Automatically Creating Google Homework Sites

We had an interesting Google Sites idea for teaching get floated yesterday....

Q. Could you automatically create a Google Site for a list of students that only they and their tutor can see? Also, at a given time could the students' permission be changed from "writer" to "reader" when the deadline had arrived? It'd be good if you could use a Template Site so that the site could be set up with the right pages and prompts to begin with.

With the GData API it is possible. Here's an example python script that shows how... First, load the libraries. You may need to download these if you don't have them already.


import atom.data
import gdata.sites.client
import gdata.sites.data
from time import sleep


I then created lots of stub functions in the hope that I could better understand how the API works. The problem ( for me ) here was that ...

a. This code uses a now old fashioned means of connecting to the API. If this was going to be "real code" we would have to use OAuth to connect to the API.
b. The API documentation is hopeless
c. It seems to work around the idea of (Atom) feeds with have <entry> objects, which have <other objects> in them, which you get to via various URLs, which just makes the whole process more cumbersome... anyway...

The first thing you do is connect to the API


client = gdata.sites.client.SitesClient(source='University of York', site='UoY') # Can be anything, I think
client.domain = "york.ac.uk" #Not needed for non Apps domains
client.ClientLogin('****@york.ac.uk', '*********', client.source)


... then I wanted to see a list of all my sites and also show the URLs I'd need to use for each.

def list_sites( ):

'''

Shows a list of URLs you will need

'''

feed = client.GetSiteFeed()

for entry in feed.entry:

print '%s (%s)' % (entry.title.text, entry.site_name.text)

if entry.summary.text:

print 'description: ' + entry.summary.text

print "\turl:", entry.find_self_link()

print "\tedit link", entry.find_edit_link()

print "\tacl_link:", entry.get_acl_link().href

print "\tclickable link:", entry.get_alternate_link().href

print "_" * 80

... this produces something like this...

Code Example: Timed Content Creation (code-experiments)

description: This site is an example that uses an App Script to grab some data from the web and save it in a Site's pages

url: https://sites.google.com/feeds/site/york.ac.uk/code-experiments

edit link https://sites.google.com/feeds/site/york.ac.uk/code-experiments

acl_link: https://sites.google.com/feeds/acl/site/york.ac.uk/code-experiments

clickable link: https://sites.google.com/a/york.ac.uk/code-experiments/

________________________________________________________________________________

Collaborative Tools Project (collaborative-tools-project)

url: https://sites.google.com/feeds/site/york.ac.uk/collaborative-tools-project

edit link https://sites.google.com/feeds/site/york.ac.uk/collaborative-tools-project

acl_link: https://sites.google.com/feeds/acl/site/york.ac.uk/collaborative-tools-project

clickable link: https://sites.google.com/a/york.ac.uk/collaborative-tools-project/

________________________________________________________________________________

Collaboratomatic (collaboratomatic)

url: https://sites.google.com/feeds/site/york.ac.uk/collaboratomatic

edit link https://sites.google.com/feeds/site/york.ac.uk/collaboratomatic

acl_link: https://sites.google.com/feeds/acl/site/york.ac.uk/collaboratomatic

clickable link: https://sites.google.com/a/york.ac.uk/collaboratomatic/

________________________________________________________________________________

Departmental 20:20s (departmental-20-20s-july-2011)

description: A place to collect resources from the Information Directorate 20:20 presentations

url: https://sites.google.com/feeds/site/york.ac.uk/departmental-20-20s-july-2011

edit link https://sites.google.com/feeds/site/york.ac.uk/departmental-20-20s-july-2011

acl_link: https://sites.google.com/feeds/acl/site/york.ac.uk/departmental-20-20s-july-2011

clickable link: https://sites.google.com/a/york.ac.uk/departmental-20-20s-july-2011/

With this list of URLs I can then either access a site for editing, or for changing its permissions ( or ACLs) like this....




def get_site(site_feed_url):
'returns a GDEntry site object that you can fiddle with'
site = client.GetEntry(site_feed_url)
print site.title.text
print "\tedit_link:", site.get_edit_link().href
print "\tacl_link:", site.find_acl_link()
return site



def get_site_by_name(name):
feed = client.GetSiteFeed()

for entry in feed.entry:
if entry.title.text == name:
print "url:", entry.find_self_link()
print "edit link", entry.find_edit_link()
print "acl_link:", entry.get_acl_link().href
return entry
break

>>> site = get_site_by_name("My Template Site")

url: https://sites.google.com/feeds/site/york.ac.uk/my-template-site

edit link https://sites.google.com/feeds/site/york.ac.uk/my-template-site

acl_link: https://sites.google.com/feeds/acl/site/york.ac.uk/my-template-site

<gdata.sites.data.SiteEntry object at 0x10143c1d0>

I can then nosey around in my site to see what it can do...

>>> dir( site )

['FindAclLink', 'FindAlternateLink', 'FindChildren', 'FindEditLink', 'FindEditMediaLink', 'FindExtensions', 'FindFeedLink', 'FindHtmlLink', 'FindLicenseLink', 'FindMediaLink', 'FindNextLink', 'FindPostLink', 'FindPreviousLink', 'FindSelfLink', 'FindSourceLink', 'FindUrl', 'GetAclLink', 'GetAlternateLink', 'GetAttributes', 'GetEditLink', 'GetEditMediaLink', 'GetElements', 'GetFeedLink', 'GetHtmlLink', 'GetId', 'GetLicenseLink', 'GetLink', 'GetNextLink', 'GetPostLink', 'GetPreviousLink', 'GetSelfLink', 'IsMedia', 'ToString', '_XmlElement__get_extension_attributes', '_XmlElement__get_extension_elements', '_XmlElement__set_extension_attributes', '_XmlElement__set_extension_elements', '__class__', '__delattr__', '__dict__', '__doc__', '__format__', '__getattribute__', '__hash__', '__init__', '__module__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', '_attach_members', '_become_child', '_get_namespace', '_get_rules', '_get_tag', '_harvest_tree', '_list_xml_members', '_members', '_other_attributes', '_other_elements', '_qname', '_rule_set', '_set_namespace', '_set_tag', '_to_tree', 'attributes', 'author', 'category', 'children', 'content', 'contributor', 'control', 'etag', 'extension_attributes', 'extension_elements', 'find_acl_link', 'find_alternate_link', 'find_edit_link', 'find_edit_media_link', 'find_feed_link', 'find_html_link', 'find_license_link', 'find_media_link', 'find_next_link', 'find_post_link', 'find_previous_link', 'find_self_link', 'find_source_link', 'find_url', 'get_acl_link', 'get_alternate_link', 'get_attributes', 'get_edit_link', 'get_edit_media_link', 'get_elements', 'get_feed_link', 'get_html_link', 'get_id', 'get_license_link', 'get_link', 'get_next_link', 'get_post_link', 'get_previous_link', 'get_self_link', 'id', 'is_media', 'link', 'namespace', 'published', 'rights', 'site_name', 'source', 'summary', 'tag', 'text', 'theme', 'title', 'to_string', 'updated']

From here I created a few more functions, mainly just trying to understand how the API works.

def create_site(title, description='', theme='slate', source_site_url=None):

if source_site_url:

#template_site = get_site(source_site_url)

#source_url = template_site.GetId()

print "copying:", source_site_url

site = client.CreateSite(title, description=description,theme=theme, source_site=source_site_url)

else:

#Create a blank site

site = client.CreateSite(title, description=description, theme=theme)

print "site created!", site.find_edit_link()

return site

def get_acls(site_acl_url):

'''

Get the site's permissions as a feed if given an acl_link, like this

'https://sites.google.com/feeds/acl/site/york.ac.uk/my-lovely-site'

'''

if "/acl/" not in site_acl_url:

print "That's probably not the right URL!"

feed = client.GetAclFeed(site_acl_url)

for entry in feed.entry:

try:

print '%s (%s) - %s' % (entry.scope.value, entry.scope.type, entry.role.value)

except Exception, err:

#print err

pass

return feed

def share_a_site_with(acl_url, email, role='writer'):

'''

Needs an acl_link to work

role can be reader, writer or owner

'''

#feed = client.GetAclFeed(acl_url)

scope = gdata.acl.data.AclScope(value=email, type='user')

role = gdata.acl.data.AclRole(value=role)

acl = gdata.acl.data.AclEntry(scope=scope, role=role)

#post(self, entry, uri, auth_token=None, converter=None, desired_class=None, **kwargs) method of gdata.sites.client.SitesClient instance

#fudgetastic!

acl_url = acl_url.replace('https://sites.google.com', '') #!!!! ??? 

acl_entry = client.Post(acl, acl_url)

print "%s %s added as a %s" % (acl_entry.scope.type, acl_entry.scope.value, acl_entry.role.value)

def list_a_sites_acls(site_url):

feed = client.GetAclFeed(site_url)

for entry in feed.entry:

  print '%s (%s) - %s' % (entry.scope.value, entry.scope.type, entry.role.value)

return feed

def list_all_copies(site_url):

''' 

Unfinished: You will need to be able to get all copies of a template site. You can iterate

through all sites and find the sites whose source is a another_site's URL.

'''

site = get_site(site_url)

#get all my sites

#for each, check site.FindSourceLink() to see if it is site_url

def remove_a_user(site_url, email):

"Shows how to remove a users access to a site"

feed = client.GetAclFeed(site_url)

for entry in feed.entry:

print entry.scope.value

if entry.scope.value == email:

print "DELETING ACCESS FOR: ", email

client.Delete(entry.GetEditLink().href, force=True)

break

print "done!"

And Finally....

I made a test function to get my Template Google Site and make copies ( each with their own name ) and allow "writer" access to them based on a list of emails.

def do_test():

template_site_url = 'https://sites.google.com/feeds/site/york.ac.uk/my-template-site'

#This might be read in from a spreadsheet.

students_emails = ['student1@york.ac.uk',

'student2@york.ac.uk', 

'student3@york.ac.uk',

'student4@york.ac.uk',]

for i,student_email in enumerate(students_emails):

#Create a copy of the template site with a new name

client = gdata.sites.client.SitesClient(source='University of York', site='UoY')

client.domain = "york.ac.uk"

client.ClientLogin('****@york.ac.uk', '********', client.source)

new_site_name = "Auntie Homework: %s" % student_email

print "creating:", new_site_name

site = create_site(new_site_name, source_site_url=template_site_url)

sleep(5)

#Get its edit_url

new_acl_url = site.get_acl_link().href

print "sharing:", new_acl_url

#Now add that user as a writer to the new site's permissions

share_a_site_with(new_acl_url, student_email, role="writer")

sleep(5)

print "Done creating", i+1, "Google Sites"

Conclusions

It worked! It made 4 new Google Sites based on my template site, and each was shared with one other person with "writer" access.

But there's still a heap of work to make this into something genuinely useful.

• Because the client.ClientLogin() method, used to connect to API is deprecated, we will have to re-work this bit, and create an online user interface for it. I think greating a GUI in Google Spreadsheets might be the easiest but my JavaScript aint that hot.
• Although each copied site keeps a track of the site it was copied from, there's no way of running this script for two separate homework assignments for separate lists of students AND have it keep track of which was in which cohort. 
• I didn't look at "group permissions" which might be handy. When the homework deadline is passed a student group, rather the writer of the site, might be given read-only access. This would mean that, once finished everyone else can see everyone else's sites. 
• I found that without the sleep() and re-initialising the client instance each time I wanted to create a new site, that the do_test() function threw a wobbly. 
• A site can't be created if one with that name/URL already exists. I'd need to create a function to clear all copies of a template site.
• Potentially, I think we may be able to either subscribe to the newly created sites, or do something with the newly created sites' content feed, so that as a tutor you'd be able to keep an eye on who was or wasn't completing their homework in one place. I didn't get that far though. At very least, the script might create a page on the Template Site that listed the copies that had been made.