Closed simonwoerpel closed 3 years ago
Hi @simonwoerpel! Thanks for noticing and fixing the bug.
The failed test is probably because of the content-type
case change. After serialization the headers are sensitive to case. I'm merging this in and will fix that in a separate commit.
Without this fix, because of using
os.path.split
instead ofos.path.splitext
, a crawled pdf file namedWD-7-028-21-pdf-data.pdf
ends up in the data folder like:
WD_7_028_21_pdf_data.WD_7_028_21_pdf_datapdf
which is probably not as it's supposed to be ;)