Bug 337982

Created attachment 88251 [details]
First screenshot of feature

Created attachment 88252 [details]
Second screenshot demonstrating the feature

(In reply to adr.fantini from comment #0)
> It would be great if KBibTeX included some kind of automatic PDF renaming as
> other referencing programs do.
The current git version has already such a feature. Please see the attached screenshots as referenced. You may want to test the git version to see if this feature satisfies your request.

(In reply to Thomas Fischer from comment #3)
> (In reply to adr.fantini from comment #0)
> > It would be great if KBibTeX included some kind of automatic PDF renaming as
> > other referencing programs do.
> The current git version has already such a feature. Please see the attached
> screenshots as referenced. You may want to test the git version to see if
> this feature satisfies your request.

That looks great! Thanks as always for adding features. But I miss the possibility to automatically select the new filename based on paper informations (I usually do year_principal-author_title.pdf, or %y_%a_%t.pdf). Is it possible to implement this automatically, or at least to be able to manually automate it on the selected papers? (eg. repeat the same action for N selected papers)

> That looks great! Thanks as always for adding features. But I miss the
> possibility to automatically select the new filename based on paper
> informations (I usually do year_principal-author_title.pdf, or
> %y_%a_%t.pdf). Is it possible to implement this automatically, or at least
> to be able to manually automate it on the selected papers? (eg. repeat the
> same action for N selected papers)
Renaming suggestion are based on a BibTeX's entry. There is a mechanism to automatically rename entries to a certain pattern based on authors, year, title etc. For an entry '@article{smith2010inference, ...}', the corresponding PDF file will be named 'smith2010inference.pdf'. Is that good enough for you?

> Renaming suggestion are based on a BibTeX's entry. There is a mechanism to
> automatically rename entries to a certain pattern based on authors, year,
> title etc. For an entry '@article{smith2010inference, ...}', the
> corresponding PDF file will be named 'smith2010inference.pdf'. Is that good
> enough for you?
Back to my question from February: Is renaming according to the BibTeX entry's id good enough for you?

(In reply to Thomas Fischer from comment #6)
> > Renaming suggestion are based on a BibTeX's entry. There is a mechanism to
> > automatically rename entries to a certain pattern based on authors, year,
> > title etc. For an entry '@article{smith2010inference, ...}', the
> > corresponding PDF file will be named 'smith2010inference.pdf'. Is that good
> > enough for you?
> Back to my question from February: Is renaming according to the BibTeX
> entry's id good enough for you?

Whoops, I'm sorry I missed it. I think it would be better to be able to choose the renaming scheme: some prefer to have pdfs stored by fisrt author name, some by bibtex id, some by year/title... it comes down to personal preference. Having the bibtex id rename feature is certainly better than nothing but I still feel that the personalized renaming would be much nicer :)

Thanks for your great work Fischer, and the discussion between adr.fantini and Fischer.

-----

I have used kbibtex for 2 years, then switch to mendeley for 2 years, mainly for the benifits for renaming (still buggy) and annotations.
Now I want to switch back due to that nearly no hope for a better support for group with annotations, as that is a major business for mendeley.
The users do not have direct control on the files.

In specific, I need the separated annotations saved in text file, such that docfetcher or baloo can index.

-----

Come back on the renaming.
I do not think reaming the file name with bibkey or author/year is the best idea.
This solution does not serve well on give much information as bibtex.
After all, we indeed need bibtex to help us free off the burden on remember all the metadata of a file.

Furthermore, the bibkey or author/year is not constant, and it does not have direct relationship with the pdf itself. This would be a disaster when using okular.

Thus, I think renaming the files with just its md5/sha1 is a good option.
It brings the benefits that all your annotations and bookmarks stay with your pdf/epub/djvu, when using okular.
Just leave the metadata to bibtex. After all, that is what bibtex expected to do.
(Do one thing at a time and do it best.)
-----

In short, it would be great if KBibTeX could rename the file automaticly with its md5.

Hi. I'm new to KDE and thrilled to see I don't have to use JabRef. I made an account just to comment on this feature request. I'll explain my use case and my current solution and why I'd like to see more control or renaming in KBibTeX.

Use case: I currently have a Library.bib with each entry linked to a PDF. I have 2000 authors and 3500 papers. I sync my entire library to a tablet and read papers there sometimes, and scrolling through a folder of 3500 pdfs, or even 2000 LASTNAME folders (each with 1-10 PDFs therein) would be problematic.

Solution: I have PDFs auto-filed into ~/Documents/Papers/L/Lastname/Lastname_Year_First-three-words.pdf.

This works well at a number of levels - It is quick to browse and find a PDF on a tablet, but also on the CLI or in Dolphin or any GUI/CLI I have used. Furthermore, having the Name_Year_TitleWords in the filename is also useful for getting a reminder of a papers contents without actually opening it, or for helping guess the cite key, etc.

This is all made possible, easily, via the AutoFile options in BibDesk, which let you choose from 10s of options of the BibTeX fields, with modifiers (e.g. n characters or words, replacements, etc.). You can see the UI here https://bibdesk.sourceforge.io/manual/BibDeskHelp_76.html and the AutoFile formatting options here: https://bibdesk.sourceforge.io/manual/BibDeskHelp_91.html#SEC159

In my opinion, this is the best solution because it covers all possible options. I would like to see this implemented in KBibTeX, and am willing to help. I'm a programmer, but new to KDE and would have a steep learning curve to implement this myself.

I'm also hopeful that it doesn't need to be implemented. Perhaps someone can explain that a folder with 3500 PDFs in it actually does make sense and I don't need more complicated auto-filing and linking options than the current feature set provides.

Interesting... I just discovered that the UI interface for "complex" PDF renaming (and filing?) *already exists* in KBibTeX. It is the same as the "Id Suggestions" interface, that lets me build custom keys. Maybe this means it is a tractable problem for a newbie to code up...

(In reply to Ken Mankoff from comment #10)
> Interesting... I just discovered that the UI interface for "complex" PDF
> renaming (and filing?) *already exists* in KBibTeX. It is the same as the
> "Id Suggestions" interface, that lets me build custom keys. Maybe this means
> it is a tractable problem for a newbie to code up...
There was some code available to associate a PDF file with an element. This code has been improved further; the current state can be tested with KBibTeX's 'master' branch.

You can test this code without interfering with your default KBibTeX installation by following the "quick start" instructions:
https://userbase.kde.org/KBibTeX/Development#Quick_Start_to_Run_KBibTeX_from_Git

To test the Git version containing the fix for your bug, please run the following command (all in one line):
bash run-kbibtex.sh