Posted under e621 Tools and Applications
Tag Master is a tagging helper for e621 (and e926) featuring:
Please see the GitHub project page for up to date instructions.
Thank you for this, this sounds very useful. I especially like the wiki snippets which will probably be the main feature I'll use. The popup of that seems to behave strangely under a few circumstances though. I'm using the latest Firefox.
I also noticed that the data.json is quite large, somewhere around 35MB. This file also includes information on tags with very little usage, maybe you could provide a lightweight version with only tags above the count of 50 (or some other arbitrary number), this should probably shave of a few MB. Omitting implies and impliedBy when there are no values and falling back to the empty array in code would also save a bit, probably not as much as the other thing thought. But thats probably why you made it both client/serverside. I just prefer running stuff locally if I have the chance.
Anyways, thanks for making this, looking forward to what else you might do.
Thank you for giving it a shot and the feedback! I'm very happy to see this project finally get some use.
The wiki snippets were a pain to parse out because of how much variation there is in the wiki page contents, but I'm very glad I did. They're extremely useful. The #related tag groups are also retrieved from the wiki text.
Rewriting the snippet popups is something I already started earlier, but ultimately I decided that I needed to focus on other features. I'll position them using JS in the future.
I agree that the data.json file is large. It exists because I know I myself wouldn't want to send my searches to a server and I'm glad to see I'm not alone on that. I was hoping that the file would be compressed in transit, but unfortunately GitHub doesn't allow that. Compressing the data is another thing I put off to release the tool earlier. Removing the empty implies and impliedBy arrays is quite an obvious improvement and I don't know how I missed it.
I initially didn't want to omit tags with low usage because a lot of them are perfectly valid tags that simply don't get much use. Having Tag Master suggest them could bring them to life. You inspired me to look at the actual numbers and I just changed my mind.
Let's look at the numbers. Dropping all tags that have only 2 posts reduces the file size by 27%. Requiring at least 5 posts reduces it by 45%. Removing the empty arrays reduces the file size by 29%. Removing empty arrays and requiring at least 2 posts adds up to a 44% reduction. At least 5 posts and no empty arrays is a 55% reduction. Both of these are quite significant improvements as loading the data can take upwards of a few seconds simply because of how long the string is.
I'll remove the empty arrays and adding a separate data set containing tags with at least a few posts. The settings menu could use some more settings anyway.
