifact@v4
+        with:
+          name: dist-s2and
+          path: dist/s2and
+
+      - name: Publish s2and
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          packages-dir: dist/s2and
+
+  publish-rust:
+    name: publish s2and-rust to PyPI
+    if: (github.event_name == 'push' && github.ref == 'refs/heads/main' && needs.detect-versions.outputs.rust_changed == 'true') || (github.event_name == 'workflow_dispatch' && github.ref == 'refs/heads/main' && needs.detect-versions.outputs.publish_rust == 'true')
+    needs: [detect-versions, wheels-windows, wheels-macos, wheels-linux, sdist]
+    runs-on: ubuntu-latest
+    environment:
+      name: pypi
+    permissions:
+      id-token: write
+
+    steps:
+      - name: Download s2and-rust dists
+        uses: actions/download-artifact@v4
+        with:
+          pattern: dist-s2and-rust-*
+          merge-multiple: true
+          path: dist/s2and-rust
+
+      - name: Publish s2and-rust
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          packages-dir: dist/s2and-rust