diff options
| author | yyamashita <yyamashita@hetzner.yyamashita.com> | 2026-06-18 12:51:55 +0900 |
|---|---|---|
| committer | yyamashita <yyamashita@hetzner.yyamashita.com> | 2026-06-18 12:51:55 +0900 |
| commit | 272f7d079cbac81c9af86b9f210ace8cabc0e98c (patch) | |
| tree | b1489cfe1628ded107d02839ebaed53b551211ef /app/scrapers | |
| parent | 243d222c35f4607ac13a400e827333f44a0f44c7 (diff) | |
Fix HTML entity leakage in FAD scraper and Pitbar _clone date bug
fad-yokohama: replace innerHTML manipulation with Cheerio .clone().text()
so that HTML entities ( etc.) are decoded before storing titles.
pitbar-nishiogikubo: strip non-numeric suffix from day part of element ID
to handle _clone-suffixed IDs that freecalend.com generates after JS renders.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Diffstat (limited to 'app/scrapers')
| -rw-r--r-- | app/scrapers/fad-yokohama.ts | 9 | ||||
| -rw-r--r-- | app/scrapers/pitbar-nishiogikubo.ts | 4 |
2 files changed, 6 insertions, 7 deletions
diff --git a/app/scrapers/fad-yokohama.ts b/app/scrapers/fad-yokohama.ts index f8f7cbc..ddf3fb7 100644 --- a/app/scrapers/fad-yokohama.ts +++ b/app/scrapers/fad-yokohama.ts @@ -71,11 +71,10 @@ function parsePageEvents( const $firstP = $b("p").first(); if (!$firstP.length) continue; - // Get text with <br/> as newlines - const rawHtml = $firstP.html() ?? ""; - const lines = rawHtml - .replace(/<br\s*\/?>/gi, "\n") - .replace(/<[^>]+>/g, " ") + // Get text with <br/> as newlines; use Cheerio .text() to decode HTML entities + const $clone = $firstP.clone(); + $clone.find("br").replaceWith("\n"); + const lines = $clone.text() .split("\n") .map((l) => l.replace(/\s+/g, " ").trim()) .filter(Boolean); diff --git a/app/scrapers/pitbar-nishiogikubo.ts b/app/scrapers/pitbar-nishiogikubo.ts index 2553002..569ce26 100644 --- a/app/scrapers/pitbar-nishiogikubo.ts +++ b/app/scrapers/pitbar-nishiogikubo.ts @@ -38,10 +38,10 @@ async function extractEvents(page: Page, dateFrom: string, dateTo: string): Prom return Array.from(cells).map((el) => { const id = el.getAttribute("id") ?? ""; const parts = id.split("-"); - // id: cal-25771-YYYY-M-D + // id: cal-25771-YYYY-M-D (sometimes suffixed with _clone after JS cloning) const year = parts[2]; const month = parts[3]; - const day = parts[4]; + const day = parts[4]?.replace(/\D.*$/, ""); if (!year || !month || !day) return null; return { date: `${year}-${month.padStart(2, "0")}-${day.padStart(2, "0")}`, |