XellossNakama's Patches?

[dylan-k]

This fork of the plugin appears to be the most up-to-date, so I thought to suggest this here. @frcooper would you consider any pull requests to integrate the patches created by XellossNakama?

Among other things, the patches make more efficient use of the ComicVine API, useful considering its tighter restrictions nowadays.

[frcooper]

#3 - how exactly does this make better use of the api?

[dylan-k]

It appears to make more efficient use of the API by making less frequent calls to it. This way the API won't hit your limits as quickly/often. The code is a available for review but a pull request might make it easier for comparison.

…

__ dk

________________________________ From: FRC ***@***.***> Sent: Wednesday, October 23, 2024 11:48:37 PM To: frcooper/comic-vine-scraper ***@***.***> Cc: Dylan Kinnett ***@***.***>; Author ***@***.***> Subject: Re: [frcooper/comic-vine-scraper] XellossNakama's Patches? (Issue #2) CAUTION: This email originated from a sender outside of Walters Art Museum! Please proceed with caution and DO NOT CLICK links, open attachments or provide ANY information if the sender is unknown, you were NOT expecting the email message or the email message contains suspicious looking content. Please pay particular attention to email messages requesting you to verify accounts of any kind or requests to change your password(s) that YOU DID NOT initiate. #3<https://cas5-0-urlprotect.trendmicro.com:443/wis/clicktime/v1/query?url=https%3a%2f%2fgithub.com%2ffrcooper%2fcomic%2dvine%2dscraper%2fpull%2f3&umid=37fecbbc-9b3b-445e-8aac-744dfae9d503&auth=1471ab15ac03abb12a16568b5dd91c9353c2869e-97ee355ec684870ded4931ebedabfcb8fa320aca> - how exactly does this make better use of the api? — Reply to this email directly, view it on GitHub<https://cas5-0-urlprotect.trendmicro.com:443/wis/clicktime/v1/query?url=https%3a%2f%2fgithub.com%2ffrcooper%2fcomic%2dvine%2dscraper%2fissues%2f2%23issuecomment%2d2434191089&umid=37fecbbc-9b3b-445e-8aac-744dfae9d503&auth=1471ab15ac03abb12a16568b5dd91c9353c2869e-7b003053eed6340fb6399c9bad8a3cb3768e91cd>, or unsubscribe<https://cas5-0-urlprotect.trendmicro.com:443/wis/clicktime/v1/query?url=https%3a%2f%2fgithub.com%2fnotifications%2funsubscribe%2dauth%2fAAQ2ILDNWC2MNRQVWCH3HT3Z5BUZLAVCNFSM6AAAAABQMVSOMWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMZUGE4TCMBYHE&umid=37fecbbc-9b3b-445e-8aac-744dfae9d503&auth=1471ab15ac03abb12a16568b5dd91c9353c2869e-9e2b7fea9cc9aae5056c7cf448abeead307bdd58>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[frcooper]

a pull request might make it easier for comparison.
Yeah. I know. That's why I made one.

[dylan-k]

how exactly does this make better use of the api?

Some of the changes, particularly in cvdb.py are likely to yield greater efficiency, especially in handling ambiguous or high-volume search terms, by improving term normalization, expanding result limits, and implementing prioritization logic. These updates should reduce API calls in scenarios where large or nuanced search queries are involved.

here are the main updates:

• Increased Result Limit (__max_search_results)
• Improved Search Term Cleaning (Unicode Normalization and Filtering)
• Unicode Normalization and Accent Removal, potentially reducing requests by preventing mismatches due to minor variations in search terms.
• New Matching Criteria Logic, improves efficiency in cases with large result sets, reducing excess queries.
• Enhanced Series Matching and Caching (series_refs2), reducing the total number of potential requests by identifying and storing only top matches.
• Error Handling Adjustments, may prevent unnecessary pagination requests

As for what's going on with images, there could be some more explanation of the changes in the earlier post:

If you use the "try to select the correct series automatically" in the script options... this doesn't only check data and select the first option if the match is good enough... IT COMPARE THE IMAGES OF THE PREVIEW AND THE COVER and only if they are similar enough, it continues automatically... It already did what I always wanted! (I never used the select the series automatically because I was afraid of incorrect matches). You can even edit the code (it is automatcher.py and tune the __MATCH_THRESHOLD (the default is 87% of similarity, I tune it to 93, more than that and the "barcodes" or things like that that are different from digitals to paper covers will make them different enough many times)