From 025a723a5a201009da392bca4c27c4eb25e9e734 Mon Sep 17 00:00:00 2001
From: ryuslash
Date: Sun, 17 Jan 2010 03:48:23 +0100
Subject: Parser seperation, progress bar

* Seperated the parser from the downloader code.
* Added a progressbar class, to make it look fancier
* Created some functions to do all the work in downloader.py, cleaner now
* Changed parser.py to htmlparser.py, since it was conflicting with a built-in module
---
 htmlparser.py | 30 ++++++++++++++++++++++++++++++
 1 file changed, 30 insertions(+)
 create mode 100644 htmlparser.py

(limited to 'htmlparser.py')

diff --git a/htmlparser.py b/htmlparser.py
new file mode 100644
index 0000000..73338dd
--- /dev/null
+++ b/htmlparser.py
@@ -0,0 +1,30 @@
+import sgmllib
+import re
+
+class MyParser(sgmllib.SGMLParser):
+    def __init__(self, verbose=0):
+        sgmllib.SGMLParser.__init__(self, verbose)
+        
+        self.hyperlinks = []
+        self.url_reg = re.compile('res/\d+\Z')
+        self.prev = ""
+        
+    def parse(self, s):
+        self.feed(s)
+        self.close()
+
+    def start_a(self, attributes):
+        for name, value in attributes:
+            if name == "href":
+                if self.url_reg.search(value) != None:
+                    if self.prev != value:
+                        self.hyperlinks.append(value)
+                        self.prev = value
+    
+    def get_hyperlinks(self):
+        return self.hyperlinks
+
+class MySubParser(MyParser):
+    def __init__(self, verbose=0):
+        MyParser.__init__(self, verbose)
+        self.url_reg = re.compile('/src/\d+\.\w{3,4}\Z')
-- 
cgit v1.2.3-54-g00ecf