3 ms·
I simply print to PDF, anything interesting I've read online. So now I've got 30+ years of my own private offline Internet experience. Some 80,000+ files in a
by aa-jv 8mo ago
I simply print to PDF, anything interesting I've read online. So now I've got 30+ years of my own private offline Internet experience.
Some 80,000+ files in a directory represents an awesome database of knowledge. "$ ls inux" to find anything Linux-related, etc.
One of these days I'll get around to setting up some ML tool that will tell me all the things I didn't already osmose from the archive .. and maybe long after I'm gone, in some hole in a wall of some grimy back alley somewhere, there'll be a ML version of me embedded in a brick, ready to have the conversation well into the future ..
- dewey 8mo agohttps://github.com/paperless-ngx/paperless-ngx https://github.com/paperless-ngx/paperless-ngx might be a nice rabbit hole for you, drop the files in there and it'll be OCR'ed and searchable. There's also some AI projects you can give access to paperless to achieve your use case.
- aa-jv 8mo agoThanks for that - I've tried it off and on over the years and it looks like its turning into quite an effective tool these days - so I'll give it another try. I've considered building my own tools to manage these files, but its always nice to see the approach others are taking too ..
- EllaneW 8mo agoI’m interested in your naming convention for these files (so many files!). Do they all have the date you saved them on, up front and centre?
- aa-jv 8mo agoThey have the datetime they were saved, and the filename is usually derived from the page title - this often needs a bit of manual tweaking on my part if the original author hasn't named it something sensible, but its a minor action to take. Every year or so I 'sanitize' the filenames of the archives, to make sure I can still search them with normal command-line (bash) tools, removing special % ^ # $ & chars, and so on .. A lot of other info is stored in the PDF metadata, too ..