UnboundLocalError: local variable 'new_value' referenced before assignment

[nbisweden]

UnboundLocalError: local variable 'new_value' referenced before assignment

[Juke34]

Could you relaunch with '-vvv' option to have more verbose information ?
Thx

[Asutu]

Of course, please attached log file.
EMBLmyGFF3.test.log

[Juke34]

Ok then it's related to a bug we are aware of: #32
We shouldn't stop the tool but just raise a WARNING and not report the value in the output.
It's because the KEGG database (from the dbxref attribute) is not among accepted databases at ENA.
Here the list of databases accepted: https://www.ncbi.nlm.nih.gov/genbank/collab/db_xref/

A way would be to remove those information or edit the legal_dbxref.json to add those DB among the list of accepted one. In the second case you will create an EMBL file that do not strictly follow the ENA specification and I'm no sure how ENA will handle it.

Another thing, I realised recently that ENA do not keep DB_xref information when it's processed by their pipeline. I'm still waiting for information about it from them, but the only information I got is: "We get rid of DB_xref information and redo our own functional information"...
At the end I don't really know how the db_xref data is handle by ENA.

[Asutu]

Great, so I guess the safest way for now is to remove this attribute from the GFF. Will try to do that.
Many thanks for the prompt support.

[Juke34]

As your annotation comes from NBIS I guess several other DB reported in the dbxref attribute are not supported by ENA. Here is the list of DB we can report that are not accepted by ENA:
"KOG",
"KEGG",
"MetaCyc",
"Reactome",
"UniPathway"

[Asutu]

Removing these values from the dbxref attribute solves this error.

[Juke34]

fixed in release 1.2.5