/data-exports-bulk

Data export module for Atlas Prod

Primary LanguagePerlMIT LicenseMIT

Atlas Data Exports module (v0.1.0)

This module provides functionality for the Atlas Production data exports processes:

  • OpenTargets json export.
  • ChEBI export.
  • EBEYE export.
  • Large Atlas tar.gz export for bulk.

Exports are normally executed between the pre-release date and the release date, as they require release data loaded in a wwwdev or equivalent web deployment.

Our central Jenkins provision has execution jobs for all of these.

Version 0.1.0 was used for the Nov 2018 release.

Open Targets JSON schema transform

bin/schema_transform.jslt file contains logic required to migrate JSON files across different schemas. To run transformation into the new Open Targets schema, use the JSLT Java library. The simplest way is to run it via docker container:

./bin/run_schema_transform_container.sh <container>

Where is either 'docker' or 'singularity' The output file will be found in the outputs dir within current directory.

The following variables must be defined:

  • INPUT_JSON - absolute path to json file to be transformed. This can either be a single JSON in 'long' format or a multi-line file where each line corresponds to a JSON object.
  • SCHEMA_TRANSFORM - absolute path to JSLT file specifying transformation logic. See dcoumentation [here]((https://github.com/schibsted/jslt) for more detail. See file data/schema_transform.jslt as example.

Optional variables:

  • PROCESSED_JSON - specify name of output processed json file
  • OUTPUT_DIR - specify absolute path to output directory
  • IMAGE_NAME - specify Docker image name to be used

It might be needed to filter out evidence strings which do not have certain fields. To do so, run

bin/run_json_filtering.sh $json_to_process $ensembl_genes $excluded_biotypes

the arguments are:

  • json_to_process - Initial txt file with multiple json objects (one per line)
  • ensembl_genes - Table with E! gene annotation (see test_data for an example)
  • excluded_biotypes - Text file with excluded biotypes which need to be filtered out (see data directory)