Fix missing / in Illusia URLs and restore broken parseNovel selectors

[lucasaugustodeveloper]

Checklist

• Update version code if an existing plugin was modified
• Test changes in Plugin Playground or the app
• Reference related issues in the PR body (e.g. Closes #xyz)
• Commit messages follow type(scope): description (e.g. feat(<generator>): add new source)

Summary

This PR fixes three related problems found in the Portuguese plugins:

1. Illusia (root cause of the reported bug): novel/chapter URLs were being built
   without a / after the domain, e.g.
   https://illusia.com.brstory/ldm-lorde-dos-misterio instead of
   https://illusia.com.br/story/ldm-lorde-dos-misterio/.
2. Illusia (found during validation): parseNovel failed to extract the novel
   name, cover, and summary because the site switched to a new theme
   (illusia-single-story__* CSS classes), invalidating the old selectors.
3. Blog do Amon Novels (found during validation): parseNovel returned zero
   chapters because the site changed its template — the #clwd element the plugin
   scraped for the series category label now only exists inside a <script> (built
   at runtime via JS), and is no longer part of the rendered DOM.

Root cause analysis

1. Illusia — missing / in URLs

The plugin normalized every novel/chapter path by stripping both the leading and
trailing slashes:

path: novelUrl
  .replace(this.site, '')
  .replace(/^\//, '')
  .replace(/\/$/, ''),
// => "story/ldm-lorde-dos-misterio"

The plugin also had no resolveUrl defined (an existing implementation was left
commented out at the bottom of the file). When a plugin does not implement
resolveUrl, the LNReader app falls back to a plain site + path concatenation,
producing:

"https://illusia.com.br" + "story/ldm-lorde-dos-misterio"
=> "https://illusia.com.brstory/ldm-lorde-dos-misterio"   // missing "/"

2. Illusia — outdated selectors after theme change

A live check against the site showed the current markup uses a custom Illusia theme:

• Title: <h1 class="illusia-single-story__title"> (was h1.story__identity-title)
• Cover: <img class="illusia-single-story__cover-img" data-src="...">
  (was figure.story__thumbnail img)
• Summary: <div class="illusia-single-story__description">
  (was section.story__summary)

Author (a[href*="/author/"]), genres (div.tag-group > a), status
(span.story__status), chapter list (li.chapter-group__list-item), and chapter
content (section#chapter-content) were verified against the live pages and still
match, so they were left untouched.

3. Blog do Amon Novels — chapter list extraction broken

The old implementation derived the Blogger series label from the text of a #clwd
element:

const cat = loadedCheerio('#clwd').text().split("'")[1];

The site's new template builds #clwd dynamically via jQuery inside a <script>
tag, so it no longer exists in the raw HTML and cat became undefined, falling
into a branch that scraped a now-absent #chapters element and returned 0 chapters.

The new template declares the label inline in the page:

<script>
  let startIndex = 2;
  const maxResults = 30;
  const categoria = "Makio";
  ...
</script>

Fetching feeds/posts/default/-/{categoria}?alt=json with that extracted label
returns the full chapter feed (confirmed against the live site: 394 entries for the
tested novel).

Changes

plugins/portuguese/illusia.ts (1.0.2 → 1.0.3)

• Enabled resolveUrl so paths keep their current normalized form and the app
  builds correct absolute URLs:

  resolveUrl = (path: string) => `${this.site}/${path}/`;

• Updated parseNovel selectors (old ones kept as fallbacks):
  • name: added h1.illusia-single-story__title
  • cover: added img.illusia-single-story__cover-img (data-src first, since
    src holds an SVG lazy-load placeholder)
  • summary: added div.illusia-single-story__description

plugins/portuguese/blogdoamonnovels.ts (1.0.1 → 1.0.2)

• Extract the series label from the inline script, keeping the old #clwd method
  as a fallback:

  const cat =
    /categoria\s*=\s*"([^"]+)"/.exec(body)?.[1] ??
    loadedCheerio('#clwd').text().split("'")[1];

Other plugins audited

The remaining plugins in plugins/portuguese/ were checked for the missing-slash
bug and do not have it, so they were left unchanged:

Plugin	Affected?	Why
tsundoku.ts	No	Paths keep the leading / (.replace(this.site, '')), so the app's site + path fallback already yields a valid URL
novelmania.ts	No	Paths always start with /novels/... and the plugin already implements resolveUrl

Testing

All four plugins in plugins/portuguese/ were exercised against their live sites
with npm run check:plugin:

Plugin	popularNovels	searchNovels	parseNovel	parseChapter
illusia.ts	✅ 10 novels	✅ 2 results	✅ 1437 chapters	✅ 18630 chars
blogdoamonnovels.ts	✅ 12 novels	✅ 1 result	✅ 95 chapters	✅ 12896 chars
tsundoku.ts	✅ 48 novels	✅ 1 result	✅ 115 chapters	✅ 881 chars
novelmania.ts	✅ 20 novels	✅ 2 results	✅ 6 chapters	✅ 32866 chars

URL construction verified in the check output:
https://illusia.com.br/story/ldm-lorde-dos-misterios/ (correct, with /).

ESLint and Prettier pass on both modified files with no warnings.

Note: plugins/portuguese/novelmania.ts has a pre-existing Prettier formatting
warning on the committed code. It was intentionally left untouched to keep this PR
focused.

[greptile-apps]

Confidence Score: 5/5

The PR appears safe to merge, with no accepted new findings or outstanding blocking findings.

Summary

This PR repairs Portuguese source plugins affected by site-template and URL-construction changes.

• Updates Illusia’s title, cover, and summary selectors while retaining legacy fallbacks.
• Adds URL resolution for Illusia and now preserves lowercase HTTP(S) absolute URLs.
• Extracts Blog do Amon Novels’ category from its inline script, retaining the former DOM-based fallback.
• Increments both modified plugin versions.

_{Reviews (2) · Last reviewed commit: "Update plugins/portuguese/illusia.ts"}

[rajarsheechatterjee]

This is superseded by the rewrite in #2567.