Closes#58
## Summary
- Add pure per-Site cover extraction beside latest-chapter parsing for all six Sites.
- Read Asura, Demonic, LightNovelWorld, and NovelFull metadata; read the Comix target detail state; read Kagane's browser-fetched `series_covers[].image_id` JSON.
- Preserve published cover URLs, percent-encode Demonic raw spaces, select Comix's smaller published `medium`, and avoid thumbnail rendition URL synthesis.
- Add live-source fixtures plus no-cover and Cloudflare challenge coverage for every Site.
## Correctness
- Scope Comix extraction to the requested series detail key, avoiding recommended posters.
- Parse Kagane's current live API shape and emit its canonical compressed image route from the published image ID; unrelated JSON fields are ignored.
- Validate Kagane image IDs against the existing UUID-shaped route constraint.
- Keep extraction pure; storage, polling, and wire integration remain outside issue #58.
## Acceptance criteria
- [x] Cover extraction exists for all six Sites in the existing latest parser module.
- [x] Each Site has a live-source fixture with source URL and date.
- [x] Comix reads the state blob, not metadata.
- [x] Demonic raw spaces are percent-encoded.
- [x] Comix returns the smaller published rendition.
- [x] No-cover pages return empty.
- [x] Cloudflare challenge pages return empty.
- [x] No thumbnail URL is synthesized by editing a published URL.
- [x] `go test ./...` passes.
## Verification
- `go test ./...`
- `go vet ./...`
- `git diff --check`
Parent issues #47 and #55 remain open as requested.
Reviewed-on: #67
Co-authored-by: Sulthan Zaki <sultankiki05@gmail.com>
Co-committed-by: Sulthan Zaki <sultankiki05@gmail.com>
sulthan
b6b88bde8a