From 43b3556588d782f2826c20b0956190007e87e113 Mon Sep 17 00:00:00 2001
From: Alain <fou.alain@gmail.com>
Date: Fri, 5 Feb 2016 13:58:21 -0500
Subject: [PATCH 1/3] Starting Phone number recognition

---
 bin/Phone.py | 56 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 56 insertions(+)
 create mode 100644 bin/Phone.py

diff --git a/bin/Phone.py b/bin/Phone.py
new file mode 100644
index 00000000..87caf772
--- /dev/null
+++ b/bin/Phone.py
@@ -0,0 +1,56 @@
+#!/usr/bin/env python2
+# -*-coding:UTF-8 -*
+"""
+    module for finding phone numbers
+"""
+
+import time
+import pprint
+import re
+from packages import Paste
+from packages import lib_refine
+from pubsublogger import publisher
+from Helper import Process
+
+
+def search_phone(message):
+    paste = Paste.Paste(message)
+    content = paste.get_p_content()
+    # regex to find phone numbers, may raise many false positives (shalt thou seek optimization, upgrading is required)
+    reg_phone = re.compile(r'(\+\d{1,3}\(\d{1,2}\)\d?)?(\d{2,4}[\W\D\s]?){4,6} ')
+    # list of the regex results in the Paste, may be null
+    results = reg_phone.findall(content)
+
+    # if the list is greater than 4, we consider the Paste may contain a list of phone numbers
+    if len(results) > 4 :
+        print results
+        publisher.warning('{} contains PID (phone numbers)'.format(paste.p_name))
+
+	if __name__ == '__main__':
+    # If you wish to use an other port of channel, do not forget to run a subscriber accordingly (see launch_logs.sh)
+    # Port of the redis instance used by pubsublogger
+    publisher.port = 6380
+    # Script is the default channel used for the modules.
+    publisher.channel = 'Script'
+
+    # Section name in bin/packages/modules.cfg
+    config_section = 'Phone'
+
+    # Setup the I/O queues
+    p = Process(config_section)
+
+    # Sent to the logging a description of the module
+    publisher.info("Run Phone module")
+
+    # Endless loop getting messages from the input queue
+    while True:
+        # Get one message from the input queue
+        message = p.get_from_set()
+        if message is None:
+            publisher.debug("{} queue is empty, waiting".format(config_section))
+            time.sleep(1)
+            continue
+
+        # Do something with the message from the queue
+        search_phone(message)
+        

From fabbfd8ae9a606ce3fe13949cf066a2129880470 Mon Sep 17 00:00:00 2001
From: Alain <fou.alain@gmail.com>
Date: Fri, 5 Feb 2016 14:00:41 -0500
Subject: [PATCH 2/3] Update module.cfg (adding Keys and Phone section)

---
 bin/packages/modules.cfg | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/bin/packages/modules.cfg b/bin/packages/modules.cfg
index e408b84f..1638f7be 100644
--- a/bin/packages/modules.cfg
+++ b/bin/packages/modules.cfg
@@ -46,3 +46,9 @@ subscribe = Redis_ValidOnion
 [Web]
 subscribe = Redis_Web
 publish = Redis_Url,ZMQ_Url
+
+[Keys]
+subscribe = Redis_Global
+
+[Phone]
+subscribe = Redis_Global

From ea52fd106818f28a1954c6bcc2b9b5794c616ea1 Mon Sep 17 00:00:00 2001
From: Alain <fou.alain@gmail.com>
Date: Fri, 5 Feb 2016 20:58:02 +0100
Subject: [PATCH 3/3] Phone regex updated

Might still need to be fixed / optimized, in case of maths or random numbers starting with a 0. Do not capture dates, hours, coordinates anymore. Captured formats are: e.g. +331234567890 ; 09 12 34 56 78 ; +4177/123.45.69 ; +352(0)6-23-23-23...
---
 bin/Phone.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/bin/Phone.py b/bin/Phone.py
index 87caf772..384040cf 100644
--- a/bin/Phone.py
+++ b/bin/Phone.py
@@ -17,7 +17,7 @@ def search_phone(message):
     paste = Paste.Paste(message)
     content = paste.get_p_content()
     # regex to find phone numbers, may raise many false positives (shalt thou seek optimization, upgrading is required)
-    reg_phone = re.compile(r'(\+\d{1,3}\(\d{1,2}\)\d?)?(\d{2,4}[\W\D\s]?){4,6} ')
+    reg_phone = re.compile(r'(\+\d{1,4}(\(\d\))?\d?|0\d?)(\d{6,8}|([-/\. ]{1}\d{2,3}){3,4})')
     # list of the regex results in the Paste, may be null
     results = reg_phone.findall(content)