Is there a way to limit number of tags, instead of characters ? booru filename

[tealkami]

Hi, I did look for this for hour or more, couldn't find it.

with examples :
"filename":"{id}---{tag_string_artist}---{tag_string_copyright[:50]|'':?/---/}{tag_string_character[:125]|'':?/---/}{filename}.{extension}"
"filename":"{id}---{tags_artist}---{tags_character[:125]}---{filename}.{extension}"

Instead using [:125] and have cut off some tags in the middle, use some other syntax for limiting number of tags (like... use only first 3 tags), I did check that most booru websites use space between tags in json, so it should be possible.
so instead {tags_character[:125]} i would use something {tags_character[:mt3]}

---

Update : This is full example if someone want to limit the number of tags (with comma) work only on danbooru:
"filename": "\fF {id}---{(', '.join(tag_string_artist.split(' ')[:1]) + '---') if tag_string_artist else ''}{(', '.join(tag_string_copyright.split(' ')[:1]) + '---') if tag_string_copyright else ''}{(', '.join(tag_string_character.split(' ')[:5]) + '---') if tag_string_character else ''}{filename}.{extension}"

---

and something less important:

is this |'':?/---/ the best/correct way to do :
|'' if the tag is not there make it empty
:?/---/ if the tag is there , after the tag add "---"

it does work, but it looks wonky to me.

[mikf]

For the time being, you'll have to manually split, cut, and re-join tags if you want to limit them by tag count. Regular format strings can't do this, so you'll have to use f-strings.

\fF {', '.join(tags.split(' ')[:3])}

Eventually, tags will always be a list and can then be used as {tags[:3]!S} in regular format strings.

---

As for your other question: tags_artist etc seem to always exist on Danbooru so you can drop the |'', but |'':?/---/ would be the correct way of doing what you described with regular format strings.

[tealkami]

Can you write that an example please ?
I have no idea how to add \fF {', '.join(tags.split(' ')[:3])} into this
"filename":"{id}---{tag_string_artist}---{tag_string_copyright[:50]|'':?/---/}{tag_string_character[:125]|'':?/---/}{filename}.{extension}"
for tag_string_artist to use only the first(1), and tag_string_character the first 5

[Hrxn]

@tealkami

Like this:
(suggestion removed, see comment by mikf below)

Eventually, tags will always be a list and can then be used as {tags[:3]!S} in regular format strings.

Definitely nice to use and will do what tealkami asked for, avoiding cutting off some tags further down eventually..

But using something like {tags|'Untagged':?/./J_/[:80]} still has benefits, in my opinion:

• Unlike the !S conversion, you can join tags with a custom separator value
• Limits the output length in the filename to actual 80, if you just go by tag count, you still can get screwed in rare cases where tags in the tag list are very long.

[mikf]

Like this:

"filename": "\fF {id}---{', '.join(tags_artist.split(' ')[:1])}---{tag_string_copyright[:50]|'':?/---/}{', '.join(tags_character.split(' ')[:5])}|'':?/---/}"

Unfortunately, that does not work. f-strings do not support | or :?/---/ or basically anything from docs/formatting, at least not with the documented syntax. It would be possible but incredibly long, complicated, and unreadable.

Each single tag category would be something like

{(', '.join(tag_string_character.split(' ')[:5]) + '---') if tag_string_character else ''}

---

I have since added some new tags metadata fields for Danbooru that provide them as a list of tags: 627ed79
(the old tag_string_character etc fields are still available)

With that, you can use {tags_character:[:5]J, /?/---/}.

[Hrxn]

With that, you can use {tags_character:[:5]J, /?/---/}.

Is it {tags_character:[:5]J, /?/---/} or {tags_character[:5]:J, /?/---/}?

[mikf]

Both work the same. [:5] can be placed directly after a field name tags_character[:5] or somewhere after :.

(tags_character[:5] was implemented first. Allowing [] after : was tacked on later (54525d2))

[Hrxn]

Unfortunately, that does not work. f-strings do not support | or :?/---/ or basically anything from docs/formatting, at least not with the documented syntax.

I was actually not aware.
Not even the ?<start>/<end>/ format specifier?

I can't give any recent examples, but I checked my own config and found an example like this:

"\fF [{','.join(a.title() for a in artist)}] {title!t:R.//}"

as the last entry in a list of strings in a "directory" setting.

Have not used this extractor in quite a while, but I remember that it did not throw any errors back then.
Is it possible that this simply fails silently?

[mikf]

It produces a nasty SyntaxError exception:

>>> f"{title!t:R.//}"
  File "<stdin>", line 1
    f"{title!t:R.//}"
                     ^
SyntaxError: f-string: invalid conversion character: expected 's', 'r', or 'a'

edit: {title!t:R.//} can be replaced with {title.strip().replace('.', '')} in f-strings