Generalize site search for all registered sources and normalize result URLs

[kmatsunami]

Motivation

• Expand the search capability to cover all registered sources instead of the previous two hard-coded sources.
• Unify duplicate per-site search logic into a single generic extractor to simplify maintenance and add per-source configuration.
• Ensure discovered result URLs are resolved, domain-checked, and canonicalized to existing seed URL normalization logic.

Description

• Replaced hard-coded SUPPORTED_SEARCH_SOURCES with REGISTERED_SOURCES and imported normalize_seed_url from manga_watch.sources and switched quote to quote_plus for query encoding.
• Added _SOURCE_SEARCH_CONFIG to declare each source's search_url and allowed_domains and implemented _search_via_public_site to drive searches for any configured source.
• Reworked anchor parsing into _extract_anchor_results and added helpers _resolve_result_url, _is_allowed_domain, and _canonical_seed_url_for_source to resolve relative links, enforce allowed domains, and produce canonical seed URLs via normalize_seed_url.
• Removed source-specific search functions and _normalize_result_url, and built _SEARCHERS dynamically so each registered source uses the generic search path.

Testing

• Ran updated unit tests in tests/test_source_search.py which assert supported sources and seed URL normalization and they passed.
• Ran tests/test_discord_command_registration_search.py which validates the expanded search choices and it passed.
• Ran tests/test_discord_supertwins.py which checks search invocation behavior and it passed.

---

Codex Task

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9f061644f0

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Handle image-only anchors when extracting titles

Search result cards that render as <a><img alt="..."></a> now lose their human-readable title because _extract_anchor_results only looks at the anchor title attribute or text body, then falls back to the URL when both are empty. This is a regression from the previous champion-cross parser (which read img alt), and it will surface in production as URL-looking option labels that make search results hard to distinguish.

Useful? React with 👍 / 👎.