Templates: various questions not worth their own thread

[ownedbycats]

Additional question: The GPM template returns backslashes. Do I need to use a re(_,/) when this is used in a save template as with a non-composite column?

[chaley]

Originally Posted by ownedbycats

Additional question: The GPM template returns backslashes. Do I need to use a re(_,/) when this is used in a save template as with a non-composite column?

Perhaps. Using backslashes works for paths in Windoes unless they appear in a context where they might be considered to be escapes. I also don't know that you can mix slashes and backslashes in the same path.

I think best would be to use slashes exclusively, which might argue for the re() you mention. However, I am not sure that save templates can use functions because they use a limited formatter, which if so means you can't use re().

Try it and see...

[BetterRed]

Suggestion: an has_data function, analogue of has_cover.

[ownedbycats]

This is something I would use.

[chaley]

Originally Posted by BetterRed

Suggestion: an has_data function, analogue of has_cover.

Originally Posted by ownedbycats

This is something I would use.

Such a function must query the file system, so it won't be fast. My suspicion is that you want to use this for column icons, which may make the performance unacceptable.

Implementation choices:

1. Fastest would be to check if "data/" exists, regardless of whether or not there is anything in it. I suspect the false positive rate would be unacceptable.
2. Next would be to check if "data/" contains anything.
3. Equally performant as 2, or perhaps I should say non-performant, would be to permit checking if particular files exist by matching the list of files against a regular expression.

This python template does check #2. Make it a stored template and try it to see if it is sufficiently performant.

python:
def evaluate(book, context):
	db = context.db.new_api
	files = db.list_extra_files_matching(book.id)
	return '1' if files else ''

[ownedbycats]

I want to run it in a saved search, so speed wouldn't have been much an issue. Will try the python though.

[chaley]

Originally Posted by BetterRed

Suggestion: an has_data function, analogue of has_cover.

Originally Posted by ownedbycats

This is something I would use.

I should point out another problem with using some API to have column icons for books with extra data: the answer might be wrong. The answer from column icon templates is cached in the GUI until something changes. Unfortunately for this: extra data can be added/removed outside of calibre, so calibre doesn't know that something changed. The "answer" will be wrong until something does change or the GUI is refreshed.

[BetterRed]

No worries, I'll continue doing what I've been doing with my !Attachment folders, putting a file:/// link to the ...author/book/data path in the #links/Links long text columns.

I guess I assumed there'd be a has_data BOOL in the books table. Should have looked.

BTW the inter-library peek (book-details) is exactly what I wanted, thanks for that.

[ownedbycats]

does has_cover stored in the database, or just a check for cover.jpg?

[chaley]

Originally Posted by ownedbycats

does has_cover stored in the database, or just a check for cover.jpg?

It checks the database. If you delete the cover by hand then the database can be wrong. This is why "Check library" has a cover check.

[ownedbycats]

That makes sense. I don't think there's an in-program way to delete extra files (at least that I found). So setting a database entry on adding an extra file isn't very reliable.

[ownedbycats]

@chaley - I saved the template to has_data(). I did change '1' to 'Yes' so it would match has_cover().

It worked pretty quickly in a saved search so as an extra test I added it to this:

program: 

formats = approximate_formats();

   strcat
   (
		if "epub" inlist formats then 'formats_epub.png:' fi,
		if "pdf" inlist formats then 'formats_pdf.png:' fi,	
		if "azw3" inlist formats then 'formats_azw3.png:' fi,
		if "mobi" inlist formats then 'formats_mobi.png:' fi,
		if "cbz" inlist formats then 'formats_cbz.png:' fi,	
		if "txt" inlist formats then 'formats_txt.png:' fi,	
		if "paperbook" inlist formats then 'formats_paperbook.png:' fi,
		if "overdrive" inlist formats then 'formats_overdrive.png:' fi,	
		if has_data() == 'Yes' then 'save.png:' fi,
		if list_difference(formats, 'epub, pdf, azw3, mobi, cbz, paperbook, overdrive, txt', ',') then 'formats_unknown.png' fi	
   )

It doesn't seem to be any slower than usual so think I will keep.

[ownedbycats]

@chaley - After updating Calibre to source, has_data() broke with this error:

EXCEPTION: Error in function evaluate on line 4 : AttributeError - 'Cache' object has no attribute 'list_extra_files_matching'

What changed?

EDIT: Maybe this change https://github.com/kovidgoyal/calibr...095525b6e24e92

[chaley]

Originally Posted by ownedbycats

@chaley - After updating Calibre to source, has_data() broke with this error:

EXCEPTION: Error in function evaluate on line 4 : AttributeError - 'Cache' object has no attribute 'list_extra_files_matching'

What changed?

EDIT: Maybe this change https://github.com/kovidgoyal/calibr...095525b6e24e92

Yes, because of that commit. The API was improved for this use case, making it faster.

The template should be:

python:
def evaluate(book, context):
	db = context.db.new_api
	files = db.list_extra_files(book.id, pattern='data/**/*', use_cache=True)
	return 'Yes' if files else ''

Note: I will be adding template functions for queries about extra files.

[ownedbycats]

Thanks

Somewhat related: What happens if a new template function shares the name as an already existing saved template?