[GUI Plugin] FanFicFare

[JimmXinu]

Originally Posted by ownedbycats

Question: is it possible to do a replace_metadata based on an author URL (rather than the author name)?

edit: actually, that might not work, as AO3 seems to use the author name in the URL rather than some sort of uid. Since it's one story I guess story URL would work.

Possible, I suppose? But I wouldn't recommend it without a very good reason.

I wouldn't recommend changing authorId either, as at least a few adapters construct authorUrl from authorId and other bits.

[JimmXinu]

New Test Version Posted

2022-04-25
- adapter_phoenixsongnet: Remove login code and changes for static author links.
- adapter_webnovelcom: Fixes for site changes. #828
- adapter_fictionhuntcom: Fixes for site changes.

[ownedbycats]

Originally Posted by JimmXinu

Possible, I suppose? But I wouldn't recommend it without a very good reason.

I wouldn't recommend changing authorId either, as at least a few adapters construct authorUrl from authorId and other bits.

Specifically, the author of a fanfic seems to change their username every time they update their fic and I wanted to retain the original one for save template reasons.

Thankfully I was able to put a match-anything replace for the author under a [https://archiveofourown.org/works/xxxx] section. Is this the best way to handle it?

[chaley]

Originally Posted by JimmXinu

That was my thought as well from the initial error "TypeError: '<' not supported between instances of 'int' and 'NoneType'".

But the FFF code only calls set_marked_ids() with a dict that contains id->string value, or an empty dict() to clear all marks. No where does it set a boolean or a None into the dict. And I haven't been able to reproduce the reported issue.

I think I know what is happening but I won't be able to work on it for a week or so. The marked set can apparently contain the ids of deleted books. These might change to a book id of None. Alternatively, someone passed a set containing {None:'aaa'}. In any event, a key of None should be filtered out.

This new named marking UI conflicts a bit with the way Reading List PI and FFF have used a unique, numbered named mark on each book to both mark and order. I wonder if having hundreds or thousands of uniquely named marks (from viewing a Reading List) will cause problems? My 5k+ library fills the screen on 'Show marked books with text' but otherwise is fine.

I will change it to show only the first 40 (arbitrary). More than that were probably produced programmatically and aren't useful, and can be done with search.

FYI, the new 'Mark with text label' action will allow an all whitespace label. And treats one space and two spaces as different labels. Doesn't seem to hurt anything, but may not be desired.

That isn't good. I will fix it.

[JimmXinu]

Originally Posted by chaley

I think I know what is happening but I won't be able to work on it for a week or so. The marked set can apparently contain the ids of deleted books. These might change to a book id of None. Alternatively, someone passed a set containing {None:'aaa'}. In any event, a key of None should be filtered out.

Originally Posted by chaley

I will change it to show only the first 40 (arbitrary). More than that were probably produced programmatically and aren't useful, and can be done with search.

That works. People naming them manually is really a different use case.

[ownedbycats]

I noticed the coloured pins with reading list and FFF too. I wondered if it might be possible to somehow set, e.g. reading_list_to_be_read_\d{4} to all be the same colour.

[JimmXinu]

Originally Posted by ownedbycats

Thankfully I was able to put a match-anything replace for the author under a [https://archiveofourown.org/works/xxxx] section. Is this the best way to handle it?

Nothing better comes to mind.

[JimmXinu]

Originally Posted by ownedbycats

I noticed the coloured pins with reading list and FFF too. I wondered if it might be possible to somehow set, e.g. reading_list_to_be_read_\d{4} to all be the same colour.

I thought that it would also be nice if Mark grouped named marks ignoring the trailing _\d{4,} RL and FFF uses, but it would be pretty specifically for those plugins. ({4,} because it's not impossible for there to more than 4 digits.)

Is there a thread for discussing the mark feature changes? I didn't immediately see one.

[ownedbycats]

I didn't see one either; I believe the change came from a request on launchpad. You could probably make a thread in the dev subforum.

[ILB]

Originally Posted by JimmXinu

ILB - What are your FFF 'Mark' settings on the FFF General config tab, and how did you produce the error? Were you doing anything else that used Marks at the time?

I don´t know if you still need this information but I got that error with "get story urls from mail"


My setting in the basic tab from FFF

Spoiler:

[chaley]

Originally Posted by JimmXinu

I thought that it would also be nice if Mark grouped named marks ignoring the trailing _\d{4,} RL and FFF uses, but it would be pretty specifically for those plugins. ({4,} because it's not impossible for there to more than 4 digits.)

There are only 6 pin colors so they are recycled for the various names.

I think that grouping marks to pins is the beginning of a never-ending chain of enhancement requests. I have to ask "What does the current implementation break?" And "How many people care?"

Is there a thread for discussing the mark feature changes? I didn't immediately see one.

No. Feel free to make one.

[JimmXinu]

ILB - I'm still not able to duplicate it. In the meantime, you could turn off marking in FFF, or switch back a version of Calibre and see if that helps.

chaley - You're not wrong--I was asking about a thread to make another suggestion.

[JimmXinu]

New Test Version Posted

2022-04-26
- Py2 vs py3 fix #828

Only matters for webnovel.com on Calibre 4 and below.

[ownedbycats]

Originally Posted by JimmXinu

Nothing better comes to mind.

While a hypothetical for now, is there a way to share settings between several story URLs rather than duplicating them?

[JimmXinu]

The only current feature I can think of is _filelist (wiki link).