[GUI Plugin] Library Codes

[DaltonST]

A prefix can be anything of any length.

An example of what can be done with the new Tag Prefix Suffix Rules Table in the Scrub Tags tool of Job Spy (that I intend to supersede entirely the 'prefix' values in Library Code's options once the new JS functionality is released):

Example: Add a prefix of "Fiction" to any tag that does not already start with "Fiction" but has "Romance" or "Suspense" or "Science" or "Mystery" in it, or otherwise matches the rule's regular expression matching requirement.

I use tags with implicit hierarchical formatting like "Factual:History-War-WWII-Africa-North Africa-Libya" and "Fiction:Science-Space Opera" and "Fiction:Romance-Paranormal-Vampires".

I develop my plug-ins for myself as their first user, and I would never want any Tag that had "--"'s in the middle arbitrarily split into 2 or 3 pieces as has been suggested. Such splitting should be rule-based on an individual user basis. I might add that capability to Job Spy too.



DaltonST

[davidfor]

Originally Posted by DaltonST

A prefix can be anything of any length.

An example of what can be done with the new Tag Prefix Suffix Rules Table in the Scrub Tags tool of Job Spy (that I intend to supersede entirely the 'prefix' values in Library Code's options once the new JS functionality is released):

Example: Add a prefix of "Fiction" to any tag that does not already start with "Fiction" but has "Romance" or "Suspense" or "Science" or "Mystery" in it, or otherwise matches the rule's regular expression matching requirement.

I use tags with implicit hierarchical formatting like "Factual:History-War-WWII-Africa-North Africa-Libya" and "Fiction:Science-Space Opera" and "Fiction:Romance-Paranormal-Vampires".

OK, so it is adding something that one of your other plugins might find useful.

I develop my plug-ins for myself as their first user, and I would never want any Tag that had "--"'s in the middle arbitrarily split into 2 or 3 pieces as has been suggested. Such splitting should be rule-based on an individual user basis. I might add that capability to Job Spy too.

Yes, and as you so capably demonstrated, adding an extra option to the configuration dialog to enable a split like this would be simple. The amount of time you took to add that prefix would be roughly the same time as to add an option to do what @knobunc wanted. And of course, where to do the split,or even if it makes sense at all, is dependent on the tag type. The FAST tags are clearly using "--" in a way that represents a hierarchy. The examples you use above appear to use just "-" (before you added the prefix). I don't know about any of the other tags that Library Code as I haven't looked at them closely.

Adding splitting like this to JobSpy is fine, but it does mean the user has to install and configure that plugin. And they have to run two separate steps each time they update metadata.

But, it is good that you finally understand what was being asked.

[DaltonST]

Library Codes users who wish to modify either the LCEAD Tags or the FAST Tags should see this post: https://www.mobileread.com/forums/sh...5&postcount=23


For example, if you want to replace any "--" with a "." for #fast tags, use the Tag String Replacement Rules table to do so.




DaltonST

[knobunc]

Thanks! I'll give that a try.

Do you have any thoughts about making it possible to skip results from WorldCat that are not Dewey codes? e.g. "FIC" or "B" often come back. Unfortunately, they are not consistently used so I usually get back a DDC, but about 30% of the books have something else.

I added a (probably horrible) bit of code to classify_web_service_api.py to skip things that don't match a regex. And that works nicely for now while I'm updating my collection to include the FAST codes.

But I wasn't sure if it was something you were interested in adding to your plugin.

               #~ ---------------------------
                ddc = recommendations.getElementsByTagName('ddc')[0]
                #~ ---------------------------
                if ddc:
                    for mostPopular in ddc.getElementsByTagName('mostPopular'):
                        holdings = mostPopular.attributes["holdings"].value
                        nsfa = mostPopular.attributes["nsfa"].value
                        sfa = mostPopular.attributes["sfa"].value
                        if DEBUG: print('DDC mostPopular: class=' + sfa + ' normalized=' + nsfa + ' holdings=' + holdings)
                        # Allow cutter codes after matches, but we want numbers, not text strings (e.g. FIC, B, etc.)
                        if re.match(r"^\d+(\.$\d+)?(\s.+)?", sfa) is not None:
                            ddc_return = sfa
                            break
                    del ddc
                    del holdings

[DaltonST]

Use the Job Spy plugin's Scrub Tags functionality to add new tags, delete undesired tags, change existing tags, create a 3d new tag if 2 certain tags exist, or a 4th if 3 exist, capitalize sub-strings within tags, and so forth.

The above can easily delete any tags for any specific custom column, like #fast, based on any matching regular expression you wish to specify.




DaltonST

[knobunc]

Sorry, Job Spy is really cool, but I'm not sure how it pertains in this particular case.

The little patch above is to make it so that the OCLC Dewey Codes are actually valid codes. Many of the entries seem to have FIC or B as the DDC, this just skips them.

[DaltonST]

Originally Posted by knobunc

Sorry, Job Spy is really cool, but I'm not sure how it pertains in this particular case.

The little patch above is to make it so that the OCLC Dewey Codes are actually valid codes. Many of the entries seem to have FIC or B as the DDC, this just skips them.

I was in a "tag-like" custom column mindset when I mentioned Job Spy. DDC of course is just a simple text field, not a tag-like field. I should have said "Use Calibre Bulk Edit Metadata to delete FIC and B if you do not want them". Job Spy also has a special tool for wiping out comments-like text, such as DDC codes, for selected books.

FIC and B are non-standard DDC codes, but commonly downloaded. I have a lot of them. Personally, I prefer a FIC or a B (for Fiction or Biography) rather no DDC at all.

LC is generic. What one user may like, another use may dislike. Fortunately, Calibre has a lot of functionality to make both happy.



DaltonST

[Dak21]

Sorry if I am being a bother I am receiving this error:

calibre, version 3.35.0
ERROR: Cannot Update DDC, LCC for Selected Books: An OCLC-OWI Identifier for at least one of the selected books, plus DDC and/or LCC Custom Columns, must exist in this library order to perform this action.

I don't understand why I would since I have the column present and added an OCLC number

Thank you in Advance, I apologize if this was addressed somewhere else I couldn't find it.

[DaltonST]

The OCLC-OWI is added by LC as an Identifier. Edit your book that you say has an OCLC (which is not the same as OCLC-OWI). Does it show under Identifiers that there is an OCLC-OWI? No? Assuming that you have an ISBN and that route did not work, try the other 2-part route (part 1 of 2, and then part 2 of 2). You can read the ToolTips for the LC main menu by hovering your mouse over it. They explain the above. Basically, the more Identifiers LC finds, the more additional Identifiers it can find.

Not all books will have anything found by LC, especially new fiction, even if the books have an ISBN. It has to be available on the web.

Good luck.


DaltonST

[Ascobis]

Greetings,
Please forgive if this is answered in the ten preceding pages.
As a programmer/developer, I am merely curious: Why does LC run in the foreground instead of as a background job?

I have learned to start page counts first and then to start the LC lookups; no worries.

Thank you for your most excellent Calibre plugin.

[DaltonST]

Originally, LC was a private plug-in that I developed just for me, and running it as multiple threads in the foreground rather than jobs in the background fit how I needed to use it for dodgy fiction books. I made it available to everyone later as an afterthought.

EPUBs always have a page count, so the Count Pages plugin is perfect running as a job. Ditto for Extract ISBN, Polish Books, Modify EPUB, etc. That is not at all the case for DDC or LCC for random fiction books with sketchy metadata that often do not have a straight-forward lookup. It takes manual fiddling with to find them, and even then fails frequently for new fiction unless a manually-discovered alternate ISBN can be found for the lookup.

I have automated the quality fixes and metadata enhancements (e.g. count pages, polishing, extract ISBN, etc.), so all of that is already done by the time I get around to running LC in my Work Bench Library. I am referring to the Job Spy plug-in. See the attached.



DaltonST

[GalacticHull]

Is anyone else experiencing the functional death of this plugin, not due to the plugin itself, it would seem, but, run a search: http://classify.oclc.org/classify2/

Or click on the cover of a book under Jump Right In...

I suppose I would like to know I'm not the only person experiencing this....

[DaltonST]

Originally Posted by GalacticHull

Is anyone else experiencing the functional death of this plugin, not due to the plugin itself, it would seem, but, run a search: http://classify.oclc.org/classify2/

Or click on the cover of a book under Jump Right In...

I suppose I would like to know I'm not the only person experiencing this....

This plugin works 100% perfectly. It does exactly what it says it does. The "Jump Right In" you refer to has nothing to do with this plugin. Your report of the death of this plugin is premature.




DaltonST

[GalacticHull]

Originally Posted by DaltonST

This plugin works 100% perfectly. It does exactly what it says it does. The "Jump Right In" you refer to has nothing to do with this plugin. Your report of the death of this plugin is premature.

Yeah. What I wrote wasn't close to perfect. The plugin you've developed is fantastic, by the way. I simply wanted someone besides myself--distinct computer, ip, etc--to see if they were having the same trouble with http://classify.oclc.org/classify2 as I.

For about 36 hours, until a few minutes ago, if I were to enter a title, isbn, etc. or click on the cover of a 'jump right in' book, it would simply reload http://classify.oclc.org/classify2 -- didn't intend to imply that this had anything to do with your plugin, just that this process was subject to the same symptom.

That said, after 36 hours or so, the site is back to functioning, and the plugin with it.

[DaltonST]

Version 1.0.39 - 2019-04-13 Removed LC's personal BeautifulSoup 3. Use standard Calibre's BeautifulSoup (currently 3, but moving to 4), and choose BS3 .findAll or BS4 .find_all dynamically based on its attributes. Both backwards and forwards compatible with Calibre.

DaltonST