Search code examples
freebasemql

Freebase not returning all available values

I'm trying to extract the link to the NYTimes Topic page (among the topic_equivalent_webpage values) for Barack Obama from Freebase, but my query doesn't return any results, although it's on the webpage (http://www.freebase.com/m/02mjmr). This is my query:

[{
 "id": "/en/barack_obama",
 "type": "/common/topic",
 "topic_equivalent_webpage": {
   "value": null,
   "value~=": "*nytimes*"
 }
}]

I've also tried extracting all of the topic_equivalent_webpage values instead, using this query:

[{
"id": "/en/barack_obama",
"type": "/common/topic",
"topic_equivalent_webpage": []
}]

For some reason it only returns one of the values (http://www.worldcat.org/wcidentities/lccn-n94-112934).

Does anyone have any tips?

Solution

  • NOTE: All Freebase APIs are going away in a few months.

    You have three choices:

    1. Download the RDF dump and filter it. This is most appropriate for a large-scale download instead of using the API. For the property name and decoding process, see #3.

    2. Use the Topic API i.e. https://www.googleapis.com/freebase/v1/topic/en/barack_obama?filter=/common/topic/topic_equivalent_webpage

    3. Query MQL for the keys in the namespace that you want (ie the NY Times namespace)

    [{ "id": "/en/barack_obama", "key": [{ "namespace": "/source/nytimes", "value": null }] }]

    Normally the result is an identifier that gets substituted into a URI template, but in the NYT case it's basically a full URI path that just gets appended to http://nytimes.com/

    The key value (e.g. top$002Freference$002Ftimestopics$002Fpeople$002Fo$002Fbarack_obama) will be MQL key encoded, so they'll need to be decoded, but in this case you can probably cheat and replace all "$002F" substrings with "/". If any other characters are encoded, just replace $dddd with the character that has that Unicode code point.