resolves STREAMS-399
Project: http://git-wip-us.apache.org/repos/asf/incubator-streams/repo Commit: http://git-wip-us.apache.org/repos/asf/incubator-streams/commit/a2b29f0a Tree: http://git-wip-us.apache.org/repos/asf/incubator-streams/tree/a2b29f0a Diff: http://git-wip-us.apache.org/repos/asf/incubator-streams/diff/a2b29f0a Branch: refs/heads/master Commit: a2b29f0a301da0c1933dffb2a70d56b022189408 Parents: d5ef370 Author: Steve Blackmon @steveblackmon <[email protected]> Authored: Fri Oct 21 12:01:32 2016 -0500 Committer: Steve Blackmon @steveblackmon <[email protected]> Committed: Sat Oct 22 15:40:25 2016 -0500 ---------------------------------------------------------------------- .../converter/util/TwitterActivityUtil.java | 27 +++++- .../src/main/jsonschema/com/twitter/tweet.json | 98 ++++++++++++++------ 2 files changed, 93 insertions(+), 32 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/incubator-streams/blob/a2b29f0a/streams-contrib/streams-provider-twitter/src/main/java/org/apache/streams/twitter/converter/util/TwitterActivityUtil.java ---------------------------------------------------------------------- diff --git a/streams-contrib/streams-provider-twitter/src/main/java/org/apache/streams/twitter/converter/util/TwitterActivityUtil.java b/streams-contrib/streams-provider-twitter/src/main/java/org/apache/streams/twitter/converter/util/TwitterActivityUtil.java index 0bfa786..ac503cb 100644 --- a/streams-contrib/streams-provider-twitter/src/main/java/org/apache/streams/twitter/converter/util/TwitterActivityUtil.java +++ b/streams-contrib/streams-provider-twitter/src/main/java/org/apache/streams/twitter/converter/util/TwitterActivityUtil.java @@ -36,6 +36,7 @@ import org.apache.streams.twitter.Url; import org.apache.streams.twitter.pojo.Delete; import org.apache.streams.twitter.pojo.Entities; import org.apache.streams.twitter.pojo.Hashtag; +import org.apache.streams.twitter.pojo.Place; import org.apache.streams.twitter.pojo.Retweet; import org.apache.streams.twitter.pojo.Tweet; import org.apache.streams.twitter.pojo.User; @@ -49,6 +50,7 @@ import java.util.HashMap; import java.util.List; import java.util.Map; +import static com.google.common.math.DoubleMath.mean; import static org.apache.streams.data.util.ActivityUtil.ensureExtensions; /** @@ -273,8 +275,7 @@ public class TwitterActivityUtil { .or(Optional.of(tweet.getId().toString())) .orNull() )); - location.put("coordinates", tweet.getCoordinates()); - extensions.put("location", location); + location.put("coordinates", boundingBoxCenter(tweet.getPlace())); extensions.put("location", location); } /** @@ -350,4 +351,26 @@ public class TwitterActivityUtil { extensions.put("keywords", tweet.getText()); } + + /** + * Compute central coordinates from bounding box + * @param place the bounding box to use as the source + */ + public static List<Double> boundingBoxCenter(Place place) { + if( place == null ) return new ArrayList<>(); + if( place.getBoundingBox() == null ) return new ArrayList<>(); + if( place.getBoundingBox().getCoordinates().size() != 1 ) return new ArrayList<>(); + if( place.getBoundingBox().getCoordinates().get(0).size() != 4 ) return new ArrayList<>(); + List<Double> lats = Lists.newArrayList(); + List<Double> lons = Lists.newArrayList(); + for( List<Double> point : place.getBoundingBox().getCoordinates().get(0)) { + lats.add(point.get(0)); + lons.add(point.get(1)); + } + List<Double> result = new ArrayList<Double>(); + result.add(new Double(mean(lats))); + result.add(new Double(mean(lons))); + return result; + } + } http://git-wip-us.apache.org/repos/asf/incubator-streams/blob/a2b29f0a/streams-contrib/streams-provider-twitter/src/main/jsonschema/com/twitter/tweet.json ---------------------------------------------------------------------- diff --git a/streams-contrib/streams-provider-twitter/src/main/jsonschema/com/twitter/tweet.json b/streams-contrib/streams-provider-twitter/src/main/jsonschema/com/twitter/tweet.json index bdd2201..e529559 100644 --- a/streams-contrib/streams-provider-twitter/src/main/jsonschema/com/twitter/tweet.json +++ b/streams-contrib/streams-provider-twitter/src/main/jsonschema/com/twitter/tweet.json @@ -2,7 +2,7 @@ "type": "object", "$schema": "http://json-schema.org/draft-03/schema";, "$license": [ - "http://www.apache.org/licenses/LICENSE-2.0"; + "http://www.apache.org/licenses/LICENSE-2.0"; ], "id": "#", "javaType" : "org.apache.streams.twitter.pojo.Tweet", @@ -30,22 +30,34 @@ "type" : "number" }] }, - "coordinates": { + "place": { "type": "object", - "javaType" : "org.apache.streams.twitter.pojo.Coordinates", + "javaType" : "org.apache.streams.twitter.pojo.Place", "javaInterfaces": ["java.io.Serializable"], - "items": { - "properties": { - "type": { - "type": "string" - }, - "coordinates": { - "type": "array", - "items": [ - { - "type": "integer" + "properties": { + "id": { + "type": "string" + }, + "bounding_box": { + "type": "object", + "javaType" : "org.apache.streams.twitter.pojo.BoundingBox", + "javaInterfaces": ["java.io.Serializable"], + "properties": { + "type": { + "type": "string" + }, + "coordinates": { + "type": "array", + "items": { + "type": "array", + "items": { + "type": "array", + "items": { + "type": "number" + } + } } - ] + } } } } @@ -88,20 +100,20 @@ "hashtags": { "type": "array", "items": { - "type": "object", - "javaType": "org.apache.streams.twitter.pojo.Hashtag", - "javaInterfaces": ["java.io.Serializable"], - "properties": { - "text": { - "type": "string" - }, - "indices": { - "type": "array", - "items": { - "type": "integer" - } - } - } + "type": "object", + "javaType": "org.apache.streams.twitter.pojo.Hashtag", + "javaInterfaces": ["java.io.Serializable"], + "properties": { + "text": { + "type": "string" + }, + "indices": { + "type": "array", + "items": [{ + "type": "integer" + }] + } + } } }, "urls": { @@ -118,7 +130,7 @@ "type": "array", "items": [ { - "type" : "integer" + "type" : "integer" } ] }, @@ -167,6 +179,10 @@ "ignore_malformed": false, "type": "integer" }, + "favorite_count": { + "ignore_malformed": false, + "type": "integer" + }, "retweet_count": { "ignore_malformed": false, "type": "integer" @@ -176,6 +192,28 @@ }, "user": { "$ref": "User.json" + }, + "is_quote_status": { + "type": "boolean" + }, + "quoted_status_id": { + "ignore_malformed": false, + "type": "integer" + }, + "quoted_status": { + "type": "object", + "required" : false, + "description" : "Describes the tweet being quoted.", + "$ref" : "tweet.json" + }, + "retweeted_status_id": { + "type": "integer" + }, + "retweeted_status": { + "type": "object", + "required" : false, + "description" : "Describes the tweet being retweeted.", + "$ref" : "tweet.json" } } -} \ No newline at end of file +}
