From 6e037a57288bce7e8f97a11fde9ac8c5bb00fbf4 Mon Sep 17 00:00:00 2001
From: Greg Wilson <gvwilson@third-bit.com>
Date: Sun, 14 Jun 2015 17:04:37 -0400
Subject: [PATCH] CSS cataloguing tool

---
 tools/catalog.py | 47 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 47 insertions(+)
 create mode 100644 tools/catalog.py

diff --git a/tools/catalog.py b/tools/catalog.py
new file mode 100644
index 0000000..4f449bb
--- /dev/null
+++ b/tools/catalog.py
@@ -0,0 +1,47 @@
+#!/usr/bin/env python
+
+'''Create YAML catalog of CSS styles using in a set of HTML documents.
+
+Usage: catalog.py file [file...]
+'''
+
+import sys
+import yaml
+from bs4 import BeautifulSoup
+
+def main(argv):
+    '''Main driver.'''
+
+    catalog = {}
+    for filename in argv[1:]:
+        with open(filename, 'r') as reader:
+            doc = BeautifulSoup(reader.read())
+            for node in doc.descendants:
+                update(catalog, node)
+    display(catalog)
+
+
+def update(catalog, node):
+    '''Record classes used in node.'''
+
+    if node.name is None:
+        return
+
+    if node.name not in catalog:
+        catalog[node.name] = set()
+
+    if 'class' in node.attrs:
+        for cls in node.attrs['class']:
+            catalog[node.name].add(cls)
+
+
+def display(catalog):
+    '''Show the catalog.'''
+
+    for name in sorted(catalog.keys()):
+        catalog[name] = sorted(catalog[name])
+    yaml.dump(catalog, stream=sys.stdout)
+
+
+if __name__ == '__main__':
+    main(sys.argv)