feat(market-research): add play.google.com charts scraper
Replaces the AppBrain scraper (Cloudflare-blocked 403, confirmed dead).
play.google.com server-rendered chart HTML works: live-validated, real
Android packages + names + ratings from both /store/apps/top and
/store/apps/category/<CAT>. CLI: <top|category> [--category CAT]
[--region R] [--limit N] [--html-file F]; output source:"google-play",
entries {rank,name,package,rating}. Package link is the stable field;
title (Epkrse class) and rating are best-effort.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
fatih.bulut
parent
c8eb023712
commit
fbfa3aece7
|
|
@ -1,20 +1,18 @@
|
|||
#!/usr/bin/env python3
|
||||
"""Fetch an AppBrain Play-store top-chart into a normalized JSON list.
|
||||
"""Fetch a play.google.com chart page into a normalized JSON list.
|
||||
|
||||
Google Play has no public chart feed and renders via obfuscated batchexecute
|
||||
JS. AppBrain (appbrain.com) publishes server-rendered HTML top-chart pages with
|
||||
REAL Android package names. This scrapes those rows. Stdlib-only, no pip.
|
||||
Google Play has no public chart feed, but its server-rendered HTML for
|
||||
/store/apps/top and /store/apps/category/<CAT> carries ranked app cards with
|
||||
REAL Android package names (unlike Apple's RSS iOS bundle ids). Each card is an
|
||||
<a href="/store/apps/details?id=PKG"> wrapping the icon, then a
|
||||
<div class="Epkrse ">NAME</div> title, then a star aria-label. Play obfuscates
|
||||
the title class, so the package LINK (stable) is the load-bearing field; name
|
||||
and rating are best-effort. Stdlib-only, no pip.
|
||||
|
||||
Unlike Apple's RSS (iOS bundle ids), entries here carry Android packages usable
|
||||
directly for a clone-app handoff.
|
||||
(AppBrain, the original plan's source, is Cloudflare-blocked 403 — dropped.)
|
||||
"""
|
||||
import sys, json, re, argparse, urllib.request, ssl, subprocess, shutil
|
||||
import sys, json, re, html as _html, argparse, urllib.request, ssl, subprocess, shutil
|
||||
|
||||
CHARTS = {
|
||||
"popular": "https://www.appbrain.com/apps/popular/",
|
||||
"top-grossing": "https://www.appbrain.com/apps/highest-grossing/",
|
||||
"top-new": "https://www.appbrain.com/apps/new/",
|
||||
}
|
||||
UA = "Mozilla/5.0"
|
||||
|
||||
def _http_get(url):
|
||||
|
|
@ -33,58 +31,66 @@ def _http_get(url):
|
|||
raise
|
||||
return out.stdout.decode("utf-8", "replace")
|
||||
|
||||
# One app row: a link to /app/<slug>/<package> followed (within the same row
|
||||
# block) by developer / category / rating / installs spans.
|
||||
ROW = re.compile(
|
||||
r'href="/app/[^"/]+/(?P<package>[\w.]+)"[^>]*>(?P<name>[^<]+)</a>'
|
||||
r'(?P<rest>.*?)(?=href="/app/|</div>\s*</div>|$)', re.DOTALL)
|
||||
DEV = re.compile(r'class="developer"[^>]*>([^<]+)<')
|
||||
CAT = re.compile(r'class="category"[^>]*>([^<]+)<')
|
||||
RAT = re.compile(r'class="rating"[^>]*>\s*([\d.]+)\s*<')
|
||||
INST = re.compile(r'class="installs"[^>]*>\s*([\d,]+\+)\s*<')
|
||||
LINK = re.compile(r'href="/store/apps/details\?id=([a-zA-Z0-9._]+)"')
|
||||
NAME = re.compile(r'class="Epkrse[^"]*"[^>]*>([^<]+)<')
|
||||
# rating is best-effort: search-style aria-label OR a star-icon followed by a number
|
||||
RATE_ARIA = re.compile(r'aria-label="Rated\s+([\d.]+)\s+star')
|
||||
RATE_STAR = re.compile(r'>star</[^>]+>\s*([\d.]+)')
|
||||
|
||||
def _first(rx, text):
|
||||
m = rx.search(text)
|
||||
return m.group(1).strip() if m else None
|
||||
def _chart_url(chart, category, region):
|
||||
if chart == "category":
|
||||
return f"https://play.google.com/store/apps/category/{category}?hl=en&gl={region}"
|
||||
return f"https://play.google.com/store/apps/top?hl=en&gl={region}"
|
||||
|
||||
def parse(html, chart, region, limit):
|
||||
def parse(html, chart_label, region, limit):
|
||||
entries = []
|
||||
for i, m in enumerate(ROW.finditer(html), start=1):
|
||||
if i > limit:
|
||||
break
|
||||
rest = m.group("rest")
|
||||
rating = _first(RAT, rest)
|
||||
seen = set()
|
||||
for m in LINK.finditer(html):
|
||||
pkg = m.group(1)
|
||||
if pkg in seen:
|
||||
continue
|
||||
win = html[m.start():m.start() + 1600] # one card's worth of markup
|
||||
nm = NAME.search(win)
|
||||
if not nm:
|
||||
continue # icon-only anchor with no title nearby; skip
|
||||
seen.add(pkg)
|
||||
rt = RATE_ARIA.search(win) or RATE_STAR.search(win)
|
||||
entries.append({
|
||||
"rank": i,
|
||||
"name": m.group("name").strip(),
|
||||
"developer": _first(DEV, rest),
|
||||
"category": _first(CAT, rest),
|
||||
"package": m.group("package"),
|
||||
"rating": float(rating) if rating else None,
|
||||
"installs": _first(INST, rest),
|
||||
"rank": len(entries) + 1,
|
||||
"name": _html.unescape(nm.group(1).strip()),
|
||||
"package": pkg,
|
||||
"rating": float(rt.group(1)) if rt else None,
|
||||
})
|
||||
return {"source": "appbrain", "chart": chart, "region": region,
|
||||
if len(entries) >= limit:
|
||||
break
|
||||
return {"source": "google-play", "chart": chart_label, "region": region,
|
||||
"count": len(entries), "entries": entries}
|
||||
|
||||
def main():
|
||||
ap = argparse.ArgumentParser()
|
||||
ap.add_argument("chart", choices=sorted(CHARTS))
|
||||
ap.add_argument("--region", default="us")
|
||||
ap.add_argument("chart", choices=("top", "category"))
|
||||
ap.add_argument("--category")
|
||||
ap.add_argument("--region", default="US")
|
||||
ap.add_argument("--limit", type=int, default=25)
|
||||
ap.add_argument("--html-file")
|
||||
args = ap.parse_args()
|
||||
|
||||
if args.chart == "category" and not args.category:
|
||||
print("ERROR: --category is required for the 'category' chart", file=sys.stderr)
|
||||
sys.exit(2)
|
||||
chart_label = f"category:{args.category}" if args.chart == "category" else "top"
|
||||
|
||||
if args.html_file:
|
||||
with open(args.html_file, encoding="utf-8") as f:
|
||||
html = f.read()
|
||||
else:
|
||||
try:
|
||||
html = _http_get(CHARTS[args.chart])
|
||||
html = _http_get(_chart_url(args.chart, args.category, args.region))
|
||||
except Exception as e:
|
||||
print(f"ERROR: failed to fetch AppBrain chart: {e}", file=sys.stderr)
|
||||
print(f"ERROR: failed to fetch Play chart: {e}", file=sys.stderr)
|
||||
sys.exit(1)
|
||||
|
||||
print(json.dumps(parse(html, args.chart, args.region, args.limit), indent=2))
|
||||
print(json.dumps(parse(html, chart_label, args.region, args.limit), indent=2))
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
|
|
|
|||
|
|
@ -1,18 +0,0 @@
|
|||
<!DOCTYPE html><html><body>
|
||||
<div class="app-list">
|
||||
<div class="app">
|
||||
<a class="app-icon-name" href="/app/habit-tracker/com.example.habit">Habit Tracker</a>
|
||||
<span class="developer">Focus Labs</span>
|
||||
<span class="category">Productivity</span>
|
||||
<span class="rating">4.6</span>
|
||||
<span class="installs">5,000,000+</span>
|
||||
</div>
|
||||
<div class="app">
|
||||
<a class="app-icon-name" href="/app/budget-buddy/com.example.budget">Budget Buddy</a>
|
||||
<span class="developer">Money Inc</span>
|
||||
<span class="category">Finance</span>
|
||||
<span class="rating">4.2</span>
|
||||
<span class="installs">1,000,000+</span>
|
||||
</div>
|
||||
</div>
|
||||
</body></html>
|
||||
|
|
@ -0,0 +1,9 @@
|
|||
<!DOCTYPE html><html><body>
|
||||
<div role="main">
|
||||
<a href="/store/apps/details?id=com.example.habit" jslog="38003; track:click"><div class="TjRVLb"><img src="https://play-lh.googleusercontent.com/x=s256"></div><div class="Epkrse ">Habit Tracker</div></a>
|
||||
<div aria-label="Rated 4.6 stars out of five stars"></div>
|
||||
<a href="/store/apps/details?id=com.example.budget" jslog="38003; track:click"><div class="TjRVLb"><img src="https://play-lh.googleusercontent.com/y=s256"></div><div class="Epkrse ">Budget Buddy</div></a>
|
||||
<div aria-label="Rated 4.2 stars out of five stars"></div>
|
||||
<a href="/store/apps/details?id=com.example.habit" jslog="38003; track:click"><div class="Epkrse ">Habit Tracker</div></a>
|
||||
</div>
|
||||
</body></html>
|
||||
|
|
@ -3,11 +3,11 @@ import json, subprocess, sys, os
|
|||
|
||||
HERE = os.path.dirname(os.path.abspath(__file__))
|
||||
SCRIPT = os.path.join(HERE, "..", "skills", "market-research", "scripts", "fetch-play-charts.py")
|
||||
FIXTURE = os.path.join(HERE, "fixtures", "appbrain-popular.html")
|
||||
FIXTURE = os.path.join(HERE, "fixtures", "play-chart.html")
|
||||
|
||||
def run():
|
||||
out = subprocess.check_output(
|
||||
[sys.executable, SCRIPT, "popular", "--region", "us", "--html-file", FIXTURE])
|
||||
[sys.executable, SCRIPT, "top", "--region", "US", "--html-file", FIXTURE])
|
||||
return json.loads(out)
|
||||
|
||||
def main():
|
||||
|
|
@ -17,20 +17,19 @@ def main():
|
|||
print(f"{'PASS' if cond else 'FAIL'}: {name}")
|
||||
if not cond: fails.append(name)
|
||||
|
||||
check("source", d["source"] == "appbrain")
|
||||
check("chart", d["chart"] == "popular")
|
||||
check("count", d["count"] == 2)
|
||||
check("source", d["source"] == "google-play")
|
||||
check("chart", d["chart"] == "top")
|
||||
check("region", d["region"] == "US")
|
||||
check("count (dedup app1 repeat)", d["count"] == 2)
|
||||
e0 = d["entries"][0]
|
||||
check("rank 1", e0["rank"] == 1)
|
||||
check("name", e0["name"] == "Habit Tracker")
|
||||
check("package", e0["package"] == "com.example.habit")
|
||||
check("developer", e0["developer"] == "Focus Labs")
|
||||
check("category", e0["category"] == "Productivity")
|
||||
check("rating", e0["rating"] == 4.6)
|
||||
check("installs", e0["installs"] == "5,000,000+")
|
||||
for k in ["rank", "name", "package", "developer", "category", "rating", "installs"]:
|
||||
for k in ["rank", "name", "package", "rating"]:
|
||||
check(f"key present: {k}", k in e0)
|
||||
check("second package", d["entries"][1]["package"] == "com.example.budget")
|
||||
check("second name", d["entries"][1]["name"] == "Budget Buddy")
|
||||
sys.exit(1 if fails else 0)
|
||||
|
||||
main()
|
||||
|
|
|
|||
Loading…
Reference in New Issue