Date from Filename
Hi,
sometimes I am getting a document without a Date on it, or a document, i know that Docspell will have problems to detect it.
Therefore I scan the Document with Genius Scan App, name it e.x. 2022-03-01 My crazy document and share it with docspell-share app.
Unfortunately, Docspell is not proposing this date.
It would be cool to have a "module" to parse the filename and add the date as primary proposal since it has been "manually approved" by naming the file accordingly.
I like the idea. There is no filename specific processing going on at all currently. I think a quick approach is to prepend the filename to the extracted text, that should then at least propose the date (actually I thought I did this, but obviously that's not the case). On a second step, a more file specific processing could be implemented (treating everything that comes out of it as "confirmed" perhaps)
One way to consider would be a hook that gets fed the filename, and can output a new filename, and a JSON dictionary with metadata it was able to extract.