From 06814ecae8937310ee32ddb2aed144fef6c7c542 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Arne=20Schl=C3=BCter?= <aesthaddicts@gmail.com>
Date: Sat, 17 Jan 2015 11:45:52 +0100
Subject: [PATCH] Start writing some analyzaton code

---
 analyze.py | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)
 create mode 100644 analyze.py

diff --git a/analyze.py b/analyze.py
new file mode 100644
index 0000000..04e45aa
--- /dev/null
+++ b/analyze.py
@@ -0,0 +1,25 @@
+def get_district(article_headline):
+    """
+    Returns a geo-coded version of a district an article is about, based on its
+    headline.
+    """
+    pass
+
+def get_categories(article_body):
+    """
+    Gives a list of categories an article falls into, which is empty if none of
+    the following are matched:
+    - sexism
+    - antisemitism
+    - homophobia
+    - racism
+    """
+    bad_words = {
+        'antisemit': 'antisemitism',
+        'homophob': 'homophobia',
+        'sexis': 'sexism',
+        'rassis': 'racism'
+    }
+    found_categories = [bad_words[key] for key in bad_words
+                                        if key in article_body.lower()]
+    return found_categories or ['other']