reactome / release-orthopairs

Data-release-pipeline: Retrieval of homology data (human to other species) from PANTHER (http://www.pantherdb.org/)
0 stars 0 forks source link

Code cannot handle changes to downloaded file paths #8

Open SolomonShorser-OICR opened 3 years ago

SolomonShorser-OICR commented 3 years ago

For Release 76, Orthologs_HCOP.tar.gz was unzipped and untarred as a set of nested directories containing Orthologs_HCOP. The full path is " target4/famlib/dev/PANTHER16.0/lib_16.0/ftp/Orthologs_HCOP"

Unfortunately, the application (class: org.reactome.release.orthopairs.OrthologyFileParser, method: parsePantherOrthologFiles) can't handle this.

https://github.com/reactome/release-orthopairs/blob/c75f7be45d8110d7e00ec1fe7ba1148d91cb9cae/src/main/java/org/reactome/release/orthopairs/OrthologyFileParser.java#L39

SolomonShorser-OICR commented 3 years ago

Immediate solution was to just copy target4/famlib/dev/PANTHER16.0/lib_16.0/ftp/Orthologs_HCOP to the application's root directory, but a better long-term solution is needed. Preferably in the application, as it's file-handling looks fragile and prone to failure, but a fix in the Jenkinsfile or in the bash script that run the application might be OK.