dlesl / gb-io Goto Github PK

So I went ahead and tested this library by parsing all GenBank entries from:

• https://ftp.ncbi.nlm.nih.gov/genbank/

The parser fails and raises an Exception on two entries out of >200000000 because of an "®" character.

Do you think one could fix this?
Thanks!

• https://ftp.ncbi.nlm.nih.gov/genbank/gbenv56.seq.gz

• https://ftp.ncbi.nlm.nih.gov/genbank/gbbct902.seq.gz

Hi there, thanks for your work on this project!

I got a compiler warning in a project that sues gb-io:

warning: the following packages contain code that will be rejected by a future version of Rust: nom v4.2.3
note: to see what the problems were, use the option `--future-incompat-report`, or run `cargo report future-incompatibilities --id 1`

This is fixed in subsequent version of nom. It looks like #7 already addresses this issue, but the PR is still open. As someone currently using this package, I figured I would make this issue to say 1) I appreciate this package and 2) it would also appreciate if #7 could be merged at some point.

Also happy to lend a hand if there are still any difficulties porting from nom 4 to nom 7.

When looking at a GenBank entry there often is interesting information in the "source" feature, which actually contains several subfields. From instance below one can see the "host" information:

I end up having to wrap the gb_io.Record class with another class to access this information. Also I end up constantly having to replace newlines with spaces.

class GenBankRecord:

    def __init__(self, rec):
        """Get a record instance from gb_io"""
        self.rec = rec

    @functools.cached_property
    def source(self):
        """Parse the first feature which is always 'source'"""
        return next(iter(self.rec.features))

    @functools.cached_property
    def fields(self):
        """Get all subfields of source."""
        return self.source.qualifiers.to_dict()

    @functools.cached_property
    def host(self):
        if 'host' in self.fields:
            return self.fields['host'][0].replace('\n', ' ')
        return ''

It would be nice to just do rec.host no?

Hi David!

In v0.7.0 I noticed there is no way to use the GbParserError type: the errors module is private, and this struct is not publicly re-exported in the ::reader module. I guess it would make sense to re-export it publicly there (the alternative, making errors public, only makes sense if you also had the errors from the ::seq module in there).