check search pdfs

partitioncloud/modules/search.py

@@ -8,6 +8,7 @@ import threading
 import socket
 import os
 
+import pypdf
 import googlesearch
 
 from .db import get_db
@@ -52,12 +53,17 @@ def local_search(query, partitions):
 def download_search_result(element, instance_path):
     uuid = element["uuid"]
     url = element["url"]
+    filename = f"{instance_path}/search-partitions/{uuid}.pdf"
 
     try:
-        urllib.request.urlretrieve(url, f"{instance_path}/search-partitions/{uuid}.pdf")
+        urllib.request.urlretrieve(url, filename)
+        pypdf.PdfReader(filename)
 
-    except (urllib.error.HTTPError, urllib.error.URLError):
-        with open(f"{instance_path}/search-partitions/{uuid}.pdf", 'a', encoding="utf8") as _:
+    except (urllib.error.HTTPError, urllib.error.URLError,
+            pypdf.errors.PdfReadError, pypdf.errors.PdfStreamError):
+        if os.path.exists(filename):
+            os.remove(filename)
+        with open(filename, 'a', encoding="utf8") as _:
             pass # Create empty file
 
 

requirements.txt

@@ -2,4 +2,5 @@ flask
 flask-babel
 google
 colorama
+pypdf
 qrcode

scripts/hooks/v1.py

@@ -179,4 +179,12 @@ def move_thumbnails():
 
 
 def install_babel():
-    utils.install_package("flask-babel")
+    utils.install_package("flask-babel")
+
+
+"""
+ v1.8.*
+"""
+
+def install_pypdf():
+    utils.install_package("pypdf")

scripts/migration.py

@@ -35,7 +35,8 @@ hooks = [
     ("v1.4.1", [("Install qrcode", v1_hooks.install_qrcode)]),
     ("v1.5.0", [("Move to instance directory", v1_hooks.move_instance)]),
     ("v1.5.1", [("Move thumbnails", v1_hooks.move_thumbnails)]),
-    ("v1.7.0", [("Install babel", v1_hooks.install_babel)])
+    ("v1.7.0", [("Install babel", v1_hooks.install_babel)]),
+    ("v1.8.2", [("Install pypdf", v1_hooks.install_pypdf)])
 ]