Generalize feature url retrieval in refactor_reportingtools_table.rb, closes #13

[fischer-hub]

• In the RNAseq_analysis_with_DESeq2_full_extended.html the ensembl url of current feature is now retrieved dynamically by constructing the base url first (get species name from ensemble REST API call and url prefix from mapping file) and adding the according feature suffix containing the ensemble stable ID
• added mapping file containing the listed species from ensemble.org and all the sister sites mapped to their sister site url prefix

[hoelzer]

really great @fischer-hub . So basically, this will now check for the base URL and three letter species code automatically and simply based on the input GTF annotation file, right?

What happens if the user provides a non-Ensembl annotation? Or let's say analyses some data of a non-model species with no available Ensembl annotation? Can the pipeline still deal w/ that and just does not provide any links to Ensemlb in the final output?

[fischer-hub]

Yep the species is identified through the ENS ID in the annotation file.
Right, in case the gene ID is not a valid ENS ID (not ensebml or not in DB) the rest api call will return a bad request and the ens_url variable will just be set to false, so the output will not have links to ensembl.

[hoelzer]

Yep the species is identified through the ENS ID in the annotation file. Right, in case the gene ID is not a valid ENS ID (not ensebml or not in DB) the rest api call will return a bad request and the ens_url variable will just be set to false, so the output will not have links to ensembl.

great! last question to this PR I think: what happens when no internet connection is available OR the API call is e.g. blocked for whatever reason (e.g. thinking about restricted government servers... ;) ).

[fischer-hub]

Good question, I think it should work because we check for a valid response e.g. response code 200 or smth like that and in all other cases the url var will just be false, but I'd have to check again that the ruby script doesn't just crash without connection..

[hoelzer]

Good question, I think it should work because we check for a valid response e.g. response code 200 or smth like that and in all other cases the url var will just be false, but I'd have to check again that the ruby script doesn't just crash without connection..

I mean should be rare: its 2022 and almost all the time users should have a solid internet connection ;) If you can try this out, nice, if not also fine. Please feel free to merge this PR then! thx!

[fischer-hub]

Good question, I think it should work because we check for a valid response e.g. response code 200 or smth like that and in all other cases the url var will just be false, but I'd have to check again that the ruby script doesn't just crash without connection..

I mean should be rare: its 2022 and almost all the time users should have a solid internet connection ;) If you can try this out, nice, if not also fine. Please feel free to merge this PR then! thx!

without connection the ruby module actually crashed! this is now catched and will just behave as if no url was found -> report will just not contain links!

[hoelzer]

Great, ok good you tested this! Please feel free to merge this PR