How to quickly create tag based on word in the comments field

[sophieturgeon]

Hello,

OK I need help again! As I have mentioned in the past I have an extra=large library that I am trying to organize. I have been able to download the metadata for almost all of my books. Now I am trying to organize the tags. I don’t want to use any of the tags given by the publishers. They have too many for the same thing and they can be vague. I am currently doing this manually which is mad because I have thousands of books. The way things are going, I’ll be a very old woman by the time I’m finished.

I am updating the tags based on the reading in the comment field in the metadata. I look for specific words; for example if the words “murder” “killer” “killed” “murderer” or a question mark appear in the comments field, I will tag the book “suspense/mystery”.

I am no programmer, never was. But I have created enough simple macros in excel to know something like that is possible. I remember creating some kind of look up tables with “if” formulas. That was over 7 years ago, so I don’t remember much from those days.

That I know of, no plugins like that has been written yet. Can you tell me where I might find the very simple written instructions, or a video or something? I have a lot of difficulty understanding the info and terms used in the calibre help site: For someone like me, it is like reading a foreign language. I look forward to any suggestions you might have. Thanks again!

[DoctorOhh]

Moderator Notice
Moved post to the proper calibre forum. More folks can help answer this question here.

[theducks]

Originally Posted by sophieturgeon

Hello,

OK I need help again! As I have mentioned in the past I have an extra=large library that I am trying to organize. I have been able to download the metadata for almost all of my books. Now I am trying to organize the tags. I don’t want to use any of the tags given by the publishers. They have too many for the same thing and they can be vague. I am currently doing this manually which is mad because I have thousands of books. The way things are going, I’ll be a very old woman by the time I’m finished.

I am updating the tags based on the reading in the comment field in the metadata. I look for specific words; for example if the words “murder” “killer” “killed” “murderer” or a question mark appear in the comments field, I will tag the book “suspense/mystery”.

I am no programmer, never was. But I have created enough simple macros in excel to know something like that is possible. I remember creating some kind of look up tables with “if” formulas. That was over 7 years ago, so I don’t remember much from those days.

That I know of, no plugins like that has been written yet. Can you tell me where I might find the very simple written instructions, or a video or something? I have a lot of difficulty understanding the info and terms used in the calibre help site: For someone like me, it is like reading a foreign language. I look forward to any suggestions you might have. Thanks again!

It will still be tedious to Tag/retag.

in Calibre
search comments:<your magic words for your books to get A tag
Select All <results>:tap 'E' (bulk meta edit): Add Tag <your tag>

another idea for the existing tag mess
In the Tag Browser: Tags section: Right click: (here you can rename (to consolidate) or simply delete those garbage tags eg. Science Fiction, Fiction
<rant> <can SF or Fantasy be non-fiction?? Many tags are a holdover from those 3x5 card catalog days </rant>
Another thing to think about is hierarchical tags
Science Fiction.Military vs Science Fiction-Military

The former can easily return results for both cases: SF or SF.Military

[BetterRed]

@sophieturgeon - maybe this PI can help ==>> [GUI Plugin] QuarantineAndScrub

I don't use it, but it seems to be targeted at cleaning up metadata.

BR

[DaltonST]

@sophieturgeon:

QuarantineAndScrub is an add-on (not just a plug-in) that has (among other things) enormous Work Tag functionality. Please read the 45 page User Guide for more details. It can be found at: https://www.mobileread.com/forums/sho...d.php?t=247425 .

The one thing it does not do (yet) is to allow scanning of Comments to derive Work Tags. The reason it does not do it is that it did not occur to me until now. So, I am going to add that functionality too.

In the meantime, you can default Work Tag by Author. It will also derive Work Tags based on keywords in the Work Title (e.g. Mystery of the Dead Man" generates a a Work Tag of Mystery).

Work Tags are passed thru the Tag Rules Table to either delete or change them based on your rules. For example, a tag of "Mystery" in my QuarantineAndScrub ecosystem is changed into "Fiction:Mystery". You own the Tag Rules Table.

Multiple Tags can be winnowed down via the Tag Priorities table, which you also own, which is used by the Q&S "Minimize Work Tag" function.

And....well, there is way too much functionality to describe here, but suffice it to say that QuarantineAndScrub was designed for people with large (e.g. 30,000 ebook) libraries that are a mess. The User Guide is 45 pages in length for a reason.

If you choose to install Q&S, you will also have to download the official Q&S library (with metadata.db and .json) from the Q&S original post, and install it as described in the OP plus the User Guide. Q&S uses 42 (currently) custom tables and 32 (currently) views that Calibre knows nothing about. By no means is Q&S "just a plug-in".

The User Guide is your friend and source of support for Q&S.


Good luck.



DaltonST

[DaltonST]

QuarantineAndScrub Version 3.4.2 now supports creating Work Tags by parsing the Comments of a book using a new table, _tags_by_comment . This new table is delivered empty because it is so powerful. You must "opt in" by adding rows to it. Refer to the attached image for examples.


DaltonST

[sophieturgeon]

Thank you DaltonST. I may not have replied right away, but I've been thinking about your tool. I just need to figure it out (reading the PPT file and learning by doing). I will let you know how it goes. Thanks again

[Gary_M_Mugford]

DaltonST,

I've had a situation that looks a lot like yours. I have my OWN tag field which is perpetually only one word. Sometimes I have to invent words, like MThriller for Medical Thriller or BSports which stands for a Sports Biography. But in general, I keep to the categories I would mentally assign the book.

On inhaling a bunch of books, I select them all and then press E to edit and change the TagsGM category to UNCAT (Uncatagorized). At the end of the month, when I'm sending out a list of what's new to the various family members, I run my series of searches that would lead me to reducing the books to the one-word classification system that appeals to me.

First, I run my search for malformed books and author information. Here's my ___KILL_EM_ALL search string:

title:"(" or title:"[" or title:"{" or title:"-" or title:":" or title:"#" or title:"Part " or title:"Series" or title:"\.doc" or title:"\.doc" or title:"\.indd" or title:"Microsoft" or title:"\:\ Book" or authors:"The\ " or title:"\-\-" or authors:"Editor" or authors:"=Owner" or title:"ebok" or title:"publish" or author:"publish" or author:"book" or author:"design" or author:"hot tree editing" or author:"warcraft"

I manually edit the obvious ones, wiping out MOST title clauses that start with any bracket and then correcting what I can from looking at the meta-data (or pre-existing cover)

After that, I run through each of the actual category search in the order that I find optimal [which changes more than it should] and then select all (of the found books) and then press e to edit. I delete all previous contents of TagsGM and then add back in the one that the search was designed to ferret out. Press save and all the books disappear since they are no longer UNCAT. And example of a search to fulfill WRITING as a tagsGM would be:

#tagsgm:"=Uncat" and (tags:"Writing" or tags:"Words" or tags:"Language" or tags:"Grammar" or tags:"Linguistics")

THRILLER would be:

#tagsgm:"=Uncat" and (tags:"Thriller" or tags:"Suspense" or tags:"=Romantic Suspense" or tags:"=Action" or tags:"Adventure")

In developing these searches, I actually created them all in a notepad file, as the process for typing there was easier than into the editor screen in calibre. It also serves as documentation.

My only issue, is that it takes about a half-hour to KillEmAll and then do each individual search and edit in the proper tagsGM classification. Really not that long, but an automated process would be better. Afterall, It IS a repetetive chore with no editorial guidance involved beyond spotting certain incongruities, such as when a sports biography is going to be misclassified.

I'm going to take a long look at your work. It READS like it is a variant of my approach and can be melded on my approach.

Regardless, thanks for taking the time to do what you've done and to share it with us. It's much appreciated.

GM