antlr-plsql

Development

ANTLR requires Java, so we suggest you use Docker when building grammars. The Makefile contains directives to clean, build, test and deploy the ANTLR grammar. It does not run Docker itself, so run make inside Docker.

Build the grammar

# Build the docker container
docker build -t antlr_plsql .

# Run the container to build the python grammar
# Write parser files to local file system through volume mounting
docker run -it -v ${PWD}:/usr/src/app antlr_plsql make build

Set up the Python module

Now that the Python parsing files are available, you can install them with pip:

pip install -r requirements.txt
pip install -e .

And parse SQL code in Python:

from antlr_plsql import ast
ast.parse("SELECT a from b")

Using the AST viewer

If you're actively developing on the ANLTR grammar or the tree shaping, it's a good idea to set up the AST viewer locally so you can immediately see the impact of your changes in a visual way.

• Clone the ast-viewer repo and build the Docker image according to the instructions.
• Spin up a docker container that volume mounts the Python package, symlink-installs the package and runs the server on port 3000:

docker run -it \
  -u root \
  -v ~/workspace/antlr-plsql:/app/app/antlr-plsql \
  -p 3000:3000 \
  ast-viewer \
  /bin/bash -c "echo 'Install development requirements in development:' \
    && pip install --no-deps -e app/antlr-plsql \
    && python3 run.py"

When simultaneously developing other packages, volume mount and install those too:

docker run -it \
  -u root \
  -v ~/workspace/antlr-ast:/app/app/antlr-ast \
  -v ~/workspace/antlr-plsql:/app/app/antlr-plsql \
  -v ~/workspace/antlr-tsql:/app/app/antlr-tsql \
  -p 3000:3000 \
  ast-viewer \
  /bin/bash -c "echo 'Install development requirements in development:' \
    && pip install --no-deps -e app/antlr-ast \
    && pip install --no-deps -e app/antlr-plsql \
    && pip install --no-deps -e app/antlr-tsql \
    && python3 run.py"

• If you update the tree shaping logic in this repo, the app will auto-update.
• If you change the grammar, you will have to first rebuild the grammar (with the antlr_plsql docker image) and restart the ast-viewer container.

Run tests

# Similar to building the grammar, but running tests
# and not saving the generated files
docker build -t antlr_plsql .
docker run -t antlr_plsql make build test

Or run the test locally, first build the grammar then run:

pytest

Travis deployment

• Builds the Docker image.
• Runs the Docker image to build the grammar and run the unit tests.
• Deploys the resulting python files to PyPi when a new release is made, so they can be installed easily.