Hi all, I only returned to looking at messytables after a good while today (see #136). It strikes me that PDF parsing support is a deeply absurd thing to try and do in a utility library like messytables.
I've taken half a dozen table-containing PDFs now to throw them at the PDF tool and none of them gave me a useful result.
On the other hand, having whacky features like this makes messytables feel bloated and heavyweight.
I wonder if there would be a strong argument against building an entry_points mechanism and then putting the PDF parser in it's own module? That would then also manage dependencies properly, instead of soft-failing upon call.
Hi all, I only returned to looking at messytables after a good while today (see #136). It strikes me that PDF parsing support is a deeply absurd thing to try and do in a utility library like messytables.
I've taken half a dozen table-containing PDFs now to throw them at the PDF tool and none of them gave me a useful result.
On the other hand, having whacky features like this makes messytables feel bloated and heavyweight.
I wonder if there would be a strong argument against building an entry_points mechanism and then putting the PDF parser in it's own module? That would then also manage dependencies properly, instead of soft-failing upon call.