When an error occurs, give information about location of error in GFF3 file

[nbisweden]

When an error occurs, give information about location of error in GFF3 file

[Juke34]

Thank you for your feedback. That would be definitely useful, we will add this information.

[Juke34]

Done.
Will be included in the next release (not planned yet. I will wait for more modifications). To have this improvement before the next EMBLmyGFF3 release, install EMBLmyGFF3 using git clone:

pip uninstall EMBLmyGFF3 #could be needed twice
git clone https://github.com/NBISweden/EMBLmyGFF3.git
cd EMBLmyGFF3/
python setup.py install

If you already installed the tool using git clone you will have to do so:

pip uninstall EMBLmyGFF3 #could be needed twice
git pull  #from the EMBLmyGFF3 git repository. It will update the code
python setup.py install #from the EMBLmyGFF3 git repository

[loraine-gueguen]

Thanks!
Would it be possible to catch all exceptions with a clue on the location into the GFF3 file ? The error examples number 1 and 3 in my comment above are not addressed by this commit, unless I'm mistaken (although, for error number 1, I admit that I can search for non ASCII character by a simple shell command)

[Juke34]

Actually having a CDS not multiple of three is allowed. So it is not an error, but translation method (which is part of biopython and not EMBLmyGFF3) raise a warning to inform you.

To fulfil your request I have now added an internal (in EMBLmyGFF3) Warning with the CDS ID when the CDS is not multiple of three. I'm not sure about the pertinence to add this warning, but it is now done.

For the following error

UnicodeDecodeError: 'ascii' codec can't decode byte 0xef in position 5: ordinal not in range(128)

I don't have any test case to work with. Could you provide me a sample of your data ?

[loraine-gueguen]

I had a look on your commit (6816aa8). I agree with you that this may not be pertinent to add this warning... I thought it could be possible quite easy to catch all kind of errors indifferently from biopython and provide the feature ID where it occurs, but maybe I am mistaken... I thank you anyway.

For the ASCII error, you should have the error only by copy-paste this character into an attribute of the 9th column of the GFF3: "α" (greek alpha). Tell me if not.

Thank you

[Juke34]

The last commit (7398bd4) should have addressed the UnicodeDecodeError error.
Thank you for having catched this issue.