first cut

mhausenblas · Jun 11, 2011 · fc74edf · fc74edf
1 parent e7627a2
commit fc74edf
Show file tree

Hide file tree

Showing 194 changed files with 38,243 additions and 0 deletions.
diff --git a/a404.html b/a404.html
@@ -0,0 +1,13 @@
+<!DOCTYPE HTML>
+<html>
+	<head>
+		<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+		<title>
+			omnidator
+		</title>
+	</head>
+	<body>
+		<div style="margin: 5em; text-align: center; font-size: 30pt">Uuuups, that a 404</div>
+		<div style="margin: 5em; text-align: center; font-size: 18pt">... sorry, seems we don't have what you're looking for ...</div>
+	</body>
+</html>
diff --git a/app.yaml b/app.yaml
@@ -0,0 +1,10 @@
+application: omnidator
+version: 1
+runtime: python
+api_version: 1
+
+handlers:
+- url: /examples
+  static_dir: examples
+- url: /.*
+  script: main.py
diff --git a/examples/solar-system.csv b/examples/solar-system.csv
@@ -0,0 +1,4 @@
+Thing,Mass,URL
+Earth,"5.9721986x10^24 kg",http://dbpedia.org/resource/Earth
+Mars,"6.4191x10^23 kg",http://dbpedia.org/resource/Mars
+Sun,"1.988435x10^30 kg",http://dbpedia.org/resource/Sun
diff --git a/handler.py b/handler.py
@@ -0,0 +1,53 @@
+"""
+The handler script.
+
+@author: Michael Hausenblas, http://sw-app.org/mic.xhtml#i
+@since: 2011-06-11
+@status: inital version
+"""
+import sys
+sys.path.insert(0, 'lib')
+import logging
+import cgi
+import os
+import platform
+import urllib
+import urllib2
+import StringIO
+import csv
+
+from google.appengine.ext import webapp
+from google.appengine.ext.webapp import template
+
+import rdflib
+
+from rdflib import Graph
+from rdflib import Namespace
+from rdflib import plugin
+from rdflib.serializer import Serializer
+
+import schema_org_processor
+
+
+class MainHandler(webapp.RequestHandler):
+	def get(self):
+		self.response.out.write(template.render('index.html', None))
+
+class NotFoundHandler(webapp.RequestHandler):
+	def get(self):
+		self.error(404)
+		self.response.out.write(template.render('a404.html', None))
+
+class SchemaOrgHandler(webapp.RequestHandler):
+	def get(self, doc_url):
+		logging.info('Trying to translate document at [%s]' %(urllib.unquote(doc_url)))
+
+		try:
+			sop = schema_org_processor.SchemaOrgProcessor()
+			sop.parse(urllib.unquote(doc_url))
+			self.response.headers.add_header("Access-Control-Allow-Origin", "*") # CORS-enabled
+			self.response.headers['Content-Type'] = 'application/rdf+xml'
+			self.response.out.write(str(sop.dump_data()))
+		except Exception, e:
+			self.error(404)
+			self.response.out.write(e)#template.render('a404.html', None))
diff --git a/index.html b/index.html
@@ -0,0 +1,13 @@
+<!DOCTYPE HTML>
+<html>
+	<head>
+		<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+		<title>
+			omnidator
+		</title>
+	</head>
+	<body>
+		<h1>API</h1>
+		<div>http://omnidator.appspot.com/2/{URL} ... from Schema.org data (microdata, CSV, etc.) to RDF</div>
+	</body>
+</html>
diff --git a/lib/html5lib/__init__.py b/lib/html5lib/__init__.py
@@ -0,0 +1,17 @@
+""" 
+HTML parsing library based on the WHATWG "HTML5"
+specification. The parser is designed to be compatible with existing
+HTML found in the wild and implements well-defined error recovery that
+is largely compatible with modern desktop web browsers.
+
+Example usage:
+
+import html5lib
+f = open("my_document.html")
+tree = html5lib.parse(f) 
+"""
+__version__ = "0.95-dev"
+from html5parser import HTMLParser, parse, parseFragment
+from treebuilders import getTreeBuilder
+from treewalkers import getTreeWalker
+from serializer import serialize
diff --git a/lib/html5lib/__init__.pyc b/lib/html5lib/__init__.pyc