How do you people manage/organize/index downloaded PDF & other type files?

B

bakul

Not specific to FreeBSD but surely some of you have run into similar problems?

I download a lot of PDFs but never got around to organizing them properly. Also, it is not just pdfs but epubs, tex files, text files, webpages, directories, images, videos etc. Is zotero the best tool for this? Any other useful tools? What I want ideally:
  • watch certain directories for new pdf (& other type) files
  • extract meta data & text data, the latter for text based search
  • maintain orig urls
  • show them in an "inbox", with suggestions as to where to file
  • use meta data to name the *symlink* (but not the original pdf)
  • filing via symlinks
  • originals may optionally get moved to a more permanent place
  • allow adding tags & notes
  • access from any unix machine
  • cli / web / gui interfaces, but prefer more modular cli approach
  • learn from prior actions
I think zotero does a lot of this.
 
Nobody really knows.

I create a title for the pdf that best identifies it if the title doesn't already do that. Then I file it in a directory where I think it best belongs; such as Drive/FreeBSD/

Later, if I'm looking for a particular subject, I'll try and find it using ls. If that doesn't work then I'll use find. If still no luck then grep

Everything should be as simple as possible.
 
drhowarddrfine said:
Later, if I'm looking for a particular subject, I'll try and find it using ls. If that doesn't work then I'll use find. If still no luck then grep
Click to expand...
I rely on my memory and MacOS spotlight but both are far from perfect and getting worse! Also with 7K-8k+ pdfs (some duplicates) it is a pain. These have been collected over decades and I haven't used a consistent naming scheme over that period. And I quickly scan downloaded tech reports but usually don't read them cover to cover. But may want to read/refer to them even years later.
 
Requires the Google ecosystem, but anyway:

notebooklm.google.com is a nice toy for pdfs. It is a LLM that you can throw a few documents in and it will answer from those documents. Questions, summaries etc. I find it to be quite good as long as the document doesn't rely too much on pictures.
 
None .. and I have total mess in those. I mean I keep them under e.books/ and e-learning/ for 20 years or so (yeah, dot vs hyphen mistake too).
But I still end up googling stuff when I need something even though I do have those books in my repository.

I'm considering just deleting it all, unlikely I will ever use them again. But for now I keep syncing them around when I move data.
 
Is that not a CMS (content management system)?

Content management system - Wikipedia

en.wikipedia.org en.wikipedia.org

I personally order my files in directories, nothing special.

Perhaps you can write your own program? I did something like that for my CDs that I backed up with cdda2wav.
I put the CD tracks in a directory hierarchy Music containing wav/mp3 files, a subdirectory may contain a file
Desc describing its contents. Then a program generates a sqlite3 database from the directory and the Desc files.
If you want to reach it from the internet, you will have to write a cgi script with authentication.
 
Just individual files organised by general subject in a directory tree. I like mupdf for viewing pdfs, it's very fast. I use vifm to give me a browser and viewer launcher. For a quick look at html I use 'links', also very fast. I do the same with mp4's etc. Vifm lets you associate a filename suffix with a specific launcher. You can do things like regex search on file names. It's fast and works well. I guess something like kde dolphin would be the gui equivalent, but I use windowmaker, I mainly use the terminal environment.

Vifm v0.14.3 - Vifm

vifm.info vifm.info

Yeah, it's easy for it to turn into a big mess, I don't spend a lot of time on it. In fact I usually use online search and read online information before I look at anything I've saved previously. Even if I remember I've saved something somewhere locally, it's often faster just to search for it online, having a fast broadband connection. Which is kind of scary, because you become dependent to some extent; and it makes you lazy. Perhaps it's the only realistic way of dealing with the information flood.
 
cracauer@ said:
notebooklm.google.com is a nice toy for pdfs. It is a LLM that you can throw a few documents in and it will answer from those documents.
Click to expand...
I have heard good things about notebooklm. I don't (yet) want to ask an LLM to interpret the contents of a PDF but may be a local LLM can help with categorizing.

This seems promising as a component: https://github.com/NanoNets/docstrange
Convert and extract data from PDF, DOCX, images, and more into clean Markdown and structured JSON. Plus: Advanced table extraction, 100% local processing, and a built-in web UI.
Click to expand...
 
Websites on local server, if you stay at home/office like me. Websites can be created by arguments and inside a website sub-arguments (sections and sub-sections). On the disk make the same tree of directories with the same name of the websites names, sections and sub-sections. If you travel and need to keep documents on working machine copy the same tree of directories.
 
As with others here, my ~/Downloads is mostly a mess. For books I want to keep and plan to read I use deskutils/calibre. It's great for organizing content, you can even add extra metadata fields, just set Calibre to use external PDF reader of your choice, internal PDF conversion can be messy.
 
Yeah, I bet a lot of us have a mess of stuff that we just never get to organizing. For what it's worth, I have pdfs directory, which has subdirectories of novels, shellscripting (which contains any tech book, not just scripting) and the shellscripting subdirectory has an epub section. And that was the result of me saying, Wow, let me organize this, then getting tired after awhile. I have a docxls directory and $DEITY help me if I ever need anything out of it, though I do have some subdirectories there, such as 2025medexp and 2025medreceipts. But then, looking through that directory in order to put something here, I had to laugh, I have 2025lease.pdf another called lease.pdf from some year or another. And yeah, those pdfs are in the docxls directory rather than the pdfs directory.

And when I try to get rid of old stuff, too often I get rid of something I'll need three months later, and go frantically through backup files, trying to remember the name.
 
You must log in or register to reply here.
Back
Top