Changeset 158 for table_parser

Timestamp:

21/11/10 02:30:03 (18 months ago)

Author:

nigel.sim

Message:

added CLI usage

File:

• table_parser/table_parser.py (modified) (2 diffs)

table_parser/table_parser.py

@@ -51 +51 @@
     def __str__(self):
         return '<a href="%s">%s</a>'%(self.href, self.text)
+    def __repr__(self):
+        """Possible not the best repr"""
+        return '{"href": "%s", "text": "%s"}'%(self.href, self.text)
 
 # Get the item on the top of a stack
@@ -148 +151 @@
         self._buf += data
 
+
+if __name__ == '__main__':
+    import sys
+    if len(sys.argv) == 2 and sys.argv[1] == '-t':
+        # Run the doctests
+        import doctest
+        doctest.testmod()
+    elif len(sys.argv) == 2:
+        # Parse the URL provided
+        import pprint
+        import urllib2
+        f = urllib2.urlopen(sys.argv[1])
+        tp = TableParser()
+        tp.feed(f.read())
+        f.close()
+        pprint.pprint(tp.doc)
+    else:
+        # Print usage
+        print "run doctests: <script> -t"
+        print "parse URL: <script> <URL>"