feat: sync scripts, scraper, and sql tools from server
This commit is contained in:
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
@@ -0,0 +1,8 @@
|
||||
nohup: ignoring input
|
||||
🏙️ [PROD] Starting full Damascus scan on Mac (Low-Resource)...
|
||||
|
||||
🚀 PASS: COMMERCIAL_PUBLIC (Low-Resource Mode)
|
||||
📋 [DEBUG] Grid: 44x66 blocks. Starting...
|
||||
🔎 [TRACE] Scanning Block 0,0 | 33.45000, 36.17000
|
||||
🛡️ Bypassing Google Consent (ID: #L2AGLb)...
|
||||
🛡️ Bypassing Google Consent (ID: #L2AGLb)...
|
||||
@@ -0,0 +1,227 @@
|
||||
import asyncio
|
||||
import os
|
||||
import requests
|
||||
import json
|
||||
import re
|
||||
import math
|
||||
import time
|
||||
from urllib.parse import unquote
|
||||
from playwright.async_api import async_playwright
|
||||
from dotenv import load_dotenv
|
||||
|
||||
# Environment Setup
|
||||
load_dotenv()
|
||||
API_URL = os.getenv("API_URL", "http://localhost:3200/api")
|
||||
API_KEY = os.getenv("API_KEY", "intaleq_secret_2026")
|
||||
|
||||
# PRODUCTION SCAN CALIBRATION (Damascus Pro - High Performance Mac Edition)
|
||||
DAMASCUS_BOUNDS = {
|
||||
"min_lat": 33.4500,
|
||||
"max_lat": 33.5800,
|
||||
"min_lng": 36.1700,
|
||||
"max_lng": 36.4000
|
||||
}
|
||||
|
||||
PASSES = {
|
||||
"RESIDENTIAL": {
|
||||
"categories": ["منزل", "بيت", "بناء سكني", "دخلة", "حارة"],
|
||||
"lat_step": 0.0015,
|
||||
"lng_step": 0.0018,
|
||||
"zoom": 20,
|
||||
"scrolls": 4
|
||||
},
|
||||
"COMMERCIAL_PUBLIC": {
|
||||
"categories": ["سوق", "محل", "مطعم", "كافيه", "صيدلية", "عيادة", "جامع", "وزارة", "مدرسة"],
|
||||
"lat_step": 0.0030,
|
||||
"lng_step": 0.0035,
|
||||
"zoom": 18,
|
||||
"scrolls": 3
|
||||
}
|
||||
}
|
||||
|
||||
class DamascusProScanner:
|
||||
def __init__(self, bounds: dict):
|
||||
self.bounds = bounds
|
||||
self.global_seen = set()
|
||||
self.coord_pattern = re.compile(r"!3d(-?\d+\.\d+)!4d(-?\d+\.\d+)|/@(-?\d+\.\d+),(-?\d+\.\d+)")
|
||||
|
||||
async def route_interceptor(self, route):
|
||||
"""السماح بكافة الموارد على الـ Mac لضمان دقة النتائج"""
|
||||
return await route.continue_()
|
||||
|
||||
def calculate_distance(self, lat1, lon1, lat2, lon2):
|
||||
R = 6371
|
||||
d_lat = math.radians(lat2 - lat1)
|
||||
d_lon = math.radians(lon2 - lon1)
|
||||
a = math.sin(d_lat / 2)**2 + math.cos(math.radians(lat1)) * math.cos(math.radians(lat2)) * math.sin(d_lon / 2)**2
|
||||
return R * (2 * math.asin(math.sqrt(a)))
|
||||
|
||||
async def bypass_consent(self, page):
|
||||
"""تخطي صفحة موافقة جوجل بمختلف أشكالها (للعمل بسلاسة في السيرفر)"""
|
||||
try:
|
||||
selectors = [
|
||||
"#L2AGLb",
|
||||
"button[aria-label*='Accept']",
|
||||
"button:has-text('أوافق')",
|
||||
"button:has-text('I agree')",
|
||||
"form[action*='consent'] button"
|
||||
]
|
||||
for selector in selectors:
|
||||
if await page.locator(selector).count() > 0:
|
||||
print(f" 🛡️ Bypassing Google Consent (Selector: {selector})...", flush=True)
|
||||
await page.click(selector)
|
||||
await page.wait_for_timeout(2000)
|
||||
return True
|
||||
return False
|
||||
except:
|
||||
return False
|
||||
|
||||
async def wait_for_results(self, page):
|
||||
"""الانتظار الذكي لنتائج البحث أو صفحة الخريطة"""
|
||||
try:
|
||||
await page.wait_for_selector('a[href*="/maps/place/"]', timeout=15000)
|
||||
return True
|
||||
except:
|
||||
return False
|
||||
|
||||
async def extract_places(self, page, center_lat, center_lng, pass_name) -> list[dict]:
|
||||
await self.bypass_consent(page)
|
||||
|
||||
# التمرير لأسفل لتحميل المزيد من النتائج
|
||||
for _ in range(PASSES[pass_name].get('scrolls', 3)):
|
||||
await page.mouse.wheel(0, 4000)
|
||||
await page.wait_for_timeout(1000)
|
||||
|
||||
# الانتظار حتى تظهر النتائج فعلياً
|
||||
await self.wait_for_results(page)
|
||||
|
||||
places = []
|
||||
elements = await page.query_selector_all('a[href*="/maps/place/"]')
|
||||
if len(elements) > 0:
|
||||
print(f" ✨ Found {len(elements)} potential results on page", flush=True)
|
||||
|
||||
for el in elements:
|
||||
try:
|
||||
href = await el.get_attribute('href')
|
||||
if not href: continue
|
||||
name_match = re.search(r"/place/([^/]+)", href)
|
||||
if not name_match: continue
|
||||
name = unquote(name_match.group(1)).replace('+', ' ').strip()
|
||||
lat, lng = None, None
|
||||
matches = self.coord_pattern.findall(href)
|
||||
for m in matches:
|
||||
valid = [val for val in m if val]
|
||||
if len(valid) == 2:
|
||||
lat, lng = float(valid[0]), float(valid[1])
|
||||
break
|
||||
if name and lat and lng:
|
||||
dist = self.calculate_distance(center_lat, center_lng, lat, lng)
|
||||
if dist > 3.0: continue # زيادة المدى قليلاً
|
||||
FORBIDDEN = ["google", "تكبير", "تصغير", "بحث", "مساحة", "إبلاغ", "عن"]
|
||||
if any(key in name.lower() for key in FORBIDDEN): continue
|
||||
dedup_key = f"{name.lower()}|{round(lat, 5)}|{round(lng, 5)}"
|
||||
if dedup_key not in self.global_seen:
|
||||
self.global_seen.add(dedup_key)
|
||||
places.append({
|
||||
"name": name, "name_ar": name, "latitude": lat, "longitude": lng,
|
||||
"category": pass_name.title(), "city": "Damascus", "source": "scraper_admin"
|
||||
})
|
||||
except: continue
|
||||
return places
|
||||
|
||||
def send_batch(self, places: list[dict]):
|
||||
if not places: return
|
||||
try:
|
||||
res = requests.post(
|
||||
f"{API_URL}/geocoding/upsert-batch",
|
||||
headers={"X-API-Key": API_KEY},
|
||||
json={"places": places},
|
||||
timeout=20
|
||||
)
|
||||
if res.status_code in [200, 201]:
|
||||
print(f" ✅ [SYNC SUCCESS] {len(places)} points added. Response: {res.text[:50]}", flush=True)
|
||||
else:
|
||||
print(f" ❌ [SYNC ERROR] Status: {res.status_code} | Body: {res.text}", flush=True)
|
||||
except Exception as e:
|
||||
print(f" ⚠️ [SYNC EXCEPTION] {str(e)}", flush=True)
|
||||
|
||||
async def run_pass(self, pass_name: str, test_mode=False):
|
||||
cfg = PASSES[pass_name]
|
||||
print(f"\n🚀 PASS: {pass_name} (Low-Resource Mode)", flush=True)
|
||||
async with async_playwright() as p:
|
||||
user_agent = "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36"
|
||||
|
||||
# استخدام مجلد دائم لحفظ الـ Cookies والحالة لضمان تجاوز الحجب
|
||||
user_data_dir = os.path.join(os.getcwd(), "browser_state")
|
||||
context = await p.chromium.launch_persistent_context(
|
||||
user_data_dir,
|
||||
headless=not test_mode,
|
||||
viewport={"width": 1024, "height": 768},
|
||||
locale="ar-SA",
|
||||
user_agent=user_agent
|
||||
)
|
||||
|
||||
# حدود البحث (تصغير المنطقة جداً في وضع الاختبار)
|
||||
bounds = {
|
||||
"min_lat": 33.5100, "max_lat": 33.5110,
|
||||
"min_lng": 36.2900, "max_lng": 36.2915
|
||||
} if test_mode else self.bounds
|
||||
|
||||
rows = math.ceil((bounds['max_lat'] - bounds['min_lat']) / cfg['lat_step'])
|
||||
cols = math.ceil((bounds['max_lng'] - bounds['min_lng']) / cfg['lng_step'])
|
||||
|
||||
print(f"📋 [DEBUG] Grid: {rows}x{cols} blocks. Starting...", flush=True)
|
||||
|
||||
for r in range(rows):
|
||||
for c in range(cols):
|
||||
lat = bounds['min_lat'] + (r * cfg['lat_step'])
|
||||
lng = bounds['min_lng'] + (c * cfg['lng_step'])
|
||||
|
||||
print(f"🔎 [TRACE] Scanning Block {r},{c} | {lat:.5f}, {lng:.5f}", flush=True)
|
||||
|
||||
page = await context.new_page()
|
||||
await page.route("**/*", self.route_interceptor)
|
||||
|
||||
for cat in cfg['categories']:
|
||||
url = f"https://www.google.com/maps/search/{cat}/@{lat},{lng},{cfg['zoom']}z?hl=ar"
|
||||
try:
|
||||
await page.goto(url, wait_until="domcontentloaded", timeout=40000)
|
||||
|
||||
# Trace & Handle Consent via CSS ID (More reliable)
|
||||
title = await page.title()
|
||||
if "قبل المتابعة" in title or "Before you continue" in title or "Consent" in title:
|
||||
print(f" 🛡️ Bypassing Google Consent (ID: #L2AGLb)...", flush=True)
|
||||
try:
|
||||
await page.click("button#L2AGLb")
|
||||
await page.wait_for_load_state("networkidle", timeout=10000)
|
||||
except: pass
|
||||
|
||||
# الانتظار حتى تظهر أول نتيجة بحث أو انتهاء الوقت
|
||||
try:
|
||||
await page.wait_for_selector('a[href*="/maps/place/"]', timeout=8000)
|
||||
except: pass
|
||||
|
||||
for _ in range(cfg['scrolls']):
|
||||
await page.mouse.wheel(0, 1000)
|
||||
await asyncio.sleep(0.5 if test_mode else 0.4)
|
||||
|
||||
found = await self.extract_places(page, lat, lng, pass_name)
|
||||
if found:
|
||||
self.send_batch(found)
|
||||
print(f" ✅ [{pass_name}] Block {r},{c} | Found {len(found)} results", flush=True)
|
||||
except Exception as e:
|
||||
print(f" ⚠️ [ERROR] {str(e)[:40]}", flush=True)
|
||||
|
||||
await page.close()
|
||||
await asyncio.sleep(1.0 if test_mode else 0.6)
|
||||
await browser.close()
|
||||
|
||||
if __name__ == "__main__":
|
||||
scanner = DamascusProScanner(DAMASCUS_BOUNDS)
|
||||
async def main():
|
||||
print("🏙️ [PROD] Starting full Damascus scan on Mac (Low-Resource)...", flush=True)
|
||||
# تشغيل المسح التجاري أولاً
|
||||
await scanner.run_pass("COMMERCIAL_PUBLIC", test_mode=False)
|
||||
# ثم المسح السكني الكثيف
|
||||
await scanner.run_pass("RESIDENTIAL", test_mode=False)
|
||||
asyncio.run(main())
|
||||
@@ -0,0 +1,173 @@
|
||||
import asyncio
|
||||
import os
|
||||
import requests
|
||||
import json
|
||||
import re
|
||||
import math
|
||||
import urllib.parse
|
||||
from urllib.parse import unquote
|
||||
from playwright.async_api import async_playwright
|
||||
from dotenv import load_dotenv
|
||||
|
||||
# Environment Setup
|
||||
load_dotenv()
|
||||
API_URL = os.getenv("API_URL", "http://localhost:3200/api") # القيمة الافتراضية للسيرفر
|
||||
API_KEY = os.getenv("API_KEY", "intaleq_secret_2026")
|
||||
|
||||
# PRODUCTION SCAN CALIBRATION (Damascus Pro - High Performance Server Edition)
|
||||
DAMASCUS_BOUNDS = {
|
||||
"min_lat": 33.4500,
|
||||
"max_lat": 33.5800,
|
||||
"min_lng": 36.1700,
|
||||
"max_lng": 36.4000
|
||||
}
|
||||
|
||||
PASSES = {
|
||||
"COMMERCIAL_PUBLIC": {
|
||||
"categories": ["سوق", "محل", "مطعم", "كافيه", "صيدلية", "عيادة", "جامع", "وزارة", "مدرسة"],
|
||||
"lat_step": 0.0030,
|
||||
"lng_step": 0.0035,
|
||||
"zoom": 18,
|
||||
"scrolls": 3
|
||||
},
|
||||
"RESIDENTIAL": {
|
||||
"categories": ["منزل", "بيت", "بناء سكني", "دخلة", "حارة"],
|
||||
"lat_step": 0.0015,
|
||||
"lng_step": 0.0018,
|
||||
"zoom": 20,
|
||||
"scrolls": 4
|
||||
}
|
||||
}
|
||||
|
||||
class DamascusProScanner:
|
||||
def __init__(self, bounds: dict):
|
||||
self.bounds = bounds
|
||||
self.global_seen = set()
|
||||
self.coord_pattern = re.compile(r"!3d(-?\d+\.\d+)!4d(-?\d+\.\d+)|/@(-?\d+\.\d+),(-?\d+\.\d+)")
|
||||
|
||||
def calculate_distance(self, lat1, lon1, lat2, lon2):
|
||||
R = 6371
|
||||
d_lat = math.radians(lat2 - lat1)
|
||||
d_lon = math.radians(lon2 - lon1)
|
||||
a = math.sin(d_lat / 2)**2 + math.cos(math.radians(lat1)) * math.cos(math.radians(lat2)) * math.sin(d_lon / 2)**2
|
||||
return R * (2 * math.asin(math.sqrt(a)))
|
||||
|
||||
async def bypass_consent(self, page):
|
||||
"""تخطي صفحة موافقة جوجل بمختلف أشكالها"""
|
||||
try:
|
||||
selectors = ["#L2AGLb", "button[aria-label*='Accept']", "button:has-text('أوافق')", "button:has-text('I agree')"]
|
||||
for selector in selectors:
|
||||
if await page.locator(selector).count() > 0:
|
||||
await page.click(selector)
|
||||
await asyncio.sleep(1)
|
||||
return True
|
||||
return False
|
||||
except: return False
|
||||
|
||||
async def extract_places(self, page, center_lat, center_lng, pass_name) -> list[dict]:
|
||||
await self.bypass_consent(page)
|
||||
places = []
|
||||
try:
|
||||
elements = await page.query_selector_all('a[href*="/maps/place/"]')
|
||||
for el in elements:
|
||||
try:
|
||||
href = await el.get_attribute('href')
|
||||
if not href: continue
|
||||
name_match = re.search(r"/place/([^/]+)", href)
|
||||
if not name_match: continue
|
||||
name = urllib.parse.unquote(name_match.group(1)).replace('+', ' ').strip()
|
||||
lat, lng = None, None
|
||||
matches = self.coord_pattern.findall(href)
|
||||
for m in matches:
|
||||
valid = [val for val in m if val]
|
||||
if len(valid) == 2:
|
||||
lat, lng = float(valid[0]), float(valid[1])
|
||||
break
|
||||
if name and lat and lng:
|
||||
dist = self.calculate_distance(center_lat, center_lng, lat, lng)
|
||||
if dist > 3.5: continue
|
||||
if any(key in name.lower() for key in ["google", "بحث", "إبلاغ"]): continue
|
||||
dedup_key = f"{name.lower()}|{round(lat, 5)}|{round(lng, 5)}"
|
||||
if dedup_key not in self.global_seen:
|
||||
self.global_seen.add(dedup_key)
|
||||
places.append({
|
||||
"name": name, "name_ar": name, "latitude": lat, "longitude": lng,
|
||||
"category": pass_name.title(), "city": "Damascus", "source": "scraper_admin"
|
||||
})
|
||||
except: continue
|
||||
except: pass
|
||||
return places
|
||||
|
||||
def send_batch(self, places: list[dict]):
|
||||
if not places: return
|
||||
try:
|
||||
res = requests.post(f"{API_URL}/geocoding/upsert-batch", headers={"X-API-Key": API_KEY}, json={"places": places}, timeout=15)
|
||||
if res.status_code in [200, 201]:
|
||||
print(f" ✅ [SYNC SUCCESS] {len(places)} points added.", flush=True)
|
||||
else:
|
||||
print(f" ❌ [SYNC ERROR] {res.status_code}", flush=True)
|
||||
except Exception as e:
|
||||
print(f" ⚠️ [SYNC EXCEPTION] {str(e)[:50]}", flush=True)
|
||||
|
||||
async def run_pass(self, pass_name: str, test_mode=False):
|
||||
cfg = PASSES[pass_name]
|
||||
print(f"\n🚀 PASS: {pass_name} (Nuclear Mode Active)", flush=True)
|
||||
|
||||
async with async_playwright() as p:
|
||||
user_data_dir = os.path.join(os.getcwd(), "browser_state_pro")
|
||||
context = await p.chromium.launch_persistent_context(
|
||||
user_data_dir,
|
||||
headless=True,
|
||||
viewport={"width": 1280, "height": 720},
|
||||
user_agent="Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36"
|
||||
)
|
||||
|
||||
# حقن الكوكيز فوراً
|
||||
await context.add_cookies([
|
||||
{"name": "CONSENT", "value": "YES+cb.20230531-17-p0.en+FX+908", "domain": ".google.com", "path": "/"},
|
||||
{"name": "SOCS", "value": "CAISHAgBEhJnd3NfMjAyMzA4MzAtMF9SQzIaAnByIAE", "domain": ".google.com", "path": "/"}
|
||||
])
|
||||
|
||||
page = await context.new_page()
|
||||
await page.set_extra_http_headers({"Accept-Language": "ar-SY,ar;q=0.9,en-US;q=0.8"})
|
||||
|
||||
rows = math.ceil((self.bounds['max_lat'] - self.bounds['min_lat']) / cfg['lat_step'])
|
||||
cols = math.ceil((self.bounds['max_lng'] - self.bounds['min_lng']) / cfg['lng_step'])
|
||||
|
||||
for r in range(rows):
|
||||
for c in range(cols):
|
||||
lat = self.bounds['min_lat'] + (r * cfg['lat_step'])
|
||||
lng = self.bounds['min_lng'] + (c * cfg['lng_step'])
|
||||
|
||||
try:
|
||||
async with asyncio.timeout(45): # مهلة 45 ثانية لكل كتلة
|
||||
print(f"🔎 [BLOCK {r},{c}] {lat:.4f}, {lng:.4f} | ", end="", flush=True)
|
||||
cat = cfg['categories'][0] # نكتفي بصنف واحد حالياً للسرعة الكلية
|
||||
url = f"https://www.google.com/maps/search/{cat}/@{lat},{lng},{cfg['zoom']}z?hl=ar"
|
||||
|
||||
await page.goto(url, wait_until="domcontentloaded", timeout=30000)
|
||||
await self.bypass_consent(page)
|
||||
|
||||
for _ in range(cfg['scrolls']):
|
||||
await page.mouse.wheel(0, 1500)
|
||||
await asyncio.sleep(0.3)
|
||||
|
||||
found = await self.extract_places(page, lat, lng, pass_name)
|
||||
if found:
|
||||
self.send_batch(found)
|
||||
print(f"✅ Found {len(found)}", flush=True)
|
||||
else:
|
||||
print(f"⚪ Empty", flush=True)
|
||||
except Exception as e:
|
||||
print(f"⚠️ Skip: {str(e)[:20]}", flush=True)
|
||||
continue
|
||||
|
||||
await context.close()
|
||||
|
||||
if __name__ == "__main__":
|
||||
scanner = DamascusProScanner(DAMASCUS_BOUNDS)
|
||||
async def main():
|
||||
print("🏙️ [PROD] Damascus Scanner PRO - Server Edition", flush=True)
|
||||
await scanner.run_pass("COMMERCIAL_PUBLIC")
|
||||
await scanner.run_pass("RESIDENTIAL")
|
||||
asyncio.run(main())
|
||||
Executable
+20
@@ -0,0 +1,20 @@
|
||||
#!/bin/bash
|
||||
# Download Existing Syria Data from Server to Mac
|
||||
# To be run on the Mac terminal
|
||||
|
||||
SERVER_IP="188.68.36.205"
|
||||
SERVER_USER="hamzadoctor"
|
||||
REMOTE_PATH="/home/hamzadoctor/app"
|
||||
KEY_PATH="/Users/hamzaaleghwairyeen/.ssh/doctory-key"
|
||||
|
||||
echo "🌍 Connecting to server to dump existing Syria landmarks..."
|
||||
|
||||
# 1. Generate dump on server
|
||||
ssh -i "$KEY_PATH" "$SERVER_USER@$SERVER_IP" "docker exec -t map-db pg_dump -U mapuser -d mapdb -t places_syria --data-only --inserts > $REMOTE_PATH/existing_syria_data.sql"
|
||||
|
||||
# 2. Download the dump to local Mac
|
||||
echo "📥 Downloading dump to Mac..."
|
||||
rsync -avz -e "ssh -i $KEY_PATH" "$SERVER_USER@$SERVER_IP:$REMOTE_PATH/existing_syria_data.sql" infrastructure/docker/postgis/
|
||||
|
||||
echo "✅ Download complete: infrastructure/docker/postgis/existing_syria_data.sql"
|
||||
echo "Now run ./infrastructure/scripts/local_db_prep.sh to merge."
|
||||
Executable
+42
@@ -0,0 +1,42 @@
|
||||
#!/bin/bash
|
||||
# Script to execute Syria data import on the server
|
||||
# To be run from /home/hamzadoctor/app
|
||||
|
||||
APP_DIR="/home/hamzadoctor/app"
|
||||
CSV_FILE="syria_final_complete.csv"
|
||||
SQL_FILE="infrastructure/docker/postgis/import_syria_csv.sql"
|
||||
|
||||
echo "📦 Transferring CSV to PostGIS container..."
|
||||
docker cp "$APP_DIR/$CSV_FILE" map-db:/tmp/syria_data.csv
|
||||
|
||||
echo "💾 Running SQL import script..."
|
||||
# We use docker compose exec db psql to run the logic
|
||||
# First, update the SQL logic to use the /tmp path for COPY
|
||||
# We'll create a temporary SQL wrapper to handle the COPY command with the correct path
|
||||
|
||||
docker exec -i map-db psql -U mapuser -d mapdb <<EOF
|
||||
-- Load schema
|
||||
\i /home/hamzadoctor/app/$SQL_FILE
|
||||
|
||||
-- Perform COPY (must be done in the container pointing to /tmp/syria_data.csv)
|
||||
CREATE TEMP TABLE staging_syria_temp (
|
||||
name TEXT,
|
||||
latitude DECIMAL(10, 8),
|
||||
longitude DECIMAL(11, 8),
|
||||
category TEXT,
|
||||
address TEXT,
|
||||
description TEXT,
|
||||
created_at TIMESTAMP
|
||||
);
|
||||
|
||||
COPY staging_syria_temp(name, latitude, longitude, category, address, description, created_at)
|
||||
FROM '/tmp/syria_data.csv'
|
||||
WITH (FORMAT csv, HEADER false, QUOTE '"', ENCODING 'UTF8');
|
||||
|
||||
INSERT INTO staging_syria SELECT * FROM staging_syria_temp;
|
||||
DROP TABLE staging_syria_temp;
|
||||
|
||||
-- Final merge logic is already in the SQL file loaded via \i
|
||||
EOF
|
||||
|
||||
echo "✅ Import completed successfully!"
|
||||
@@ -0,0 +1,52 @@
|
||||
#!/bin/bash
|
||||
# Script to import Egypt and Jordan OSM data into the PostGIS database
|
||||
# نص برمجي لاستيراد بيانات مصر والأردن إلى قاعدة البيانات
|
||||
|
||||
set -e
|
||||
|
||||
DATA_DIR="./infrastructure/osm-data"
|
||||
DB_USER=${POSTGRES_USER:-mapuser}
|
||||
DB_NAME=${POSTGRES_DB:-mapdb}
|
||||
|
||||
echo "🌍 Starting MENA Regional Data Import..."
|
||||
echo "🌍 البدء في استيراد البيانات الإقليمية..."
|
||||
|
||||
# 1. Check for files
|
||||
if [ ! -f "$DATA_DIR/jordan-latest.osm.pbf" ]; then
|
||||
echo "❌ Jordan PBF missing. Please run setup-osm.sh first."
|
||||
exit 1
|
||||
fi
|
||||
|
||||
if [ ! -f "$DATA_DIR/egypt-latest.osm.pbf" ]; then
|
||||
echo "📥 Downloading Egypt OSM PBF..."
|
||||
curl -L https://download.geofabrik.de/africa/egypt-latest.osm.pbf -o "$DATA_DIR/egypt-latest.osm.pbf"
|
||||
fi
|
||||
|
||||
# 2. Detect Docker Compose Command
|
||||
if command -v docker-compose &> /dev/null; then
|
||||
DOKCER_COMPOSE="docker-compose"
|
||||
else
|
||||
DOKCER_COMPOSE="docker compose"
|
||||
fi
|
||||
|
||||
echo "🐳 Using $DOKCER_COMPOSE..."
|
||||
|
||||
# 3. Import Jordan (Append)
|
||||
echo "🇯🇴 Importing Jordan data (Append mode)..."
|
||||
$DOKCER_COMPOSE --profile import run --rm osm-import osm2pgsql \
|
||||
--append --slim --cache 1000 \
|
||||
--database "$DB_NAME" --host db --user "$DB_USER" \
|
||||
/data/jordan-latest.osm.pbf
|
||||
|
||||
# 4. Import Egypt (Append)
|
||||
echo "🇪🇬 Importing Egypt data (Append mode)..."
|
||||
$DOKCER_COMPOSE --profile import run --rm osm-import osm2pgsql \
|
||||
--append --slim --cache 1000 \
|
||||
--database "$DB_NAME" --host db --user "$DB_USER" \
|
||||
/data/egypt-latest.osm.pbf
|
||||
|
||||
echo "✅ Import complete. Restarting tile and routing services..."
|
||||
$DOKCER_COMPOSE restart martin routing
|
||||
|
||||
echo "🚀 MENA Regional mapping is now live!"
|
||||
echo "🚀 تم تفعيل خرائط المنطقة بنجاح!"
|
||||
Executable
+110
@@ -0,0 +1,110 @@
|
||||
#!/bin/bash
|
||||
# Local Database Preparation & Merge Script
|
||||
# To be run on the Mac terminal after download_server_data.sh
|
||||
|
||||
set -e
|
||||
|
||||
echo "🚀 Starting local database merge and preparation..."
|
||||
|
||||
# 1. Start the DB container
|
||||
echo "📦 Starting PostGIS container..."
|
||||
docker compose up -d db
|
||||
|
||||
# 2. Wait for DB to be ready
|
||||
echo "⏳ Waiting for database to be ready..."
|
||||
until docker compose exec -T db pg_isready -U mapuser -d mapdb; do
|
||||
sleep 2
|
||||
done
|
||||
|
||||
# 3. Initialize Schema & Import Existing Server Data
|
||||
echo "🛠️ Initializing schema and loading server data..."
|
||||
docker compose exec -T db psql -U mapuser -d mapdb -c "
|
||||
DROP TABLE IF EXISTS places_syria;
|
||||
CREATE TABLE places_syria (
|
||||
id SERIAL PRIMARY KEY,
|
||||
latitude DECIMAL(10, 8),
|
||||
longitude DECIMAL(11, 8),
|
||||
name TEXT,
|
||||
name_ar TEXT,
|
||||
name_en TEXT,
|
||||
address TEXT,
|
||||
category TEXT,
|
||||
neighbourhood TEXT,
|
||||
city TEXT,
|
||||
description TEXT,
|
||||
created_at TIMESTAMP DEFAULT NOW(),
|
||||
source TEXT,
|
||||
location GEOMETRY(Point, 4326)
|
||||
);
|
||||
|
||||
CREATE OR REPLACE FUNCTION sync_place_location() RETURNS trigger AS \$\$
|
||||
BEGIN
|
||||
IF NEW.latitude IS NOT NULL AND NEW.longitude IS NOT NULL THEN
|
||||
NEW.location := ST_SetSRID(ST_MakePoint(CAST(NEW.longitude AS FLOAT), CAST(NEW.latitude AS FLOAT)), 4326);
|
||||
END IF;
|
||||
RETURN NEW;
|
||||
END;
|
||||
\$\$ LANGUAGE plpgsql;
|
||||
|
||||
DROP TRIGGER IF EXISTS trg_sync_place_location ON places_syria;
|
||||
CREATE TRIGGER trg_sync_place_location
|
||||
BEFORE INSERT OR UPDATE ON places_syria
|
||||
FOR EACH ROW EXECUTE FUNCTION sync_place_location();
|
||||
"
|
||||
|
||||
# Load existing data if available
|
||||
if [ -f "infrastructure/docker/postgis/existing_syria_data.sql" ]; then
|
||||
echo "📥 Loading existing server data..."
|
||||
docker compose exec -T db psql -U mapuser -d mapdb < infrastructure/docker/postgis/existing_syria_data.sql
|
||||
fi
|
||||
|
||||
# 4. Import New CSV Data
|
||||
echo "📥 Importing new CSV landmark data..."
|
||||
docker cp infrastructure/docker/postgis/syria_final_complete.csv map-db:/tmp/syria_data.csv
|
||||
|
||||
docker compose exec -T db psql -U mapuser -d mapdb -c "
|
||||
DROP TABLE IF EXISTS staging_syria;
|
||||
CREATE TABLE staging_syria (
|
||||
name TEXT,
|
||||
latitude DECIMAL(10, 8),
|
||||
longitude DECIMAL(11, 8),
|
||||
category TEXT,
|
||||
address TEXT,
|
||||
description TEXT,
|
||||
created_at TIMESTAMP
|
||||
);
|
||||
|
||||
-- HEADER true fixes the 'invalid input syntax' error
|
||||
COPY staging_syria(name, latitude, longitude, category, address, description, created_at)
|
||||
FROM '/tmp/syria_data.csv'
|
||||
WITH (FORMAT csv, HEADER true, QUOTE '\"', ENCODING 'UTF8');
|
||||
|
||||
-- Merge into places_syria while avoiding duplicates
|
||||
-- We check for name + spatial proximity (approx 50m)
|
||||
INSERT INTO places_syria (name, name_ar, latitude, longitude, category, address, description, source, created_at)
|
||||
SELECT
|
||||
name,
|
||||
name,
|
||||
latitude,
|
||||
longitude,
|
||||
category,
|
||||
address,
|
||||
description,
|
||||
'csv_import_2026_04',
|
||||
COALESCE(created_at, NOW())
|
||||
FROM staging_syria s
|
||||
WHERE NOT EXISTS (
|
||||
SELECT 1 FROM places_syria p
|
||||
WHERE (p.name = s.name OR p.name_ar = s.name)
|
||||
AND ST_DWithin(
|
||||
ST_SetSRID(ST_MakePoint(CAST(p.longitude AS FLOAT), CAST(p.latitude AS FLOAT)), 4326)::geography,
|
||||
ST_SetSRID(ST_MakePoint(CAST(s.longitude AS FLOAT), CAST(s.latitude AS FLOAT)), 4326)::geography,
|
||||
50
|
||||
)
|
||||
);
|
||||
|
||||
DROP TABLE staging_syria;
|
||||
"
|
||||
|
||||
echo "✅ Local database merge complete."
|
||||
echo "Run ./infrastructure/scripts/local_verify.sh to review the results."
|
||||
Executable
+37
@@ -0,0 +1,37 @@
|
||||
#!/bin/bash
|
||||
# Local Verification Script for Syria Landmarks
|
||||
# To be run on the Mac terminal
|
||||
|
||||
set -e
|
||||
|
||||
echo "📊 --- Local Syria Data Audit ---"
|
||||
|
||||
# 1. Total Count
|
||||
TOTAL=$(docker compose exec -T db psql -U mapuser -d mapdb -t -c "SELECT count(*) FROM places_syria;")
|
||||
echo "📍 Total Landmarks: $TOTAL"
|
||||
|
||||
# 2. Category Distribution
|
||||
echo "🗂️ Category Distribution:"
|
||||
docker compose exec -T db psql -U mapuser -d mapdb -c "
|
||||
SELECT category, count(*) as count
|
||||
FROM places_syria
|
||||
GROUP BY category
|
||||
ORDER BY count DESC
|
||||
LIMIT 10;
|
||||
"
|
||||
|
||||
# 3. Spatial Bounds Check (Damascus region)
|
||||
echo "🌍 Spatial Check (Damascus):"
|
||||
DM_COUNT=$(docker compose exec -T db psql -U mapuser -d mapdb -t -c "
|
||||
SELECT count(*) FROM places_syria
|
||||
WHERE latitude BETWEEN 33.4 AND 33.6 AND longitude BETWEEN 36.2 AND 36.4;
|
||||
")
|
||||
echo "🏙️ Landmarks in Damascus area: $DM_COUNT"
|
||||
|
||||
# 4. Generate Export if user is satisfied
|
||||
echo "💾 Generating SQL Export..."
|
||||
docker compose exec -T db pg_dump -U mapuser -d mapdb -t places_syria --data-only --inserts > infrastructure/docker/postgis/syria_export.sql
|
||||
|
||||
echo "--------------------------------"
|
||||
echo "✅ Verification complete. Export saved to: infrastructure/docker/postgis/syria_export.sql"
|
||||
echo "If you are happy with the results, run sync_to_server.sh to push the data."
|
||||
@@ -0,0 +1,24 @@
|
||||
#!/bin/bash
|
||||
# Server Restoration Script for Syria Landmarks
|
||||
# To be run on the server terminal
|
||||
|
||||
set -e
|
||||
|
||||
APP_DIR="/home/hamzadoctor/app"
|
||||
EXPORT_FILE="infrastructure/docker/postgis/syria_export.sql"
|
||||
|
||||
echo "🔄 Restoring Syria landmarks to production database..."
|
||||
|
||||
# 1. Clean existing manual imports to avoid PK conflicts
|
||||
echo "🧹 Clearing existing manual entries..."
|
||||
docker exec -i map-db psql -U mapuser -d mapdb -c "DELETE FROM places_syria WHERE source = 'manual_import_2026_04';"
|
||||
|
||||
# 2. Inject the SQL dump
|
||||
echo "📥 Injecting SQL dump..."
|
||||
docker exec -i map-db psql -U mapuser -d mapdb < "$APP_DIR/$EXPORT_FILE"
|
||||
|
||||
# 3. Refresh API cache
|
||||
echo "🧹 Flushing Redis..."
|
||||
docker exec -i map-redis redis-cli flushall
|
||||
|
||||
echo "✅ Restoration complete! Syria landmarks are live."
|
||||
Executable
+33
@@ -0,0 +1,33 @@
|
||||
#!/bin/bash
|
||||
# script to download Jordan OSM data
|
||||
# نص برمجبي لتحميل بيانات الخرائط للأردن
|
||||
|
||||
set -e
|
||||
|
||||
DATA_DIR="./infrastructure/osm-data"
|
||||
PBF_URL="https://download.geofabrik.de/asia/jordan-latest.osm.pbf"
|
||||
|
||||
echo "📍 Starting Jordan Map Setup..."
|
||||
echo "📍 البدء في إعداد خرائط الأردن..."
|
||||
|
||||
mkdir -p $DATA_DIR
|
||||
|
||||
if [ ! -f "$DATA_DIR/jordan-latest.osm.pbf" ]; then
|
||||
echo "📥 Downloading Jordan OSM PBF from Geofabrik..."
|
||||
echo "📥 جاري تحميل بيانات الخرائط من Geofabrik..."
|
||||
curl -L $PBF_URL -o "$DATA_DIR/jordan-latest.osm.pbf"
|
||||
else
|
||||
echo "✅ Jordan PBF already exists."
|
||||
echo "✅ ملف البيانات موجود بالفعل."
|
||||
fi
|
||||
|
||||
# Note on MBTiles:
|
||||
# Tileserver-GL requires an .mbtiles file to serve tiles locally.
|
||||
# You can generate one from the .pbf using 'tilemaker' or download a free extract
|
||||
# from MapTiler (https://maptiler.com/data) and place it in $DATA_DIR/jordan.mbtiles
|
||||
|
||||
echo "⚠️ Important: To serve tiles locally, you need 'jordan.mbtiles' in $DATA_DIR"
|
||||
echo "⚠️ تنبيه: لتشغيل الخرائط محلياً، يجب توفير ملف 'jordan.mbtiles' في المجلد المذكور"
|
||||
|
||||
echo "🚀 Setup complete. Please ensure jordan.mbtiles is present before running 'docker-compose up'."
|
||||
echo "🚀 اكتمل الإعداد. تأكد من وجود ملف mbtiles قبل تشغيل النظام."
|
||||
@@ -0,0 +1,60 @@
|
||||
#!/bin/bash
|
||||
# --------------------------------------------------------------------------
|
||||
# Intaleq Map Platform - 10-Day Update Script
|
||||
# سكربت تحديث خرائط انطلاقة - التحديث الدوري (كل 10 أيام)
|
||||
# --------------------------------------------------------------------------
|
||||
|
||||
set -e # Exit on error
|
||||
|
||||
echo "🚀 Starting 10-day map update..."
|
||||
|
||||
# 1. Configuration (From .env or defaults)
|
||||
PBF_FILE="/data/jordan-latest.osm.pbf"
|
||||
SOURCE_URL="https://download.geofabrik.de/asia/jordan-latest.osm.pbf"
|
||||
APP_DIR="/home/hamzadoctor/app"
|
||||
|
||||
cd "$APP_DIR"
|
||||
|
||||
# 2. Download latest OSM data (Running on Host)
|
||||
# تحميل أحدث البيانات للأردن وسوريا
|
||||
echo "🌍 Downloading latest OpenStreetMap data for Jordan & Syria..."
|
||||
wget -O "infrastructure/osm-data/jordan-latest.osm.pbf.new" "https://download.geofabrik.de/asia/jordan-latest.osm.pbf"
|
||||
wget -O "infrastructure/osm-data/syria-latest.osm.pbf.new" "https://download.geofabrik.de/asia/syria-latest.osm.pbf"
|
||||
|
||||
mv "infrastructure/osm-data/jordan-latest.osm.pbf.new" "infrastructure/osm-data/jordan-latest.osm.pbf"
|
||||
mv "infrastructure/osm-data/syria-latest.osm.pbf.new" "infrastructure/osm-data/syria-latest.osm.pbf"
|
||||
|
||||
# 3. Import new data to PostGIS
|
||||
# استيراد البيانات إلى قاعدة البيانات
|
||||
echo "💾 Importing Jordan (Create)..."
|
||||
docker compose --profile import run --rm osm-import osm2pgsql --create --slim --cache 1000 --database mapdb --host db --user mapuser /data/jordan-latest.osm.pbf
|
||||
|
||||
echo "💾 Importing Syria (Append)..."
|
||||
docker compose --profile import run --rm osm-import osm2pgsql --append --slim --cache 1000 --database mapdb --host db --user mapuser /data/syria-latest.osm.pbf
|
||||
|
||||
# 4. Merge Data (Jordan + Syria)
|
||||
OSM_FILE="infrastructure/osm-data/region.osm.pbf"
|
||||
echo "🗺️ Merging Jordan and Syria data into $OSM_FILE..."
|
||||
osmium merge infrastructure/osm-data/jordan-latest.osm.pbf infrastructure/osm-data/syria-latest.osm.pbf -o $OSM_FILE --overwrite
|
||||
|
||||
# 5. Spatial Integrity check for Geocoding (Landmarks)
|
||||
# Purge any legacy landmarks outside the expanded region
|
||||
echo "📍 Syncing user-submitted landmarks and purging invalid coordinates..."
|
||||
DELETED_COUNT=$(docker compose exec -T db psql -U mapuser -d mapdb -t -c "DELETE FROM places_syria WHERE longitude < 34 OR longitude > 43 OR latitude < 29 OR latitude > 38;")
|
||||
echo " ✅ Purged ${DELETED_COUNT//[[:space:]]/} invalid landmarks outside the expanded region."
|
||||
|
||||
# Force Spatial Geometry Update
|
||||
docker compose exec -T db psql -U mapuser -d mapdb -c "UPDATE places_syria SET location = ST_SetSRID(ST_MakePoint(longitude, latitude), 4326) WHERE location IS NULL OR latitude IS NOT NULL;"
|
||||
|
||||
# 6. Rebuild Routing Index (GraphHopper)
|
||||
echo "🚗 Rebuilding GraphHopper routing index (This may take ~5-8 minutes)..."
|
||||
docker compose stop routing
|
||||
# Delete old cache - VERY IMPORTANT to force full re-index
|
||||
rm -rf infrastructure/osm-data/graph-cache infrastructure/osm-data/default-gh
|
||||
docker compose up -d routing
|
||||
|
||||
# 7. Final Cleanup & Cache Flush
|
||||
echo "🧹 Clearing Redis traffic cache..."
|
||||
docker compose exec -T redis redis-cli flushall
|
||||
|
||||
echo "Done! Map platform is now fully synchronized with Jordan & Syria roads (including Damascus)."
|
||||
Reference in New Issue
Block a user