fix(javstashAutofill): correct name casing, store the stash-box id, simplify the source setting

[dude-stash]

Fixes #773. Three fixes for Javstash Autofill, version bumped to 1.4.

1. The name is now corrected to the scraper's spelling

I usually type performer names without the right casing, and the plugin never fixed them. The rename check was:

prefer_scraper = bool(use_scraper_name) and bool(cand_name) and norm(cand_name) != norm(name)

norm() lowercases and strips spaces, so yua mikami and Yua Mikami are equal there and nothing happens.

Now:

same_name = norm(cand_name) == norm(name)
prefer_scraper = (bool(cand_name) and nfc(cand_name) != nfc(name)
                  and (bool(use_scraper_name) or same_name))

A difference in casing or spacing is always corrected, because it is the same name. A really different name still needs the "use scraper name as primary" toggle, as before. find_by_name() compares with norm() now as well, so an existing "Yua Mikami" is found as the duplicate and merged, instead of the rename hitting it.

2. The stash-box id is stored

remote_site_id was missing from the scrape query, so the id was never available. It is requested now and, when the source is a stash-box, written into stash_ids in all three paths (fill, rename, merge), without touching ids from other endpoints. Identify and the Tagger recognise the performer after autofill.

urls was added to the query too, because build_update() was reading cand.get("url") and that field was never fetched.

3. The source setting no longer needs the endpoint URL

Before, the setting wanted the full endpoint URL that Stash already has under Settings → Metadata Providers. Worse, resolveStashBox only accepts an endpoint that matches a configured one exactly, so a trailing slash or a small typo failed with "stash box not found" and the hook only logged a scrape error.

resolve_source() now looks the value up in configuration.general.stashBoxes and listScrapers(types: [PERFORMER]), in this order:

1. stash-box name, e.g. StashDB
2. exact endpoint URL
3. any URL, passed through as before
4. part of a name or endpoint, e.g. stashdb
5. performer scraper id or name, e.g. Minnano-AV-JP

Empty uses javstash if it is configured, otherwise the first configured stash-box. The log line now says which source was used, for example via stash-box 'StashDB' (by name).

Old values keep working: a full URL and a scraper_id are still accepted, so nobody has to change their settings.

Testing

• Tested against my own library: a manually created performer typed in lowercase gets the correct name, the StashDB id in stash_ids, and the fields filled.
• The source resolution and the update building were also checked with a stubbed GraphQL client for every case listed above, including a stash-box that is not configured and a scraper name that does not exist.
• node validate.js plugins/javstashAutofill/javstashAutofill.yml passes.

README and CHANGELOG updated.

AI assistance was used for this change.

[garfolino]

@suzuhiroruri Love the plugin, its been very helpful. I added a few improvements. :)

[discourse-stashapp]

This pull request has been mentioned on Stash Forum. There might be relevant details there:

https://discourse.stashapp.cc/t/javstash-autofill/11502/2