chore: run tests with uv (#106)

switches to using uv lockfile to run tests to make ci environment more
reproducible.

Author: tokoko

.github/workflows/test.yml

@@ -22,14 +22,13 @@ jobs:
         uses: actions/checkout@v5
         with:
           submodules: recursive
-      - name: Set up Python
-        uses: actions/setup-python@v6
+      - name: Install uv with python
+        uses: astral-sh/setup-uv@v7
         with:
           python-version: ${{ matrix.python }}
       - name: Install package and test dependencies
         run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[test]"
+          uv sync --frozen --extra test
       - name: Run tests
         run: |
-          python -m pytest tests
+          uv run pytest

CONTRIBUTING.md

@@ -6,11 +6,10 @@ git clone --recursive https://github.com/<your-fork>/substrait-python.git
 cd substrait-python
 ```
 
-## Conda env
-Create a conda environment with developer dependencies.
+## Development environment
+Activate environment with uv.
 ```
-conda env create -f environment.yml
-conda activate substrait-python-env
+uv sync --extra test
 ```
 
 ## Update the substrait submodule locally
@@ -21,43 +20,45 @@ git submodule update --init --recursive
 ```
 
 
-# Upgrade the substrait protocol definition
+# Code generation
 
-## a) Use the upgrade script
+## Protobuf stubs
 
 Run the upgrade script to upgrade the submodule and regenerate the protobuf stubs.
 
 ```
-./update_proto.sh <version>
+uv sync --extra gen_proto
+uv run ./update_proto.sh <version>
 ```
 
-## b) Manual upgrade
+## Antlr grammar
 
-### Upgrade the Substrait submodule
+Substrait uses antlr grammar to derive output types of extension functions. Make sure java is installed and ANTLR_JAR environment variable is set. Take a look at .devcontainer/Dockerfile for example setup.
 
 ```
-cd third_party/substrait
-git checkout <version>
-cd -
-git commit . -m "Use submodule <version>"
+make antlr
 ```
 
-### Generate protocol buffers
-Generate the protobuf files manually. Requires protobuf `v3.20.1`.
+## Extensions stubs
+
+Substrait uses jsonschema to describe the data model for extension files.
+
 ```
-./gen_proto.sh
+make codegen-extensions
 ```
 
+# Lint & Format
+
+Run the following make commands to lint and format with ruff.
 
-# Build
-## Python package
-Editable installation.
 ```
-pip install -e .
+make lint
+make format
 ```
 
 # Test
 Run tests in the project's root dir.
 ```
-pytest
+uv sync --extra test
+uv run pytest
 ```

environment.yml

@@ -5,7 +5,7 @@
 from substrait.builders.plan import read_named_table
 import pytest
 from substrait.gen.proto.extensions.extensions_pb2 import AdvancedExtension
-from google.protobuf import any
+from google.protobuf import any_pb2
 from google.protobuf.wrappers_pb2 import StringValue
 
 struct = stt.Type.Struct(
@@ -80,7 +80,10 @@ def test_read_rel_schema_nullable():
 
 
 def test_read_rel_ae():
-    extension = AdvancedExtension(optimization=[any.pack(StringValue(value="Opt1"))])
+    any = any_pb2.Any()
+    any.Pack(StringValue(value="Opt1"))
+
+    extension = AdvancedExtension(optimization=[any])
 
     actual = read_named_table(["example_db", "example_table"], named_struct, extension)(
         None