Consuming Learning Registry Records

damonregan edited this page Aug 22, 2013 · 2 revisions

The Learning Registry provides RESTful services to access metadata and paradata about educational resources. The guidance below is intended to help you write software that will use the Learning Registry’s services to consume metadata and paradata. The guidance is broken down into two steps. The first step is harvesting records from the Learning Registry. The second step is parsing the harvested records. Each step provides python code examples. The full python code is available from:

1. Harvesting Records

The harvest service is the recommended service to use when getting started (other services that can be used are obtain and slice). The harvest listrecords verb returns (in batches of 100 documents) records within an optional specific time/date range. Both GET and POST methods are supported. The GET method is shown below.

GET <node url>/harvest/listrecords[?from=<date>&until=<date>&resumption_token=<token>]


NOTE: If clicking the example in a browser, an extension like jsonview is recommended.

Results are returned as an array of JSON. The main elements of the result object are:

  • “OK”: boolean // T if successful
  • “error”: “string” // only present if NOT OK describing error
  • “responseDate”: “string” // timestamp
  • “listrecords”: [] // array of records resulting from request
  • “request”: {} // object containing the original request
  • “resumption_token”: “string” // flow control resumption token, NULL if end

Below is a python code example that harvests records using resumption tokens:

def harvest(start_url):
    #start by adding the root URL to the list of urls to harvest from
    urls = [start_url]
    #while we have url to harvest from continue
    while len(urls) > 0:        
        #remove the first URL to pull the LR documents from
        lr_url = urls.pop()
        # make an HTTP GET request to the LR harvest interface
        resp = urllib2.urlopen(lr_url)
            #parse json from the response body
            data = json.loads(
            # iterate over the results
            for i in data['listrecords']:
                #for the rest of this code we only care about the LR envelope portion of the harvest result
                envelope = i['record']['resource_data']
                # process the envelope
                # if there is a resumption token
                if "resumption_token" in data and \
                        data['resumption_token'] is not None and \
                        data['resumption_token'] != "null":
                    #parse the origional URL and update the query string to contain the resumption_token
                    url_parts = urlparse.urlparse(lr_url)
                    new_query = urllib.urlencode({"resumption_token": data['resumption_token']})
                    next_url = urlparse.urlunparse((url_parts.scheme,
                    #add the URL for the next page of results to the urls array
        except Exception as ex:

2. Parsing Records

Each record in the Learning Registry is based on the resource data description data model. The primary elements in the data model are:

  • “resource_data”: “string” or JSON object // describes the resource itself (metadata or paradata).
  • “resource_locator”: “string” // URL of the resource

The “resource_data” comes in different forms that are described by the following elements of the resource data description data model:

  • “payload_schema_format”: “string” // schema MIME type
  • “resource_data_type”: “string” // vocabulary of types [“metadata”, “paradata”, …]

While there are many payload_schema_formats that could be encountered, it is important to be familiar with the common data schema formats that are found in the Learning Registry.

Below is a python code example that parses records based on the common data schema formats specified in payload_schema_format:

def process_envelope(envelope):    
    #normalize casing on all the schemas in the payload_schema array, if payload_schema isn't present use an empty array
    schemas = {schema.lower() for schema in envelope.get('payload_schema', [])}
        if 'lom' in schemas:
        elif 'nsdl_dc' in schemas:
        elif 'lrmi' in schemas:
        elif 'comm_para 1.0' in schemas:
    except Exception as ex:
        print("Error In Payload")