Fix stripped image query string in Feedbin and GReader sync mappers

[prof18]

What changed

Introduced a shared ContentImageUrlExtractor in the core module and wired it into both sync mappers:

• feedSync/feedbin — EntryDTOMapper
• feedSync/greader (FreshRSS / Miniflux) — ItemContentDTOMapper

The extractor ports RSS-Parser's current image-extraction logic: it keeps the full image URL including the query string, decodes HTML entities, and filters out emoji/smiley images. New unit tests cover the extractor and both mappers.

Why

Related to #1304 (blurry feed images).

The Feedbin and GReader mappers used an outdated image-extraction regex that truncated image URLs at the file extension, dropping the query string. For tagesschau.de feeds the image URL looks like:

https://images.tagesschau.de/image/.../schnieder-144.jpg?width=1920

Without the ?width=1920 query, the CDN returns a literal 16x9-pixel image, which is why the issue reports such extreme blur. This only affected Feedbin and GReader sync accounts — the local RSS path was already correct because RSS-Parser (>= 6.0.12) fixed the same regex upstream.

Verification

• ./gradlew --quiet --console=plain detekt allTests green
• New unit tests for ContentImageUrlExtractor, EntryDTOMapper, and ItemContentDTOMapper
• Note: existing already-synced items keep their old stored (truncated) URL until they are re-fetched; the fix applies to newly synced content going forward.

Not covered

This PR does not fix the Europapress part of #1304. That feed genuinely only publishes a 120px image, so a proper fix there would need an og:image fallback, which is a separate feature. The issue stays open.