Пример #1
0
def download():
	setproperties.load_properties()

	for i in xrange(35):
		print "page %s" % i
		response = requests.get(sources[1] + "?page=%s" % i, verify=False)
		properties = get_url_title_and_description_from_html(response.text)
		
		for filename in properties:
			print filename
			filepath = os.path.join("data", "raw", filename)
			if not os.path.exists(filepath):
				with open(filepath, "wb") as datafile:
					r = requests.get(properties[filename]["url"])
					for chunk in r.iter_content(1024):
						datafile.write(chunk)
			
			setproperties.update_for_file(filename, properties[filename])
			
	setproperties.save_properties()
Пример #2
0
def get_args(form_dict):
	setproperties.load_properties()
	group_list = setproperties.groups.keys()
	group_list.sort()
	group_datasets = {}
	for group in group_list:
		group_datasets[group] = []
	
	for fname in setproperties.properties:
		p = setproperties.properties[fname]
		if not p.get("title"):
			p["title"] = fname
		p["file_name"] = fname
		if p.get("groups"):
			for group in p.get("groups"):
				group_datasets[group].append(p)
		else:
			group_datasets["Other"].append(p)
			
	return {
		"group_list": group_list,
		"groups": setproperties.groups,
		"group_datasets": group_datasets
	}