Spaces:
Sleeping
Sleeping
Commit ·
6f24875
1
Parent(s): 2706084
Fix search endpoint file permissions
Browse files- app/main.py +1 -2
- test_search.py +31 -0
app/main.py
CHANGED
|
@@ -352,6 +352,5 @@ def web_search(q: str, max_results: int = 8):
|
|
| 352 |
return results
|
| 353 |
except Exception as e:
|
| 354 |
import traceback
|
| 355 |
-
|
| 356 |
-
f.write(traceback.format_exc())
|
| 357 |
raise HTTPException(status_code=500, detail=f"حدث خطأ أثناء البحث: {str(e)}")
|
|
|
|
| 352 |
return results
|
| 353 |
except Exception as e:
|
| 354 |
import traceback
|
| 355 |
+
traceback.print_exc()
|
|
|
|
| 356 |
raise HTTPException(status_code=500, detail=f"حدث خطأ أثناء البحث: {str(e)}")
|
test_search.py
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import urllib.request
|
| 2 |
+
import urllib.parse
|
| 3 |
+
import json
|
| 4 |
+
import re
|
| 5 |
+
|
| 6 |
+
q = "محمد نجيب"
|
| 7 |
+
max_results = 9
|
| 8 |
+
|
| 9 |
+
is_arabic = bool(re.search(r'[\u0600-\u06FF]', q))
|
| 10 |
+
lang = "ar" if is_arabic else "en"
|
| 11 |
+
|
| 12 |
+
url = f"https://{lang}.wikipedia.org/w/api.php?action=query&list=search&srsearch={urllib.parse.quote(q)}&utf8=&format=json&srlimit={max_results}"
|
| 13 |
+
req = urllib.request.Request(url, headers={'User-Agent': 'SmartReader/1.0'})
|
| 14 |
+
|
| 15 |
+
with urllib.request.urlopen(req, timeout=10) as response:
|
| 16 |
+
data = json.loads(response.read().decode('utf-8'))
|
| 17 |
+
|
| 18 |
+
results = []
|
| 19 |
+
for item in data.get('query', {}).get('search', []):
|
| 20 |
+
title = item.get('title', '')
|
| 21 |
+
snippet_html = item.get('snippet', '')
|
| 22 |
+
snippet = re.sub(r'<[^>]+>', '', snippet_html)
|
| 23 |
+
article_url = f"https://{lang}.wikipedia.org/wiki/{urllib.parse.quote(title.replace(' ', '_'))}"
|
| 24 |
+
results.append({
|
| 25 |
+
"title": title,
|
| 26 |
+
"url": article_url,
|
| 27 |
+
"snippet": snippet,
|
| 28 |
+
"source": "ويكيبيديا" if is_arabic else "Wikipedia"
|
| 29 |
+
})
|
| 30 |
+
|
| 31 |
+
print(json.dumps(results, indent=2, ensure_ascii=False))
|