Save to Disk Template Issue with Series vs Standalone

[tsparr]

I have a Save to Disk template that seems like it should work; however, is throwing an error when actually running the task. It does work successfully for books with a series, but fails on the standalones with the following error.

Traceback (most recent call last):
      File "calibre\library\save_to_disk.py", line 282, in get_path_components
      File "calibre\library\save_to_disk.py", line 247, in get_components
      File "calibre\utils\formatter.py", line 1930, in unsafe_format
      File "calibre\utils\formatter.py", line 1847, in evaluate
      File "string.py", line 194, in vformat
      File "string.py", line 203, in _vformat
    ValueError: Single '}' encountered in format string
    
    During handling of the above exception, another exception occurred:
    
    Traceback (most recent call last):
      File "calibre\gui2\save.py", line 139, in do_one_collect
      File "calibre\gui2\save.py", line 147, in collect_data
      File "calibre\library\save_to_disk.py", line 288, in get_path_components
    ValueError: Failed to calculate path for save to disk. Template: {series:'re(ifempty($,field('title') & ' - ' & field('authors')),':',' -')'}/{series:'re(ifempty($,field('title') & ' - ' & field('authors')),':',' -')'}{series_index:0>2s| - | - {title}}
    Error: Single '}' encountered in format string

The template appears as if it should work in the editor, so not sure what I'm missing.

[chaley]

From the Advanced formatting section of the template language manual:

Do not use subtemplates (`{ … }`) or functions (see below) in the prefix or the suffix.

The last template expression

{series_index:0>2s| - | - {title}}

does exactly that. I suspect you want

{series_index:0>2s| - | - }{title}

It also isn't obvious why you want the series twice in the path

IMO a template of this complexity should be written in Template Program Mode (TPM). Not only is branching easier, TPM templates are much easier to debug.

[theducks]

You did not close series index before Title
{ } are always pairs and are not nested

[tsparr]

Originally Posted by chaley

From the Advanced formatting section of the template language manual:
The last template expression

{series_index:0>2s| - | - {title}}

does exactly that. I suspect you want

{series_index:0>2s| - | - }{title}

It also isn't obvious why you want the series twice in the path

IMO a template of this complexity should be written in Template Program Mode (TPM). Not only is branching easier, TPM templates are much easier to debug.

The problem if I do {title} outside of the template expression is that I would then have the title of the book in the filename twice if it's not in a series. The intent is to match up with a program I used called Kavita and their intended library structure of:
Library Root
┖── Series Name
┖── Series Name.epub

I'll look into the TPM template.

[tsparr]

Originally Posted by theducks

You did not close series index before Title
{ } are always pairs and are not nested

I think if this were the case, wouldn't the template editor complain? Again, it works fine for saving books in a series. It's only the standalone titles that it fails for. Simply removing {title} allows this template to work:

{series:'re(ifempty($,field('title') & ' - ' & field('authors')),':',' -')'}/{series:'re(ifempty($,field('title') & ' - ' & field('authors')),':',' -')'}{series_index:0>2s| - |}

[tsparr]

Sorry for the triple post at this point, but I ended up moving to Python because it just made way more sense to me. It's a little bit slower to run, but I'm only occasionally saving books out to disk.

python:
def evaluate(book, context):
	if book.series is None:
		author_str = ""
		for author in range(len(book.authors)):
			author_str += book.authors[author]
		book_str = book.title + ' - ' + author_str + '/' + book.title + ' - ' + author_str
	else:
		book_str = book.series + '/' + book.series + ' - ' + f"{int(book.series_index):02}" + ' - ' + book.title

	return book_str

[chaley]

Your python template smashes all the author names together without separation. Is that what you want? I think you want this:

python:
def evaluate(book, context):
	if book.series is None:
		author_str = ' & '.join(book.authors)
		book_str = book.title + ' - ' + author_str + '/' + book.title + ' - ' + author_str
	else:
		book_str = book.series + '/' + book.series + ' - ' + f"{int(book.series_index):02}" + ' - ' + book.title
	return book_str

or perhaps this, simplified somewhat using 'f' strings:

python:
def evaluate(book, context):
	if book.series is None:
		author_str = ' & '.join(book.authors)
		book_str = f'{book.title} - {author_str}/{book.title} - {author_str}'
	else:
		book_str = f'{book.series}/{book.series} - {int(book.series_index):02} - {book.title}'
	return book_str

For completeness, here is the equivalent template in GPM:

program:
	if $series then
		res = template('{series}/{series} - {series_index:0>2s} - {title}')
	else
		res = template('{title} - {authors}/{title} - {authors}')
	fi;
	res

For both templates: it isn't clear that you will get what you want if the series index is a floating point number such as 2.3. The python template drops the fraction. The GPM template includes it but doesn't respect the length. If you really want to ignore the fraction then the series index part of the GPM template should be

{series_index:0>2.0f}

[tsparr]

You're absolutely right in that I didn't want the names smashed together. I had actually already fixed that on my side, but didn't want to post a 4th time in a row and didn't see a way to edit my existing post.

Appreciate the simplified version and agreed it looks much cleaner. I was mainly just on a kick of adding each piece one at a time and validating the path in the template editor was showing as expected.

I don't currently have anything with a decimal in my library, so that's definitely an edge case I didn't consider. I'll manually adjust some metadata and play with that a little bit. I don't think I necessarily care if it's 00.5 vs 0.5. I just think it's cleaner to be 03 vs 3 especially when you get into series that go up into the double digits.

[tsparr]

Looks like this will handle the decimal series indexes:

python:
def evaluate(book, context):
	if book.series is None:
		author_str = ' & '.join(book.authors)
		book_str = f'{book.title} - {author_str}/{book.title} - {author_str}'
	else:
		if book.series_index.is_integer():
			book_str = f'{book.series}/{book.series} - {int(book.series_index):02} - {book.title}'
		else:
			book_str =  f'{book.series}/{book.series} - {book.series_index} - {book.title}'
	return book_str