App Store Scraper With Star Breakdown & In-App Prices

Same six-mode App Store scraper as the base (Search by AppId / Search by Keyword / Search by DeveloperId / Get Reviews / List Collection / Get Privacy Data) — same inputs, same output shapes. The difference: for the three app-listing modes, this variant also opens each app's public product page and reads Apple's own numbers instead of the base's estimate.

The base ships an histogram field, but it's not scraped — normalize_app() invents it from the average rating with hard-coded percentages, and even says so in a comment. It never reads the in-app purchase list at all. This variant adds the real ones alongside it, without touching histogram itself.

---

🔍 What this variant adds

Field	What it does
ratingBreakdown	Apple's real published 5★/4★/3★/2★/1★ counts and total, read off the product page. counts is only labelled when the weighted average recomputed from the counts matches Apple's own printed average (±0.05) — otherwise only rawCounts is filled, counts stays null, so a star is never mislabeled.
inAppPurchases	The in-app purchase names and prices Apple lists on the product page (available: false when the app has none — never a guessed empty list).
pageInformation	Seller, size, category, compatibility, location note, languages, age rating, copyright — the rest of the product page's information panel, printed the way Apple prints them.
latestVersion	The current version's "What's New" note and the release date as shown on the page.
productPageUrl / productPageFetched	The exact URL fetched and whether the fetch succeeded.

Four new input toggles, appended after every base field — includeRatingBreakdown, includeInAppPurchases, includeProductPageInfo (all default on), and minTotalRatings (default 0) to drop apps whose real rating count is below a threshold — the fast way to clear out a flattering 5.0 average built on eleven ratings.

The base's histogram field is untouched. It still gets estimated the same way it always has — this variant places the real numbers in ratingBreakdown beside it and never overwrites or hides the old field.

⚠️ Honest limits

• histogram stays wrong on purpose. Overwriting it would change base behavior; this variant adds the real distribution as new fields instead.
• Only the three app-listing modes are enriched (Search by AppId / Search by Keyword / Search by DeveloperId). Get Reviews, List Collection and Get Privacy Data return exactly what the base returns — those rows don't carry a single app record to attach a product page to.
• Apps with no ratings shelf (brand-new or unrated apps) get ratingBreakdown.available: false; minTotalRatings treats that as 0 ratings.
• inAppPurchases lists the top items Apple's page shows, not necessarily every product a developer sells.
• One extra request per app, cached per app + storefront within a run.
• maximum stays clamped at 200 by the base runtime — this variant does not raise it.

📤 Sample (real data, live-verified 2026-08-18, WhatsApp Messenger id310633997)

{
  "id": 310633997,
  "title": "WhatsApp Messenger",
  "histogram": { "1": 368504, "2": 552755, "3": 921258, "4": 3685032, "5": 12897612 }, // base, still estimated
  "ratingBreakdown": {
    "available": true,
    "average": 4.7,
    "totalRatings": 18425161,
    "counts": { "fiveStar": 15558502, "fourStar": 1434185, "threeStar": 557560, "twoStar": 248212, "oneStar": 626702 },
    "rawCounts": [15558502, 1434185, 557560, 248212, 626702],
    "source": "productPage"
  },
  "inAppPurchases": {
    "available": true,
    "itemCount": 10,
    "items": [ { "name": "No ads (1 account)", "price": "$4.00" }, /* ... */ ]
  },
  "pageInformation": {
    "seller": "WhatsApp Inc.", "sizeText": "392.2 MB", "category": "Social Networking",
    "copyright": "© 2026 WhatsApp LLC"
    // ...compatibility, location, languagesText, ageRating
  },
  "latestVersion": { "version": "Version 26.32.73", "releaseDateText": "...", "whatsNew": "..." },
  "productPageUrl": "https://apps.apple.com/us/app/id310633997",
  "productPageFetched": true
}