Adding custom fields to ComicInfo.xml?

[chowbok]

Have you given any thought to adding custom tags to ComicInfo.xml? Would this break things for other apps? anansi-project/comicinfo seems to be dead, or I would ask there as well.

[maforget]

On a technical side, unknown data in XML should normally be ignored by other programs. But we can't be certain it would be the case. Even if I've modified it in the past, I do try to keep compatibility in mind.

Also everyone seems to want something different they would want to add. I've heard of more manga centric fields, multiples values for some fields, saving user data like read progress, etc. I don't want to change a format that was used for more than a decade by various programs. There is also all the related effect some of these changes could have on the program.

And to be frank the current format suits me perfectly, I see no big need to mess with it. There are also plenty of workarounds that already exist. Plugins like StorePlus or CompleteMetadata that let's you store and restore custom values, leaving you with pretty much unlimited custom fields.

I would be more keen to add support for other metadata format, that wouldn't mess with the ComicInfo.xml format. With the change for adding support for MetronInfo, it is pretty easy to add support for new formats that might arise. Might even be possible to create a new plugin type that could handle custom metadata files.

[chowbok]

Sure, I understand. Actually, CompleteMetadata would suit my purposes perfectly if only I could get it to work.

[maforget]

That script requires Python 3.10 to be installed separately on your computer. That's probably the issue you have.

The StorePlus is just a simple python file you drop in your script folder, so It's simpler since it only saves the extra data in the Notes field. CompleteMetadata actually updates your file on it's own and clears your dirty flag. But adds the new field directly in the XML, so compatibility issue with 3rd party program might be a thing (if any exists).

[chowbok]

Duh, it actually says to install it and I just glossed over that. Thanks for pointing that out. It works fine, I don't need a change made to ComicRackCE then.

[maforget]

I've made some work in adding a new metadata file (ComicBook.xml) that contains all (most) of the library metadata. I have a test build you can test if you want. See #261

Do make sure to do a backup of the database and don't convert all your files until you are certain everything is fine. I would start testing it with test files at first.

[SteveHutchinsASBFA]

Hello,

I have a TON of custom value field metadata, so I, too, would love to see this functionality. My understanding is the plug-ins are old and don't work well (though I haven't tried any yet so I won't swear to that). Compatibility with other programs is an issue. So, is dumping the custom value fields to a separate "ComicInfoCustom.xml" file an option? That would make it exclusive to CRCE. Or would that be too much effort?

--Steve

[maforget]

My understanding is the plug-ins are old and don't work well

The plugins are old but doesn't mean that they don't work well. It's more the tooling around IronPython that is non existent, so it's harder to create, so the issue isn't really with the user. A simple notepad and print() is enough to create something, it's still just a python script. If any errors in plugins exists it's usually due to bugs in said script and just requires the maintainer to update them. Sadly most of these scripts are old and the original creator have lost interests, so somebody needs to just update them.

• CompleteMetadata I mentioned for example simply bypasses IronPython by just calling a regular python 3.10 script and do all the work in modern python. Or my AmazonScrapper & FindImageResolution plugins that just use pure C# and also bypass IronPython.
• The StorePlus is a single python file you drop in your scripts dir and let's you save all the extra data directly into the Notes field.

The issue with these plugins is more that it requires an action and it isn't automatic.

---

So, is dumping the custom value fields to a separate "ComicInfoCustom.xml" file an option?

While we are adding a new file, why simply restrict it to the custom values? Why not add other stuff?

In the code there is the ComicBook object that is used through out the program. That class in fact inherits from the ComicInfo class, which is what creates the current XML file and contains only the current properties. All the other properties are part of the ComicBook class instead. So the most way forward way to fix this is simply to add a ComicBook.xml file. This already exists when syncing via Android. The XML file that is created is in fact the full ComicBook and not simply the ComicInfo.

That file would include everything that is saved in the database covering pretty much everyone use case. Just need to add new options in the UI to also include the ComicBook.xml, disable all non library restrictions in the UI when this option is enabled & update read/write/update to also add the new file.

The potential issue I see is what happens when a ComicBook.xml differs from the ComicInfo.xml. Normally all the same data in the ComicInfo.xml would be included in the ComicBook.xml, but what if a 3rd party program updated the ComicInfo.xml independently? Normally the ComicBook.xml would be preferred over the ComicInfo.xml, since it's more complete. But in this case which to choose? Load the ComicBook.xml and replace everything with what the ComicInfo.xml contains after it? It would make it difficult to track discrepancies between both files.

[SteveHutchinsASBFA]

Thank you for considering this. I think the ComicVine plug-in has an option for updating existing metadata if it differs or leaving it in place. That may be a workable solution. It wouldn't be specific (it's all or nothing), but it would be an option and could be used on a book-by-book or groups of books or simply bulk as needed.

[maforget]

That option has no relevance to whether or not the auto update is done. It simply says if at any moment you do an update also update the ComicBook.xml (and enable some UI elements in the Info dialog).

Any change to these Library Info field will not trigger the dirty mark at all. So under no circumstance will it auto update just for changing the fields. Otherwise you would update just by reading. It still only auto update when the ComicInfo.xml field are changed. That remains unchanged.

That being said it isn't practical of not knowing if an update is required. But I fear that it may be too often. Open a book the OpenCount goes up & the Opened Time is changed, BAM an update. Read a book, the LastPageRead/CurrentPage is updated. These fields also seem less important to have updated at all cost. Having another indicator could probably overwhelm the user visually. Like how many star do we have? I don't even know what they all mean.

I could probably reuse the same indicator but simply not trigger it for all the fields. Probably exclude things that can change often, but keep it for others like custom values or catalog fields.

You can also just create an export profile with the Library Info option enabled and just export them like that instead of marking as dirty simply to update.

[chowbok]

That makes sense. I forgot that it would be updating based on read position.

[maforget]

I've added a new dirty flag to track some of the ComicBook flags. I've excluded some of the ones that change often. I've made a list of the properties here.

There is also a new smart list ("Modified Library Info") you can use to track if this flag is set or not.

Caution

Do be careful of when trying the new smart list "Modified Library Info".

Having new matchers in your library WILL result in data lost when opened in older versions where they don't exists. So if you try the test build with your library and add the new matcher, then do back to the regular nightly branch that hasn't been updated yet, your library will reset.

So when trying test builds ALWAYS maintain backups or be aware you might need to manually modify files to from the Corrupted backup done when it resets.

PS. Also note that testing this build will prevent you from getting update notification when the nightly is updated.

[chowbok]

Is there a binary of this build? I've not much experience with compiling stuff for Windows.

[maforget]

Same place, same comment, same download link. It gets updated automatically.