Regex help on reading Metadata from file name.

[JohnnyBook]

90% of my files are in the format: (The other 10% do not include the pub date)

Format: Series-series number title (author) Pub date.txt
(Series 2-4 letters)
(Series number 2-4 digits)
(pub date-year only)

ex.

BA-123 How it works (John Smith) 1989.txt
ROT-4089 Make it this way (Jane Smith) 2009.txt

Playing around with the regex I was able to separate the series and number but I could not work out the title and author Typicaly I ended up with

Title: How it works (John Smith
Author: )

And pubdate does not work at all.

Unfortunetly I kept changing it around and now it does not work at all and I cant remember what I had that almost worked.

One thing I have had trouble with is the "(" and ")" and trying to search for them in the title. I CAN search and replace the titles to remove them to substitute them for another character to make it easier to run a regex if necessary (just not "-" as some titles have a "-" in them.

Anyone have any clue how to do this?

edit: This is as close as I can come to what I had

(?P<series>[^_0-9-]*)-(?P<series_index>[0-9]*)(?P<title>[^_-]+) \(?(?P<author>[^_].+) -?(?P<date>[^_].+) ?

[eschwartz]

(?P<series>[A-Za-z ]+)-(?P<series_index>[0-9]+) (?P<title>.+) \((?P<author>[A-Za-z. ]+)\) (?P<published>[0-9]+)

Pubdate seems to be making up its own month/day

[JohnnyBook]

Great that seems to have done it. (At least in the test box) I will run some books through and see if importing works... But I am sure it will.

Thanks,

[eschwartz]

You're welcome.

[JohnnyBook]

Ok, in the test box it works perfectly. (and as you noted the strange bug, all the "published" years have Mar-15 appended for month and day)

and once files are processed, in Calibre the data is correct except....

The titles are the original file name, not the book tittle from the add books regex.

ex,
ROT-4089 Make it this way (Jane Smith) 2009.txt
in the regex box is:

Series: ROT [4089]
Author: Jane Smith
Title: Make it this way
Published: 2009-03-15

But once imported in Calibre
Title:ROT-4089 Make it this way (Jane Smith) 2009

Why is it ignoring the regex for the tittle?

EDIT: Never mind. I forgot to uncheck the box for read from metadata instead of file name.

[eschwartz]

Glad you figured it out!

Also, regarding your PM re: covers -- no, there is no good solution for getting covers for a TXT file.
You can import using but you would need a metadata.opf also.

Alternatively, do a metadata download (shortcut key is CTRL+D) which redownloads covers from various sources... which also takes time, although you can set it running automatically.

[JohnnyBook]

Yes, that is how I import them, I may have to use the method posted in that thread I linked in my PM to generate covers and opfs, then do a cover substitution and re-import.

Thanks for your assistance.

[JohnnyBook]

Originally Posted by eschwartz

(?P<series>[A-Za-z ]+)-(?P<series_index>[0-9]+) (?P<title>.+) \((?P<author>[A-Za-z. ]+)\) (?P<published>[0-9]+)

Pubdate seems to be making up its own month/day

Ok. it works in Calibre, except, if the book does not have a published date... then it fails and no parsing is done at all and the title is the file name and the rest are blank.

Is there a way to make the published date part optional?

[eschwartz]

Yes, append that group with a question mark.

(?P<series>[A-Za-z ]+)-(?P<series_index>[0-9]+) (?P<title>.+) \((?P<author>[A-Za-z. ]+)\) (?P<published>[0-9]+)?

If I had known the published date was supposed to be optional, I would've done that in the first place. You get what you asked for.
If I didn't have short-term memory loss and forget part of the OP...

[JohnnyBook]

Originally Posted by eschwartz

Yes, append that group with a question mark.

(?P<series>[A-Za-z ]+)-(?P<series_index>[0-9]+) (?P<title>.+) \((?P<author>[A-Za-z. ]+)\) (?P<published>[0-9]+)?

If I had known the published date was supposed to be optional, I would've done that in the first place. You get what you asked for.

First line in my OP.

"90% of my files are in the format: (The other 10% do not include the pub date)"

In any case, Thanks bunches for all your help.

EDIT: Nope still did not do it... Maybe since it does not find the space after the Author, it still fails?

ROT-4089 Make it this way (Jane Smith) 2009.txt
ROT-4089 Make it this way (Jane Smith).txt

The first works, the second is not parsed.

[eschwartz]

Heh, overlooked that, sorry. That is indeed my fault then.


Dur -- because it was still trying to find a space at the end. This time I did the right thing and tested it.

(?P<series>[A-Za-z ]+)-(?P<series_index>[0-9]+) (?P<title>.+) \((?P<author>[A-Za-z. ]+)\) ?(?P<published>[0-9]+)?

[JohnnyBook]

That did it. it works great now.

And it looks like that other thread, to generate covers and opfs, then do a cover substitution and re-import is actually pretty fast and easy to do.

https://www.mobileread.com/forums/sho...d+cover&page=2