From 8221b81609e1c25b2f8931b6a534c3314abcdf14 Mon Sep 17 00:00:00 2001
From: Marcio Mazza <marciomazza@gmail.com>
Date: Mon, 15 Jun 2015 16:50:29 -0300
Subject: [PATCH] Add script to extract labels from sapl forms

---
 legacy/scripts/extract_labels.py | 28 ++++++++++++++++++++++++++++
 requirements.txt                 |  1 +
 2 files changed, 29 insertions(+)
 create mode 100644 legacy/scripts/extract_labels.py

diff --git a/legacy/scripts/extract_labels.py b/legacy/scripts/extract_labels.py
new file mode 100644
index 000000000..42baacf1c
--- /dev/null
+++ b/legacy/scripts/extract_labels.py
@@ -0,0 +1,28 @@
+import string
+from bs4 import BeautifulSoup
+from itertools import chain
+
+
+def _label_from_td(td):
+    return td.text.strip().split('\n')[0].strip(u'\xa0' + string.whitespace)
+
+
+def get_labels(filename, flat=True):
+    """Extract labels from a file containg the html source of a rendered
+    legacy sapl form
+    """
+    with open(filename, 'r') as f:
+        cont = f.read()
+    html_doc = cont.decode('utf-8')
+    soup = BeautifulSoup(html_doc, 'html.parser')
+    forms = soup.find_all('form')
+    [form] = [f for f in forms if (u'method', u'post') in f.attrs.items()]
+
+    labels = [[_label_from_td(td) for td in tr.find_all('td')] for tr in form.find_all('tr')]
+    for line in labels:
+        print ', '.join("u'%s'" % l for l in line)
+    if flat:
+        return list(chain(*labels))
+    else:
+        return labels
+
diff --git a/requirements.txt b/requirements.txt
index dcaea4c9f..6b9e3618c 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,6 +1,7 @@
 Django
 MySQL-python
 psycopg2
+beautifulsoup4
 
 ipdb
 django-extensions