/event/{event_id}
This page is the bonus dev page for getting the graph data
A good example doc to look at with all the changes is: http://nsight-api-1.rr.ikanow.com/event/53680852e4b08812fcb36d61?graph_data=true&fake_geo=true
To use this call the event/{id} function in the same way with the query option ?graph_data=true e..g. nsight-api-1.rr.ikanow.com/nsight/event/12345?graph_data=true
*you can also add on a &fake_geo=true to have it create a random docGeo item for every document
This will turn on the api to add the new data model (eventually I'll just roll the regular call over to default graph_data to true so we don't have to make any changes, I just didn't want to break the current UI).
This new data model will make a few changes:
1. The event will have a new visualizations field to let the gui know which visualizations to turn on for this event (in the analysis view), the field is located: event.metadata.json[0].visualizations e.g.: { "event": { ...lots of fields blah blah... "metadata": { "json": [ [ { "_id": "5347106ce4b051fe39633d2b", "key": { "id": "russia/country||sanction||ukraine/country", "cat": "diplomatic relations" }, "description": "2 documents containing association: russia/country sanction ukraine/country \n'Euphoria over Crimea won't last'\nRussia's euphoria over Crimea won't last\n", "title": "Russia & Ukraine - Diplomatic Relations", "titles": [ "'Euphoria over Crimea won't last'", "Russia's euphoria over Crimea won't last" ], "stats": { "median_date": "Mon Mar 31 2014 00:00:00 GMT-0400 (EDT)" }, "dates": [ "Mar 31, 2014 02:27:07 PM UTC", "Mar 31, 2014 02:24:51 PM UTC" ], "verbs": [], "visualizations": [ "MEDIATYPE_VOLUME_OVER_TIME", "GEO_MAP", "ENT_TYPE_SENTIMENT_FREQUENCY_RELEVANCE" ] } ] ] } } }
2. Depending on what visualizations are turned on in the above list, the data will be aggregated in new fields currently: doc_info_aggregation ent_info_aggregation ...more if we need them They are located at the root level e.g. { "doc_info_aggregation":[], "ent_info_aggregation":[] } doc_info will be objects like: { public String mediatype; public String publishedDate; public GeoPojo docGeo; //only exists if doc had a docGeo public String docTitle; public String docId; } ent_info will be objects like: { public EntityPojo entity; //entire normal entity objects just copied in here public String arch_type; public Double aggregated_sentiment; public Long aggregated_frequency; public Double average_relevance; public List<ObjectId> doc_ids; }
Add Comment