tweak connection pooling for news

screenless/bureau/publications/publications.py

@@ -3,6 +3,7 @@ from base64 import b64encode
 from datetime import datetime
 import io
 import json
+import multiprocessing
 import os
 import random
 import string
@@ -313,9 +314,7 @@ class Publications(Bureau):
                 en_count += 1
 
         # do this multi-threaded cuz downloads can be slow
-        threads = []
-        for i in range(len(entries)):
-            entry = entries[i]
+        # NOTE: this could be further optimized with 2 threads per host (chunks)
         def fetch_og(entry):
             """
             get OpenGraph data for entry
@@ -337,16 +336,10 @@ class Publications(Bureau):
                 urllib.request.urlretrieve(entry.img, filename)
                 entry.img = "file://" + filename
 
+        fetcher = multiprocessing.pool.ThreadPool(processes=2)
+        entries_fetched = fetcher.map(fetch_og, entries)
 
-            thread = threading.Thread(target=fetch_og, args=(entry,))
-            threads.append(thread)
-            thread.start()
-
-        # wait till we're done
-        for thread in threads:
-            thread.join()
-
-        return entries
+        return entries_fetched
 
 
 def main():