[GUI Plugin] Library Codes

[DaltonST]

Originally Posted by isarl

If my comment came across as entitled, or as anything other than an attempt to provide helpful information to lessen your own work in creating a plugin whose utility I have already effusively praised in my comments in this thread, then I apologize. I have no interest in demanding anything of you; as I've said before, I greatly appreciate all the work you've already put into this plugin, whatever becomes of its future.

On that note, I hope the following bug reports are helpful. I have noticed that, while FAST tags with a comma in them (e.g. “Short stories, American”) used to have the comma replaced with a semicolon, this is not happening anymore, and as a result, instead of one tag like “Short stories; American”, the tag is split on the comma. I can't recall if this was something I had configured for myself or if it was something Library Codes was doing for me, because I don't see any relevant configuration options in the configuration dialogue accessed from the context menu -> “Configuration & ToolTip-Instructions” and if I try to configure the plugin from Preferences -> Plugins -> Library Codes -> Customize plugin, then I get the second issue I've noticed, which is the following error:

calibre, version 6.15.1
ERROR: Unhandled exception: <b>TypeError</b>:setTextFormat(self, a0: Qt.TextFormat): argument 1 has unexpected type 'int'

calibre 6.15.1  embedded-python: False
Linux-6.1.23-1-lts-x86_64-with-glibc2.37 Linux ('64bit', 'ELF')
('Linux', '6.1.23-1-lts', '#1 SMP PREEMPT_DYNAMIC Thu, 06 Apr 2023 10:56:12 +0000')
Python 3.10.10
Interface language: None
Successfully initialized third party plugins: [...snip...] && Library Codes (1, 0, 61) && [...snip...]
Traceback (most recent call last):
  File "/usr/lib/calibre/calibre/gui2/preferences/plugins.py", line 301, in customize_plugin
    self.modify_plugin(op='customize')
  File "/usr/lib/calibre/calibre/gui2/preferences/plugins.py", line 378, in modify_plugin
    if plugin.do_user_config(self.gui):
  File "/usr/lib/calibre/calibre/customize/__init__.py", line 156, in do_user_config
    config_widget = self.config_widget()
  File "calibre_plugins.library_codes.__init__", line 30, in config_widget
    return ConfigWidget()
  File "calibre_plugins.library_codes.config", line 62, in __init__
    self.label1.setTextFormat(1)
TypeError: setTextFormat(self, a0: Qt.TextFormat): argument 1 has unexpected type 'int'

I hope this bug report is helpful; if you need any more information then I'm happy to provide what I can.

On it. Thanks.


DaltonST

[DaltonST]

Version 1.0.62 - 2023-04-13 Miscellany

The old LC replaced a comma within a single FAST Tag with a semi-colon. When I rewrote LC to use webscraping instead of the old API that is deprecated now, that got dropped. This release adds it back.

The hard-to-find config window that says "do not use this config window, rather use the LC Menu configuration menu item instead" had a very old bug. Fixed.


DaltonST

[isarl]

Originally Posted by DaltonST

Version 1.0.62 - 2023-04-13 Miscellany

The old LC replaced a comma within a single FAST Tag with a semi-colon. When I rewrote LC to use webscraping instead of the old API that is deprecated now, that got dropped. This release adds it back.

The hard-to-find config window that says "do not use this config window, rather use the LC Menu configuration menu item instead" had a very old bug. Fixed.


DaltonST

Thank you! I suspected that I remembered some kind of “do not use this config window,” dialogue as you mention here, but I didn't want to make up stories, so in my bug report I limited myself to reproducible observations. I can confirm that I am again seeing, “Please Customize Directly Within Library Codes” when trying to customize via the plugins menu.

The fix for colon -> semicolon replacement does not appear to be working yet. Example OCLC work ID: 53061538 (link to Classify results for this title). “Short stories, English” gets stored as two tags, “Short stories” and “English”, rather than as “Short stories; English”.

[isarl]

With the 1.0.63 update, FAST tags now appear to have their commas transformed into semicolons again. Thank you very much!

[DaltonST]

Version 1.0.64 - 2023-04-15 Automatically get oclc-wi if needed to obviate need to get oclc-wi via Author/Title to re-run to download DDC et al. This need automatically turns on saving all LC Identifiers, regardless of customization.

LC now gets oclc-wi if it needs it and is available immediately after the failure of ISBN/ISSN to retrieve DDC. LC now automatically does a re-do of the download of DDC et al using the oclc-wi that it just got during the aforementioned failure of ISBN/ISSN.

If you have ISBN or ISSN configured as a Composite Custom Column so you can see it just like your would DDC etc., you must also add an oclc-wi Composite Custom Column.

The oclc-wi Identifier is a peer to ISBN/ISSN as far as LC is concerned. In fact, oclc-wi is better than ISBN for downloading DDC etc. ISBN Identifiers are often unresponsive; oclc-wi Identifiers never are.

If ISBN/ISSN does not work, LC must use oclc-wi. You need to know if your book has it or not via the Composite Custom Column.

If your book does not already have oclc-wi, and LC cannot get it automatically using ISBN/ISSN, you must get oclc-wi using Author/Title.

If using Author/Title does not yield an oclc-wi that is successful getting DDC et al, you will have to resort to searching the web for alternative ISBNs for the book, and trying each of them.


DaltonST


n.b. The Classify system that LC looks at to get DDC etc. states that its db is current through June 2022. Bear that in mind. LC cannot get what is not there.

[DaltonST]

Version 1.0.65 - 2023-04-16 Change in multiple-response re-do trigger. Bug fixed.


DaltonST

[SidJnsn]

xxxxx

[DaltonST]

Sid:

Your uploaded library_codes.zip is LC version "1.0.64", NOT "1.0.65", which has been the latest since before your created your .zip. That means that any testing you did using the old version is irrelevant.

Your uploaded .zip is obsolete, so no one should be using it at all. Anyone downloading and installing it should reinstall "1.0.65" using the standard mechanism to do so.


DaltonST

[SidJnsn]

xxxxx

[DaltonST]

Originally Posted by SidJnsn

Do you plan to address the bug below that I PM'd you on 5/13?

"Some wi's, if sent in the stdnbr field, will return two books, one with the wi = the number, and the second with an owi that matches the number. To avoid this, a separate call is required for wi's apart from the stdnbr call, as mentioned in an earlier post when we had the api.

The two calls to the demo site are:
base_stdnr = "http://classify.oclc.org/classify2/ClassifyDemo?search-standnum-txt=[STDNBRGOESHERE]&startRec=0"

base_winbr = "http://classify.oclc.org/classify2/ClassifyDemo?wi=[WINBRGOESHERE]" "

No, because as I last posted, you tested using, and wrote your messages based on, LC Version 1.0.94 instead of 1.0.95 (released on April 16th). Your test results are of no value.





DaltonST

[SidJnsn]

xxxxx

[DaltonST]

Originally Posted by SidJnsn

Even a cursory glance would have shown you that the problem I outlined was not addressed by the 65 release. My intent in sending you the code was only to show that the second call solves the problem.

If your intent is to alienate every grateful user who is trying to help, you're doing a great job.

Maybe you will find this more "relevant" and "of value":

wi = 39578615; title = "Edward Said and the Post-Colonial" Author = Bill Ashcroft & Hussein Kadhim

The 65 ver will return nothing for this wi. The demo cite returns two books, one which is the wi of that number, and the other that is the owi. If you click on the first entry on the demo site, it will make the call that I proposed with wi = 39578615 vs. stdnbr=39578615.

"Helping" yourself, you mean.

I now have lost all interest in LC. Unless a new version of Calibre breaks something in LC, I won't be wasting any time on it.


DaltonST

[isarl]

Originally Posted by SidJnsn

Even a cursory glance would have shown you that the problem I outlined was not addressed by the 65 release. My intent in sending you the code was only to show that the second call solves the problem.

If your intent is to alienate every grateful user who is trying to help, you're doing a great job.

Maybe you will find this more "relevant" and "of value":

wi = 39578615; title = "Edward Said and the Post-Colonial" Author = Bill Ashcroft & Hussein Kadhim

The 65 ver will return nothing for this wi. The demo cite returns two books, one which is the wi of that number, and the other that is the owi. If you click on the first entry on the demo site, it will make the call that I proposed with wi = 39578615 vs. stdnbr=39578615.

This is a strange degree of entitlement and hostility to bring to somebody who made you something for free. It is a very reasonable expectation for DaltonST to ask that bug reports be made from a current version of the software and instead of simply doing so, you chose to be argumentative. If you are as competent as your bug report makes you seem, why not patch the bug yourself, and contribute your patch, instead of talking down and making demands?

If I were supporting free software, and a user responded to such a request with hostility and entitlement, I would be demotivated to continue supporting that software, too.

DaltonST: whatever your future plans are, thanks again for all the work you've already put into this tool.

[SidJnsn]

isarl,

You do not have the full context, so are mischaracterizing my post. Specifically, you have not seen the PM's where I have provided fixes and offered to help in any way possible. Had I been asked to provide the changes to the latest version, I would have gladly done so. Instead, I got a couple of snarky public messages, so I replied in kind. My question had to do with whether he planned to make the fix, so I could create my own version if not.

[SidJnsn]

The attached plugin will find classification info for books that have a "wi" after running the 65 version, but no ddc. After extensive testing, I believe it does everything that the Classify demo site does, except for displaying FIC or E as the DDC. The 65 version does not do that either.
- can be installed manually without uninstalling the 65 version.
- will display as Library Codes MODIFIED 1.0.65
- will not interfere with notifications of future version upgrades, although those upgrades will wipe out my changes, unless DaltonST chooses to incorporate them.
- will not interfere with installing future upgrades through the normal process.
- my changes affect only the ini, classify, and ui files and are clearly marked with "####"