search

omerwe / polyfun

PolyFun (POLYgenic FUNctionally-informed fine-mapping)
MIT License
96 stars 22 forks source link

`compression = None` on line 193 of parse.py #22

Closed davidaknowles closed 4 years ago

davidaknowles commented 4 years ago

Is this deliberate? It seems to mean you can't read compressed files since it overrides the automatic compression detection.

omerwe commented 4 years ago

Thanks for the question. I'm pretty sure that this flag is irrelevant because pandas can infer compression automatically (this code is taken directly from ldsc, which was written ~6 years ago before pandas could do this). Do you have an example where this flag is required? I just verified that the code can work with .gzipped files on my system (using pandas 0.25.3, though support for compressed files existed for several years I think).

On Mon, Feb 17, 2020 at 3:47 PM David A Knowles notifications@github.com wrote:

Is this deliberate? It seems to mean you can't read compressed files since it overrides the automatic compression detection.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/omerwe/polyfun/issues/22?email_source=notifications&email_token=ACNCB423V6DWL5EAGHMAKYTRDLZVTA5CNFSM4KWYIT32YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IOEILMA, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACNCB43DMXYSDGJLRWWIW6LRDLZVTANCNFSM4KWYIT3Q .

davidaknowles commented 4 years ago

The problem is compression = None gets passed to pandas which overrides the auto compression-type detection (at least on my system). The compression = "infer" is the default for pd.read_csv.

omerwe commented 4 years ago

You're right, my bad --- my test was incorrect. I just pushed an updated version that disables this line (I made sure that I can now read gzipped ldscore files). Hope it's fine now!

On Mon, Feb 17, 2020 at 7:35 PM David A Knowles notifications@github.com wrote:

The problem is compression = None gets passed to pandas which overrides the auto compression-type detection (at least on my system). The compression = "infer" is the default for pd.read_csv.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/omerwe/polyfun/issues/22?email_source=notifications&email_token=ACNCB447SO5ZU3WKDGH3QL3RDMULXA5CNFSM4KWYIT32YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEMADBTY#issuecomment-587215055, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACNCB4Z4GL3GZBSM5JK2IHDRDMULXANCNFSM4KWYIT3Q .