Skip to content

Latest commit

 

History

History
116 lines (80 loc) · 4.91 KB

DEVELOP.md

File metadata and controls

116 lines (80 loc) · 4.91 KB

Getting Involved

Even if you do not plan to contribute to Apache Arrow itself or Arrow integrations in other projects, we'd be happy to have you involved:

We prefer to receive contributions in the form of GitHub pull requests. Please send pull requests against the github.com/apache/arrow repository.

If you are looking for some ideas on what to contribute, check out the GitHub issues for the Apache Arrow project. Comment on the issue and/or contact [email protected] with your questions and ideas.

If you’d like to report a bug but don’t have time to fix it, you can still post it on GitHub issues, or email the mailing list [email protected]

The package.json scripts

We use yarn to install dependencies and run scrips.

  • yarn clean - cleans targets
  • yarn build - cleans and compiles all targets
  • yarn test - executes tests against built targets

These scripts accept argument lists of targets × modules:

  • Available targets are es5, es2015, esnext, ts, and all (default: all)
  • Available modules are cjs, esm, umd, and all (default: all)

Examples:

  • yarn build -- builds all ES targets in all module formats
  • yarn build -t es5 -m all -- builds the ES5 target in all module formats
  • yarn build -t all -m cjs -- builds all ES targets in the CommonJS module format
  • yarn build -t es5 -t es2015 -m all -- builds the ES5 and ES2015 targets in all module formats
  • yarn build -t es5 -m cjs -m esm -- builds the ES5 target in CommonJS and ESModules module formats

This argument configuration also applies to clean and test scripts.

To run tests on the bundles, you need to build them first. To run tests directly on the sources without bundling, use the src target (e.g. yarn test -t src).

  • yarn doc

Compiles the documentation with Typedoc. Use yarn doc --watch to automatically rebuild when the docs change.

Running the Performance Benchmarks

You can run the benchmarks with yarn perf. To print the results to stderr as JSON, add the --json flag (e.g. yarn perf --json 2> perf.json).

You can change the target you want to test by changing the imports in perf/index.ts. Note that you need to compile the bundles with yarn build before you can import them.

Testing Bundling

The bundles use apache-arrow so make sure to build it with yarn build -t apache-arrow. To bundle with a variety of bundlers, run yarn test:bundle or yarn gulp bundle.

Run yarn gulp bundle:webpack:analyze to open Webpack Bundle Analyzer.

Updating the Arrow format flatbuffers generated code

  1. Once generated, the flatbuffers format code needs to be adjusted for our build scripts (assumes gnu-sed):

    cd $ARROW_HOME

# Create a tmpdir to store modified flatbuffers schemas
tmp_format_dir=$(mktemp -d)
cp ./format/*.fbs $tmp_format_dir

# Remove namespaces from the flatbuffers schemas
sed -i '+s+namespace org.apache.arrow.flatbuf;++ig' $tmp_format_dir/*.fbs
sed -i '+s+org.apache.arrow.flatbuf.++ig' $tmp_format_dir/*.fbs

# Generate TS source from the modified Arrow flatbuffers schemas
flatc --ts -o ./js/src/fb $tmp_format_dir/{File,Schema,Message,Tensor,SparseTensor}.fbs

# Remove the tmpdir
rm -rf $tmp_format_dir

  2. Manually fix the unused imports and add // @ts-ignore for other errors

  3. Add .js to the imports. In VSCode, you can search for ^(import [^';]* from '(\./|(\.\./)+)[^';.]*)'; and replace with $1.js';.

  4. Execute yarn lint from the js directory to fix the linting errors