Help with "Guessing metadata from file names"

[Justy]

I have over 800 eBooks which I really really want to import into Calibre. However I'm scared to, I previously attempt to load my library into Calibre using one of the 0.5 released and was dismayed at how many, over half, of my collection had no author or title information at all. These eBooks all have authors and titles on my Cybook and BeBook, so I'm confused. The vast majority of my eBooks are Mobipocket .prc format, I've been told this could be part of the problem.

I would love to use the 'Guess metadata from file name' option but I have 2 distinct naming conventions, one with series information and one without. Since the Gui only has 1 option, I'm not sure how to handle this.

• with series: Genre_AuthorLastName_Series_Title
• w/out series: Genre_AuthorLastName_Title

Should I start adding an extra "_" to all the titles without series, or is there an easy way to program this using a script that runs on command line? I'm not afraid of downloading new programs and having a go at it if someone can get me started. I would also love to be able to retain the Genre information as a tag during import if that is at all possible.

[kovidgoyal]

You can actually write a regex to handle both those cases. I don't have the time to do it for you, but hopefully someone who does will come along. There's also a thread on this forum about developing file name import regexes that you may find useful.

[user_none]

The following regex will match both cases provided they're no _ within each section.

[^_]+_(?P<author>[^_]+)_((?P<series>[^_]+)_)?(?P<title>[^_]+)

At least it matches your Genre_AuthorLastName_Series_Title and Genre_AuthorLastName_Title examples.

[Justy]

Originally Posted by user_none

The following regex will match both cases provided they're no _ within each section.

Thank you! I'll give it a try tonite.

[Justy]

Originally Posted by user_none

The following regex will match both cases provided they're no _ within each section.

[^_]+_(?P<author>[^_]+)_((?P<series>[^_]+)_)?(?P<title>[^_]+)

Thanks for the code, I finally got a chance to test it out today. It works great except for the fact that I forgot to mention that I put in the series index after the series name.

• with series:
  [**]Genre_AuthorLastName_Series1_Title
  [**]Genre_AuthorLastName_Series2_Title
  [**]where 1 and 2 are the positions in the series.
• w/out series: Genre_AuthorLastName_Title

Any ideas?

[user_none]

This will work provided there are no numbers in the series name:

[^_]+_(?P<author>[^_]+)_((?P<series>[^_\d]+)(?P<series_index>\d+)?_)?(?P<title>[^_]+)

[Justy]

Originally Posted by user_none

This will work provided there are no numbers in the series name:

[^_]+_(?P<author>[^_]+)_((?P<series>[^_\d]+)(?P<series_index>\d+)?_)?(?P<title>[^_]+)

You are amazing, thanks! That worked wonderfully!

I don't see a test field for Genre or Tag in the regex creation screen. Does this mean that I can't use a <tag> or <tags> at the begining of the above expression to retain my Genre tags, or would them come in anyway if I had the correct field identifier?

[user_none]

Tag is not supported. The only identifiers that can be pulled out of the file name are what are on that screen (Title, Authors, Series, Series Index, and ISBN).

[Justy]

Originally Posted by user_none

Tag is not supported. The only identifiers that can be pulled out of the file name are what are on that screen (Title, Authors, Series, Series Index, and ISBN).

Thanks again! You have been very very helpful!

[Justy]

Originally Posted by user_none

This will work provided there are no numbers in the series name:

[^_]+_(?P<author>[^_]+)_((?P<series>[^_\d]+)(?P<series_index>\d+)?_)?(?P<title>[^_]+)

This regex worked great last year but now I've had to completely reload my netbook. I am now running Calibrie 0.7.12. When I plugged the above regex into the 'Add files' screen and tested it on one of my file names it put the entire file name into the Title field instead of splitting out the components. Has something changed in how the filenames are handled?

[kovidgoyal]

Make sure you put the file extension as well as the name.

[Justy]

I entered this as the Regular Expression:

Originally Posted by user_none

[^_]+_(?P<author>[^_]+)_((?P<series>[^_\d]+)(?P<series_index>\d+)?_)?(?P<title>[^_]+)

My test file names were:

1. Genre_AuthorLastName_Title.prc
2. Genre_AuthorLastName_Series01_Title.prc

which gave the following results respectively:

1. Title = Genre AuthorLastName Title
2. Title = Genre AuthorLastName Series01 Title

It's been a while since I played with Calibre but since I'm not working right now I thought I'd finally do that first mass import of all my eBooks. Unfortunately for me, the coding I learned in school for C+ didn't look anything like this.

[kovidgoyal]

Have you checked the option to read metadata only from filenames.

[Starson17]

Originally Posted by Justy

My test file names were:

1. Genre_AuthorLastName_Title.prc
2. Genre_AuthorLastName_Series01_Title.prc

Try this:

([^_]+ )(?P<author>[A-Za-z]+) ((?P<series>[^_\d]+)(?P<series_index>\d+)? )?(?P<title>[^_]+)

[mumdigau]

Hi,

I've a very similar situation here, same as above, but without genre. The filenames are 'Author FirstName LastName_SeriesnameSeriesindex_Title'.

I slightly modified your regex to

(?P<author>[A-Za-z ]+) ((?P<series>[^_\d]+)(?P<series_index>\d+)? )?(?P<title>[A-Za-z ][^_]+)

which more or less works besides two pitfalls:

1. The seriesindex isn't correct, e.g. 0051 is shown as 51.0
2. Spaces in seriesname aren't handled correctly, e.g. seriesname 'A B' is shown as 'B', and author as 'author A'.

Can you help correct this?

Best regards

mumdigau