bugrahans's picture
Google Maps Leads Scraper Space
b5bc75e verified
Raw History Blame Contribute Delete
22.2 kB
from __future__ import annotations
import csv
import html
import tempfile
from typing import Any, Dict, List, Optional, Tuple
import gradio as gr
from tool import SCRAPE_DO_SIGNUP_URL, SORT_OPTIONS, ScrapeDoGoogleMapsTool
# ZeroGPU hardware refuses to keep a Space alive unless it finds at least one
# @spaces.GPU function at startup. This app is pure HTTP and never touches a GPU,
# so the probe below exists only to satisfy that check. It is never called.
try:
import spaces
@spaces.GPU(duration=1)
def _zerogpu_probe() -> str:
return "ok"
except ImportError:
pass
UTM = "utm_source=huggingface&utm_medium=google-maps-leads-scraper"
NO_TOKEN_MD = (
"### Paste your Scrape.do API token above\n\n"
f"No token yet? [Create a free account]({SCRAPE_DO_SIGNUP_URL}) — "
"1,000 free credits a month, no credit card. Every call is 10 credits; one search page returns 20 places."
)
SORT_LABELS = {"newest": "Newest", "most_relevant": "Most relevant", "highest_rating": "Highest rating", "lowest_rating": "Lowest rating"}
PLACE_HEADERS = ["Name", "Category", "Rating", "Reviews", "Phone", "Website", "Address", "Hours", "Place ID"]
# --------------------------------------------------------------------- helpers
def _esc(value: Any) -> str:
return html.escape("" if value is None else str(value), quote=True)
def _safe_url(url: Any) -> str:
url = str(url or "")
return url if url.startswith("https://") else ""
def _tool(token: str) -> Optional[ScrapeDoGoogleMapsTool]:
token = (token or "").strip()
return ScrapeDoGoogleMapsTool(api_token=token) if token else None
def _write_csv(rows: List[Dict[str, Any]], prefix: str) -> Optional[str]:
if not rows:
return None
fields: List[str] = []
for row in rows:
for key in row:
if key not in fields:
fields.append(key)
handle = tempfile.NamedTemporaryFile(mode="w", suffix=".csv", prefix=f"{prefix}-", delete=False, newline="", encoding="utf-8-sig")
writer = csv.DictWriter(handle, fieldnames=fields, extrasaction="ignore")
writer.writeheader()
for row in rows:
writer.writerow({k: (" | ".join(map(str, v)) if isinstance(v, list) else (str(v) if isinstance(v, dict) else v)) for k, v in row.items()})
handle.close()
return handle.name
def _error_md(data: Dict[str, Any]) -> str:
out = f"**That didn't go through.**\n\n{data.get('errorMessage') or 'Unknown error'}"
if data.get("hint"):
out += f"\n\n{data['hint']}"
return out
def _stars(rating: Any) -> str:
try:
r = float(rating)
except (TypeError, ValueError):
return '<span class="gm-muted">No rating</span>'
return f'<span class="gm-star">★ {r:.1f}</span>'
def render_places(places: List[Dict[str, Any]]) -> str:
cards = []
for p in places:
site = _safe_url(p.get("website"))
maps = _safe_url(p.get("maps_url"))
name = _esc(p.get("name"))
title = f'<a href="{_esc(maps)}" target="_blank" rel="noopener noreferrer">{name}</a>' if maps else name
reviews = f' <span class="gm-muted">({p["reviews"]:,})</span>' if isinstance(p.get("reviews"), int) else ""
lines = [f'<div class="gm-line gm-muted">{_esc(p.get("category"))}</div>']
if p.get("address"):
lines.append(f'<div class="gm-line">{_esc(p["address"])}</div>')
contact = []
if p.get("phone"):
contact.append(_esc(p["phone"]))
if site:
short = site.split("//", 1)[-1].split("/", 1)[0]
contact.append(f'<a href="{_esc(site)}" target="_blank" rel="noopener noreferrer">{_esc(short)}</a>')
if contact:
lines.append(f'<div class="gm-line">{" · ".join(contact)}</div>')
if p.get("open_state"):
lines.append(f'<div class="gm-line gm-open">{_esc(p["open_state"])}</div>')
cards.append(f'<div class="gm-card"><div class="gm-card-top"><span class="gm-name">{title}</span>'
f'<span>{_stars(p.get("rating"))}{reviews}</span></div>{"".join(lines)}</div>')
if not cards:
cards.append('<div class="gm-empty">No places came back for this search.</div>')
return f'<div class="gm-wrap"><div class="gm-grid">{"".join(cards)}</div></div>'
def render_place(p: Dict[str, Any]) -> str:
breakdown = p.get("rating_breakdown") or {}
total = sum(breakdown.values()) or 1
bars = "".join(
f'<div class="gm-bar-row"><span>{s}★</span><div class="gm-bar"><div style="width:{100 * breakdown.get(s, 0) / total:.0f}%"></div></div>'
f'<span class="gm-muted">{breakdown.get(s, 0):,}</span></div>'
for s in ("5", "4", "3", "2", "1") if breakdown
)
attrs = "".join(f'<div class="gm-attr"><b>{_esc(k.replace("_", " ").title())}</b> {_esc(", ".join(v))}</div>'
for k, v in (p.get("attributes") or {}).items())
site = _safe_url(p.get("website"))
rows = [
("Category", _esc(", ".join(p.get("categories") or [])) or "—"),
("Rating", f'{_stars(p.get("rating"))} <span class="gm-muted">{p.get("reviews") or 0:,} reviews</span>'),
("Phone", _esc(p.get("phone")) or "—"),
("Website", f'<a href="{_esc(site)}" target="_blank" rel="noopener noreferrer">{_esc(site)}</a>' if site else "—"),
("Address", _esc(p.get("address")) or "—"),
("Plus code", _esc(p.get("plus_code")) or "—"),
("Coordinates", f'{p.get("latitude")}, {p.get("longitude")}' if p.get("latitude") is not None else "—"),
("Hours", _esc(p.get("hours")).replace("; ", "<br>") or "—"),
("Place ID", f'<code>{_esc(p.get("place_id"))}</code>'),
("Feature ID", f'<code>{_esc(p.get("data_id"))}</code>'),
]
specs = "".join(f'<div class="gm-spec-name">{k}</div><div class="gm-spec-value">{v}</div>' for k, v in rows)
maps = _safe_url(p.get("maps_url"))
button = f'<a class="gm-btn" href="{_esc(maps)}" target="_blank" rel="noopener noreferrer">Open in Google Maps</a>' if maps else ""
extra = f'<div class="gm-bars">{bars}</div>' if bars else ""
extra += f'<div class="gm-attrs">{attrs}</div>' if attrs else ""
return (f'<div class="gm-wrap"><div class="gm-profile"><div class="gm-p-name">{_esc(p.get("name"))}</div>'
f'<div class="gm-specs">{specs}</div>{extra}{button}</div></div>')
def render_reviews(reviews: List[Dict[str, Any]]) -> str:
items = []
for r in reviews:
reply = f'<div class="gm-reply"><b>Owner reply</b> {_esc(r.get("owner_reply"))}</div>' if r.get("owner_reply") else ""
guide = ' <span class="gm-badge">Local Guide</span>' if r.get("local_guide") else ""
items.append(f'<div class="gm-review"><div class="gm-card-top"><span class="gm-name">{_esc(r.get("author"))}{guide}</span>'
f'<span>{_stars(r.get("rating"))} <span class="gm-muted">{_esc(r.get("date"))}</span></span></div>'
f'<div class="gm-line">{_esc(r.get("text")) or "<span class=gm-muted>(rating only)</span>"}</div>{reply}</div>')
if not items:
items.append('<div class="gm-empty">No reviews came back.</div>')
return f'<div class="gm-wrap">{"".join(items)}</div>'
# ---------------------------------------------------------------- handlers
def run_search(token: str, query: str, location: str, page: float):
tool = _tool(token)
empty = gr.Dataframe(value=[], headers=PLACE_HEADERS)
hidden = gr.Dropdown(choices=[], value=None, visible=False)
if tool is None:
return NO_TOKEN_MD, "", empty, {}, None, hidden
data = tool.search(query, location, int(page or 1))
if data.get("status") != "success":
return _error_md(data), "", empty, data, None, hidden
places = data["results"]
phones = sum(1 for p in places if p.get("phone"))
sites = sum(1 for p in places if p.get("website"))
summary = (f"**{len(places)}** places on page {data['page']} · **{phones}** with a phone · **{sites}** with a website"
f"\n\n1 request · {data['credits_used']} credits. Pick a place below to send it to the Place and Reviews tabs.")
table = gr.Dataframe(value=[[p["name"], p["category"], p["rating"], p["reviews"], p["phone"], p["website"], p["address"], p["hours"], p["place_id"]]
for p in places], headers=PLACE_HEADERS)
choices = [(f'{p["name"]} · {p["address"]}', p["place_id"] or p["data_id"]) for p in places if p.get("place_id") or p.get("data_id")]
return summary, render_places(places), table, data, _write_csv(places, "google-maps-places"), gr.Dropdown(choices=choices, value=None, visible=bool(choices))
def send_place(picked: Optional[str], p_cur: str, r_cur: str) -> Tuple[str, str, str]:
if not picked:
return p_cur, r_cur, ""
return picked, picked, "Sent. Open the **Place** or **Reviews** tab and press the button."
def run_place(token: str, ref: str):
tool = _tool(token)
if tool is None:
return NO_TOKEN_MD, "", {}, None
if not (ref or "").strip():
return "Paste a place ID or Google Maps link, or pick a place in the Search tab.", "", {}, None
data = tool.place(ref)
if data.get("status") != "success":
return _error_md(data), "", data, None
p = data["place"]
return f"1 request · {data['credits_used']} credits", render_place(p), data, _write_csv([p], "google-maps-place")
def run_reviews(token: str, ref: str, sort_label: str, pages: float):
tool = _tool(token)
if tool is None:
return NO_TOKEN_MD, "", {}, None
if not (ref or "").strip():
return "Paste a place ID or feature ID, or pick a place in the Search tab.", "", {}, None
sort_by = next((k for k, v in SORT_LABELS.items() if v == sort_label), "newest")
data = tool.reviews(ref, sort_by, int(pages or 1))
if data.get("status") != "success":
return _error_md(data), "", data, None
reviews = data["reviews"]
replies = sum(1 for r in reviews if r.get("owner_reply"))
topics = ", ".join(f'{t["keyword"]} ({t["mentions"]})' for t in data.get("topics") or [] if t.get("keyword"))
summary = (f"**{len(reviews)}** reviews · **{replies}** with an owner reply"
+ (f"\n\nTopics: {topics}" if topics else "") + f"\n\n{data['credits_used']} credits used")
return summary, render_reviews(reviews), data, _write_csv(reviews, "google-maps-reviews")
# ----------------------------------------------------------------------- UI
STYLE = """
<style>
.gradio-container, .gradio-container button, .gradio-container input, .gradio-container textarea, .gradio-container select {
font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", "Helvetica Neue", Arial, sans-serif; }
.gm-hero { position: relative; overflow: hidden; border-radius: 18px; padding: 30px 32px 28px 32px; margin-bottom: 16px;
background: radial-gradient(circle at 90% 15%, rgba(26,115,232,.40) 0%, rgba(26,115,232,0) 45%),
linear-gradient(135deg, #10161f 0%, #1b2430 60%, #24303e 100%); }
.gm-hero, .gm-hero * { color: #ffffff !important; }
.gm-hero-inner { position: relative; z-index: 1; max-width: 640px; }
.gm-hero h1 { font-size: 2.05rem !important; font-weight: 800 !important; margin: 0 0 10px 0 !important; }
.gm-hero h1 span { color: #8ab4f8 !important; }
.gm-hero p { margin: 0 0 8px 0 !important; font-size: 1rem !important; line-height: 1.55 !important; opacity: .96; }
.gm-hero a { text-decoration: underline !important; }
.gm-hero a.gm-cta { display: inline-block; margin-top: 12px; padding: 11px 24px; border-radius: 10px; background: #1a73e8 !important;
color: #ffffff !important; font-weight: 800 !important; text-decoration: none !important; }
.gm-hero-card { position: absolute; right: 34px; top: 50%; transform: translateY(-50%); width: 270px; z-index: 1;
background: #ffffff; border-radius: 14px; padding: 14px 16px; box-shadow: 0 10px 28px rgba(0,0,0,.35); }
.gm-hero-card, .gm-hero-card * { color: #202124 !important; }
.gm-hc-title { font-size: .72rem; font-weight: 800; color: #5f6368 !important; text-transform: uppercase; letter-spacing: .05em; margin-bottom: 8px; }
.gm-hc-row { display: grid; grid-template-columns: 1fr auto; gap: 8px; font-size: .84rem; margin: 7px 0; }
.gm-hc-row b { color: #1a73e8 !important; }
@media (max-width: 900px) { .gm-hero-card { display: none; } }
.gm-note { border-left: 3px solid #1a73e8; background: #eef4fe; padding: 12px 16px; border-radius: 0 10px 10px 0; font-size: .9rem;
margin-bottom: 12px; color: #202124 !important; }
.gradio-container .primary, .gradio-container button.primary, .gradio-container button[variant="primary"] {
background: #1a73e8 !important; background-image: none !important; border-color: #1a73e8 !important; color: #ffffff !important;
border-radius: 10px !important; font-weight: 800 !important; }
.gradio-container .primary:hover, .gradio-container button.primary:hover { background: #1557b0 !important; }
.gm-wrap { color: #202124 !important; }
.gm-muted { color: #5f6368 !important; }
.gm-empty { background: #ffffff; border: 1px solid #e8eaed; border-radius: 12px; padding: 20px; color: #5f6368 !important; }
.gm-grid { display: grid; grid-template-columns: repeat(auto-fill, minmax(min(290px, 100%), 1fr)); gap: 12px; }
.gm-card, .gm-review { background: #ffffff; border: 1px solid #e8eaed; border-radius: 14px; padding: 12px 14px; }
.gm-review { margin-bottom: 10px; }
.gm-card:hover { box-shadow: 0 3px 12px rgba(32,33,36,.10); }
.gm-card-top { display: flex; justify-content: space-between; gap: 8px; align-items: baseline; }
.gm-name, .gm-name a { font-weight: 800; color: #202124 !important; text-decoration: none !important; }
.gm-name a:hover, .gm-line a { color: #1a73e8 !important; }
.gm-star { color: #b06000 !important; font-weight: 800; white-space: nowrap; }
.gm-line { font-size: .84rem; margin-top: 4px; color: #202124 !important; }
.gm-open { color: #188038 !important; }
.gm-badge { background: #e8f0fe; color: #1a73e8 !important; font-size: .7rem; border-radius: 6px; padding: 1px 6px; font-weight: 700; }
.gm-reply { margin-top: 8px; padding: 8px 10px; background: #f8f9fa; border-radius: 8px; font-size: .82rem; color: #202124 !important; }
.gm-profile { background: #ffffff; border: 1px solid #e8eaed; border-radius: 16px; padding: 18px 22px; }
.gm-p-name { font-size: 1.4rem; font-weight: 800; color: #202124 !important; }
.gm-specs { display: grid; grid-template-columns: minmax(120px, .3fr) 1fr; border-top: 1px solid #ececec; margin-top: 12px; }
.gm-spec-name, .gm-spec-value { padding: 8px 10px; border-bottom: 1px solid #ececec; font-size: .86rem; }
.gm-spec-name { color: #5f6368 !important; }
.gm-spec-value, .gm-spec-value a { color: #202124 !important; }
.gm-bars { margin-top: 14px; max-width: 360px; }
.gm-bar-row { display: grid; grid-template-columns: 28px 1fr 50px; gap: 8px; align-items: center; font-size: .8rem; margin: 3px 0; color: #202124 !important; }
.gm-bar { background: #e8eaed; border-radius: 4px; height: 8px; overflow: hidden; }
.gm-bar div { background: #fbbc04; height: 8px; }
.gm-attrs { margin-top: 12px; font-size: .82rem; }
.gm-attr { margin: 3px 0; color: #202124 !important; }
.gm-btn { display: inline-block; margin-top: 14px; padding: 10px 22px; border-radius: 10px; background: #1a73e8 !important;
color: #ffffff !important; font-weight: 800; text-decoration: none !important; font-size: .9rem; }
footer { display: none !important; }
</style>
"""
HERO = f"""
<div class="gm-hero">
<div class="gm-hero-inner">
<h1>Google Maps <span>Leads</span> Scraper</h1>
<p><strong>Search a category around any point and get every business as a lead.</strong> Name, phone, website,
address, rating, review count, hours and coordinates, then full place details and reviews with owner replies.</p>
<p>Google Maps has no data in its HTML and answers scripts with consent walls and CAPTCHAs. This Space uses the
<a href="https://scrape.do/?{UTM}" target="_blank" rel="noopener noreferrer">Scrape.do</a> Google Maps API, which returns parsed JSON.</p>
<a class="gm-cta" href="{SCRAPE_DO_SIGNUP_URL}" target="_blank" rel="noopener noreferrer">Get a free API token →</a>
</div>
<div class="gm-hero-card" aria-hidden="true">
<div class="gm-hc-title">One query · 6 pages · 60 credits</div>
<div class="gm-hc-row"><span>Places</span><span><b>104</b></span></div>
<div class="gm-hc-row"><span>With a phone</span><span><b>77</b></span></div>
<div class="gm-hc-row"><span>With a website</span><span><b>96</b></span></div>
</div>
</div>
"""
CREDIT_NOTE = """
<div class="gm-note">
Every call is 10 Scrape.do credits: a search page returns 20 places, a place lookup is one place, a reviews page is about 10 reviews.
One query stops at about 120 places (6 pages); to cover a whole city, search again around other points.
Reviews carry the names of the people who wrote them: store only what you need and follow the privacy rules that apply to you.
</div>
"""
FOOTER_MD = f"""
---
### Use it as a smolagents Tool
```python
from smolagents import CodeAgent, InferenceClientModel
from tool import ScrapeDoGoogleMapsTool
agent = CodeAgent(
tools=[ScrapeDoGoogleMapsTool(api_token="YOUR_SCRAPEDO_TOKEN")],
model=InferenceClientModel(),
)
agent.run("Find dentists around 30.2672,-97.7431 on Google Maps and list the ones without a website.")
```
Or set `SCRAPE_DO_API_TOKEN` in the environment and instantiate with no arguments.
Prefer plain Python? `pip install google-maps-leads-scraper` ([GitHub](https://github.com/bugrahansss/google-maps-leads-scraper)) ·
No-code: [Apify actor](https://apify.com/scrapedo/google-maps-leads-scraper) ·
Guide: [Google Maps Scraper: How to Scrape Google Maps for Leads](https://scrape.do/blog/google-maps-scraping/?{UTM})
[Scrape.do](https://scrape.do/?{UTM}) · [Documentation](https://scrape.do/documentation/) · [Sign up free]({SCRAPE_DO_SIGNUP_URL})
"""
def build_demo() -> gr.Blocks:
with gr.Blocks(title="Google Maps Leads Scraper — Scrape.do") as demo:
gr.HTML(STYLE)
gr.HTML(HERO)
api_token = gr.Textbox(label="Scrape.do API token", placeholder="Paste your token here", type="password",
info="Free tier: 1,000 credits a month. 10 credits per call.")
gr.HTML(CREDIT_NOTE)
with gr.Tabs():
with gr.Tab("Search"):
with gr.Row():
query = gr.Textbox(label="What to search", value="coffee shop", scale=3,
info="As you would type it into Google Maps. You can add a place: 'dentist in Denver, CO'.")
location = gr.Textbox(label="Around (lat,lng)", value="30.2672,-97.7431", scale=2,
info="Optional. Leave empty to use the place in the query.")
page = gr.Number(label="Page (1-6)", value=1, minimum=1, maximum=6, precision=0, scale=1)
search_btn = gr.Button("Search Google Maps", variant="primary")
search_summary = gr.Markdown()
search_view = gr.HTML()
search_table = gr.Dataframe(headers=PLACE_HEADERS, wrap=True, label="Places")
search_csv = gr.File(label="Download leads as CSV")
pick = gr.Dropdown(label="Send a place to the Place and Reviews tabs", choices=[], value=None, visible=False)
pick_note = gr.Markdown()
with gr.Accordion("Raw JSON response", open=False):
search_json = gr.JSON()
with gr.Tab("Place"):
place_ref = gr.Textbox(label="Place ID, CID, feature ID or Google Maps link", value="",
placeholder="ChIJ... or https://www.google.com/maps/place/...",
info="Pick a place in the Search tab, or paste one.")
place_btn = gr.Button("Get place details", variant="primary")
place_summary = gr.Markdown()
place_view = gr.HTML()
place_csv = gr.File(label="Download as CSV")
with gr.Accordion("Raw JSON response", open=False):
place_json = gr.JSON()
with gr.Tab("Reviews"):
with gr.Row():
review_ref = gr.Textbox(label="Place ID or feature ID", value="", scale=3, placeholder="ChIJ... or 0x...:0x...")
sort = gr.Dropdown(choices=[SORT_LABELS[s] for s in SORT_OPTIONS], value="Newest", label="Order", scale=1)
rpages = gr.Number(label="Pages (about 10 each)", value=1, minimum=1, maximum=5, precision=0, scale=1)
reviews_btn = gr.Button("Get reviews", variant="primary")
reviews_summary = gr.Markdown()
reviews_view = gr.HTML()
reviews_csv = gr.File(label="Download reviews as CSV")
with gr.Accordion("Raw JSON response", open=False):
reviews_json = gr.JSON()
s_in = [api_token, query, location, page]
s_out = [search_summary, search_view, search_table, search_json, search_csv, pick]
search_btn.click(fn=run_search, inputs=s_in, outputs=s_out)
query.submit(fn=run_search, inputs=s_in, outputs=s_out)
pick.change(fn=send_place, inputs=[pick, place_ref, review_ref], outputs=[place_ref, review_ref, pick_note])
p_out = [place_summary, place_view, place_json, place_csv]
place_btn.click(fn=run_place, inputs=[api_token, place_ref], outputs=p_out)
place_ref.submit(fn=run_place, inputs=[api_token, place_ref], outputs=p_out)
r_out = [reviews_summary, reviews_view, reviews_json, reviews_csv]
reviews_btn.click(fn=run_reviews, inputs=[api_token, review_ref, sort, rpages], outputs=r_out)
gr.Markdown(FOOTER_MD)
return demo
# Hugging Face Spaces launches the module-level `demo` object rather than
# calling launch() itself, so build it at import time.
demo = build_demo()
if __name__ == "__main__":
demo.queue(default_concurrency_limit=4).launch(ssr_mode=False)