From 65b9a0164409568e912517f7c35869423cf52240 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rapha=C3=ABl=20Vinot?= <raphael@vinot.info>
Date: Wed, 17 Sep 2014 17:19:03 +0200
Subject: [PATCH] Add config file for DomainClassifier, proper reporting

---
 bin/DomClassifier.py           | 16 ++++++++++++----
 bin/packages/config.cfg.sample |  4 ++++
 2 files changed, 16 insertions(+), 4 deletions(-)

diff --git a/bin/DomClassifier.py b/bin/DomClassifier.py
index c2b9f4ef..f1f59f72 100755
--- a/bin/DomClassifier.py
+++ b/bin/DomClassifier.py
@@ -28,6 +28,10 @@ def main():
     publisher.info("""ZMQ DomainClassifier is Running""")
 
     c = DomainClassifier.domainclassifier.Extract(rawtext="")
+
+    cc = p.config.get("DomClassifier", "cc")
+    cc_tld = p.config.get("DomClassifier", "cc_tld")
+
     while True:
         try:
             message = p.get_from_set()
@@ -44,12 +48,16 @@ def main():
                 c.text(rawtext=paste)
                 c.potentialdomain()
                 c.validdomain(rtype=['A'], extended=True)
-                localizeddomains = c.include(expression=r'\.lu$')
+                localizeddomains = c.include(expression=cc_tld)
                 if localizeddomains:
-                    print (localizeddomains)
-                localizeddomains = c.localizedomain(cc='LU')
+                    print(localizeddomains)
+                    publisher.warning('DomainC;{};{};{};Checked {} located in {}'.format(
+                        PST.p_source, PST.p_date, PST.p_name, localizeddomains, cc_tld))
+                localizeddomains = c.localizedomain(cc=cc)
                 if localizeddomains:
-                    print (localizeddomains)
+                    print(localizeddomains)
+                    publisher.warning('DomainC;{};{};{};Checked {} located in {}'.format(
+                        PST.p_source, PST.p_date, PST.p_name, localizeddomains, cc))
         except IOError:
             print "CRC Checksum Failed on :", PST.p_path
             publisher.error('Duplicate;{};{};{};CRC Checksum Failed'.format(
diff --git a/bin/packages/config.cfg.sample b/bin/packages/config.cfg.sample
index 6ddd5f20..ac8488f8 100644
--- a/bin/packages/config.cfg.sample
+++ b/bin/packages/config.cfg.sample
@@ -34,6 +34,10 @@ db = 1
 [Url]
 cc_critical = DE
 
+[DomClassifier]
+cc = DE
+cc_tld = r'\.de$'
+
 # Indexer configuration
 [Indexer]
 type = whoosh