Depmap Omics Long Read RNA

This repository contains WDL workflows and a Python module to continuously run workflows on new RNA long read samples and sync BAM files between Terra and Gumbo.

Installation

1. Install uv.

2. Create a new virtualenv and install the project dependencies:

   uv venv
   uv sync

3. Copy env.dist into a new file .env and fill it out:

   • WOMTOOL_JAR is the path to the .jar file downloaded from https://github.com/broadinstitute/cromwell/releases
   • FIRECLOUD_OWNERS is a list of Terra users (email addresses or group names) that should be considered owners of workflow method configs

Credentials

Your GCP DEFAULT_APPLICATION_CREDENTIALS must already be configured in order to run commands.

Development

Configure your editor or IDE to automatically format your code with Ruff and check static types with Pyright by running uv run pyright.

Whenever possible, function/method arguments and return values should be validated with Pydantic or Pandera (if a data frame).

GraphQL code generation

This repo uses ariadne-codegen to generate the gumbo_gql_client module. It uses the folder of GraphQL queries (./gql) and the current GraphQL schema for a particular Gumbo environment to generate all of the Python classes, Pydantic models, and query/mutation methods for interacting with the Gumbo GraphQL Service. To regenerate the module using the current production schema:

HASURA_ADMIN_SECRET=... uv run ariadne-codegen --config ariadne-prod.toml

Scratch files

Some Python files in ./scratch are available to seed existing data (i.e. legacy uBAM files).