From 0c6b09f379ec2a6e7d3332874904220d1080d514 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rapha=C3=ABl=20Vinot?= <raphael@vinot.info>
Date: Mon, 1 Sep 2014 16:18:06 +0200
Subject: [PATCH] Fix the onion module, log the valid onions.

---
 bin/DumpValidOnion.py    | 35 +++++++++++++++++++++++++++++++++++
 bin/Onion.py             | 21 ++++++++++++++++++---
 bin/packages/modules.cfg |  5 ++++-
 3 files changed, 57 insertions(+), 4 deletions(-)
 create mode 100755 bin/DumpValidOnion.py

diff --git a/bin/DumpValidOnion.py b/bin/DumpValidOnion.py
new file mode 100755
index 00000000..dad5ea9b
--- /dev/null
+++ b/bin/DumpValidOnion.py
@@ -0,0 +1,35 @@
+#!/usr/bin/env python2
+# -*-coding:UTF-8 -*
+
+from pubsublogger import publisher
+from Helper import Process
+import datetime
+import time
+
+if __name__ == "__main__":
+    publisher.port = 6380
+    publisher.channel = "Script"
+
+    config_section = 'DumpValidOnion'
+    dump_file = 'dump.out'
+
+    p = Process(config_section)
+
+    # FUNCTIONS #
+    publisher.info("Script subscribed to channel ValidOnion")
+
+    while True:
+        message = p.get_from_set()
+        if message is not None:
+            f = open(dump_file, 'a')
+            while message is not None:
+                print message
+                date = datetime.datetime.now()
+                if message is not None:
+                    f.write(date.isoformat() + ' ' + message + '\n')
+                else:
+                    break
+                message = p.get_from_set()
+            f.close()
+        else:
+            time.sleep(20)
diff --git a/bin/Onion.py b/bin/Onion.py
index db814631..525507e8 100755
--- a/bin/Onion.py
+++ b/bin/Onion.py
@@ -29,12 +29,15 @@ import datetime
 import os
 import base64
 import subprocess
+import redis
 
 from Helper import Process
 
 
-def fetch(p, urls, domains, path):
+def fetch(p, r_cache, urls, domains, path):
     for url, domain in zip(urls, domains):
+        if r_cache.exists(url):
+            continue
         to_fetch = base64.standard_b64encode(url)
         process = subprocess.Popen(["python", './tor_fetcher.py', to_fetch],
                                    stdout=subprocess.PIPE)
@@ -42,6 +45,8 @@ def fetch(p, urls, domains, path):
             time.sleep(1)
 
         if process.returncode == 0:
+            r_cache.setbit(url, 0, 1)
+            r_cache.expire(url, 3600)
             tempfile = process.stdout.read().strip()
             with open(tempfile, 'r') as f:
                 filename = path + domain
@@ -54,7 +59,9 @@ def fetch(p, urls, domains, path):
                     os.makedirs(dirname)
                 with open(save_path, 'w') as ff:
                     ff.write(content)
-                p.populate_set_out(save_path)
+                p.populate_set_out(save_path, 'Global')
+                p.populate_set_out(url, 'ValidOnion')
+                yield url
             os.unlink(tempfile)
         else:
             print 'Failed at downloading', url
@@ -71,6 +78,10 @@ if __name__ == "__main__":
     config_section = 'Onion'
 
     p = Process(config_section)
+    r_cache = redis.StrictRedis(
+        host=p.config.get("Redis_Cache", "host"),
+        port=p.config.getint("Redis_Cache", "port"),
+        db=p.config.getint("Redis_Cache", "db"))
 
     # FUNCTIONS #
     publisher.info("Script subscribed to channel onion_categ")
@@ -121,7 +132,11 @@ if __name__ == "__main__":
                                         str(now.month).zfill(2),
                                         str(now.day).zfill(2),
                                         str(int(time.mktime(now.utctimetuple()))))
-                    fetch(p, urls, domains_list, path)
+                    to_print = 'Onion;{};{};{};'.format(PST.p_source,
+                                                        PST.p_date,
+                                                        PST.p_name)
+                    for url in fetch(p, r_cache, urls, domains_list, path):
+                        publisher.warning('{}Valid: {}'.format(to_print, url))
                 else:
                     publisher.info('{}Onion related'.format(to_print))
 
diff --git a/bin/packages/modules.cfg b/bin/packages/modules.cfg
index e6241f40..97ed166b 100644
--- a/bin/packages/modules.cfg
+++ b/bin/packages/modules.cfg
@@ -34,7 +34,10 @@ subscribe = Redis_Mail
 
 [Onion]
 subscribe = Redis_Onion
-publish = Redis_Global
+publish = Redis_Global,Redis_ValidOnion
+
+[DumpValidOnion]
+subscribe = Redis_ValidOnion
 
 [Web]
 subscribe = Redis_Web