Initial regex for ballot extractionHEADmaster

1 files changed, 17 insertions, 0 deletions

diff --git a/ballots.py b/ballots.py
new file mode 100644
index 0000000..d822d84
--- /dev/null
+++ b/ballots.py
@@ -0,0 +1,17 @@
+import re
+
+def read_file(filename):
+    with open(filename, 'r') as f:
+        return f.read()
+
+def get_ballots(f):
+    BALLOT_DATA_RE = re.compile(
+        r'(\d\d)\n(.*)\n(\d\d\d\d)\n(.*)\n([\d\.]+)\n(.*)\n(.*)')
+    m = re.findall(BALLOT_DATA_RE, f)
+    for x in m:
+        commitee_id, commitee, city_id, city, ballot_id, ballot_address, ballot_location = x
+        print commitee_id, commitee, city_id, city, ballot_id, ballot_address, ballot_location
+
+f = read_file('ballots.txt')
+data = get_ballots(f)
+print data