* revised and reorganized dump scripts

3 files changed, 32 insertions, 0 deletions

diff --git a/dumps/_sites/__init__.py b/dumps/_sites/__init__.py
new file mode 100644
index 0000000..e69de29
--- /dev/null
+++ b/dumps/_sites/__init__.py
diff --git a/dumps/_sites/chicane-f1.py b/dumps/_sites/chicane-f1.py
new file mode 100644
index 0000000..07d90c9
--- /dev/null
+++ b/dumps/_sites/chicane-f1.py
@@ -0,0 +1,14 @@
+import urllib
+import urllib2
+import urlparse
+
+from lxml import html
+
+def fetch(url):
+    contents = urllib2.urlopen(url).read()
+    tree = html.fromstring(contents)
+    title = tree.xpath("//title")[0].text
+    tables = tree.xpath("//table[@cellpadding=6]")
+    print url
+    print title
+    return title, tables
diff --git a/dumps/_sites/second-a-lap.py b/dumps/_sites/second-a-lap.py
new file mode 100644
index 0000000..52591d6
--- /dev/null
+++ b/dumps/_sites/second-a-lap.py
@@ -0,0 +1,18 @@
+import json
+import urllib
+import urllib2
+import urlparse
+
+from lxml import html
+
+def fetch(url):
+    url = urlparse.urlparse(url).path
+    contents = json.loads(urllib2.urlopen('http://second-a-lap.blogspot.com/feeds/posts/default?' +
+                          urllib.urlencode({'alt': 'json', 'v': 2, 'dynamicviews': 1, 'path': url})).read())
+    title = contents['feed']['entry'][0]['title']['$t']
+    text = contents['feed']['entry'][0]['content']['$t']
+    tree = html.fromstring(text)
+    tables = tree.xpath("//table[@bordercolor]")
+    print url
+    print title
+    return title, tables