Transforming Filenames With Special Characters

[Rand Brittain]

Is there a way to use templates or plugboards to set Calibre up so that when I send a file to a device that has an illegal character in the metadata (specifically, a colon), it transforms that character into something other than _?

I'm trying to use Calibre to manage my collection of RPG PDFs, and RPGs love having titles with colons in them, like Vampire: the Masquerade and its ilk. I'd like to be able to set things up so that the transfer sends these files to device with the filename of, more or less "Vampire — the Masquerade" instead of "Vampire_ the Masquerade". That is, it turns a colon into space-emdash-space.

Can Calibre do this automatically? I've looked at the documents but I'm not too great at regular expressions.

[theducks]

Colon is illegal in windows, there is a FAQ on WHY Calibre prefers simple ASCII names (it is only allowd as part of a device (A:, Com1:, Aux: LPT1: )
Other symbols have meaning to various OS. Calibre works across many OS, so K.I.S.S. is the rule.

[BetterRed]

I think the OP knows why the '_' (underscore) character is substituted for disallowed characters (e.g. <>:"/\|?*) in file names.

What they want to do is to use another substitute character in Save to Disk, Send to Device templates and plugboards etc. Or even better a substitute string e.g. ' — '.

Personally, I would get rid of the ':' in the Title when I added the book, either via use of a subtitle column or by changing the ':' to ' — '. But that wouldn't suit everyone.

BR

[davidfor]

If you are fiddling with the metadata that will be inside the book, you can use the metadata plugboard to make the changes you want. But, these won't affect the file names.

For the file names, the save template is used. There is a default save template on the "Sending books to devices" page of the preferences. That will be used for any device you have not done any configuration to. As this is probably a devices specific thing, configure this in the driver used by your device. The simplest way to do that, is to connect the device, wait for the device jobs to run and then right-click the device button and select the option to configure the device. That will make sure you get the right driver.

The default save template is:

{author_sort}/{title} - {authors}

For what you want, you could do is:

{author_sort}/{title: re(: , - )} - {authors}

Though I would do it in general program mode as:

program:
strcat($author_sort, "/", re($title, ":" , " -"), " - ", $authors)

Which I find that version a lot easier to read.

In both cases, the "re" is a regex replacement, so the parameters are regex to match and the string to replace all the matches with.

After this has been executed, calibre will still sanitise result. Which will make sure whatever is produced is valid on the filesystem and the device in question.

[Rand Brittain]

Originally Posted by davidfor

If you are fiddling with the metadata that will be inside the book, you can use the metadata plugboard to make the changes you want. But, these won't affect the file names.

For the file names, the save template is used. There is a default save template on the "Sending books to devices" page of the preferences. That will be used for any device you have not done any configuration to. As this is probably a devices specific thing, configure this in the driver used by your device. The simplest way to do that, is to connect the device, wait for the device jobs to run and then right-click the device button and select the option to configure the device. That will make sure you get the right driver.

The default save template is:

{author_sort}/{title} - {authors}

For what you want, you could do is:

{author_sort}/{title: re(: , - )} - {authors}

Though I would do it in general program mode as:

program:
strcat($author_sort, "/", re($title, ":" , " -"), " - ", $authors)

Which I find that version a lot easier to read.

In both cases, the "re" is a regex replacement, so the parameters are regex to match and the string to replace all the matches with.

After this has been executed, calibre will still sanitise result. Which will make sure whatever is produced is valid on the filesystem and the device in question.

Yes, thank you; this is just what I needed!

What's the correct way to do the second format if what I was currently using was:

{author_sort}/{#line}/{#subline}/{#edition}

Would it be:

program:
strcat($author_sort, "/", $#line, "/" $#subline, "/", $#edition, "/", re($title, ":" , " -"), " - ", $authors)

[davidfor]

Originally Posted by Rand Brittain

Yes, thank you; this is just what I needed!

What's the correct way to do the second format if what I was currently using was:

{author_sort}/{#line}/{#subline}/{#edition}

Would it be:

program:
strcat($author_sort, "/", $#line, "/" $#subline, "/", $#edition, "/", re($title, ":" , " -"), " - ", $authors)

Other than the missing comma before $#subline, that should be correct.

But, I feel like there should be a better way to do that, but, I can't see a "format_as_path" or the equivalent of "os.join()".

[Rand Brittain]

Thanks a bunch!

[Rand Brittain]

Hm, after testing this, it looks like doing this with an endash actually results in the filename being transformed as requested, but it outputs a hyphen instead. If I use an emdash, it outputs two hyphens.

Odd! I know the filesystems I use are okay with endash and emdash, but maybe some of them aren't and so Calibre won't use them? Or could it be a bug?

[DNSB]

Originally Posted by Rand Brittain

Hm, after testing this, it looks like doing this with an endash actually results in the filename being transformed as requested, but it outputs a hyphen instead. If I use an emdash, it outputs two hyphens.

Odd! I know the filesystems I use are okay with endash and emdash, but maybe some of them aren't and so Calibre won't use them? Or could it be a bug?

Very likely that calibre is using the ASCII character set which does not include an en or em dash. As mentioned above, this helps to avoid issues with OS where one of those characters might be invalid in a file name.