log parsed/flattened json data on file input

Author: hmsklee

index.html

@@ -78,6 +78,7 @@
 <script src="libs/lodash.js"></script>
 <script src="libs/chroma.min.js"></script>
 
+<script src="js/parser.js"></script>
 <script src="js/GraphLayout.js"></script>
 <script src="js/enums.js"></script>
 <script src="js/style.js"></script>

js/parser.js

@@ -0,0 +1,183 @@
+function readJson(path) {
+
+  var json = path || './data/angus-ex.json';
+
+  function loadJSON(path, callback) {
+      var httpRequest = new XMLHttpRequest();
+      httpRequest.onreadystatechange = function() {
+          if (httpRequest.readyState === 4) {
+              if (httpRequest.status === 200) {
+                  var data = JSON.parse(httpRequest.responseText);
+                  if (callback) callback(data);
+              }
+          }
+      };
+      httpRequest.open('GET', path);
+      httpRequest.send(); 
+  }
+
+  /* parse json file*/
+  loadJSON(json, parseData);
+}
+
+
+/** init words and links from json data
+ *  - first hierarchy:  documents--sentences--words (+associated syntax tags)
+ *  -                   edges connect parts of speech
+ *  - second hierarchy: events--paths | arguments--themes
+ */
+function parseData(data) {
+  let wordDataArray = [];
+  let wordDataMap = {};
+  let syntaxDataArray = [];
+  let mentionDataArray = [];
+
+  for (var i in data.documents) {
+    let doc = data.documents[i];
+
+    doc.sentences.forEach(function(sentence, j) {
+      // parse word data
+      let sentenceData = sentence.words.map(function(word, k) {
+
+        let wordDataObject = {
+          text: word,
+          documentId: i,
+          sentenceId: j,
+          locationInSentence: k,
+          charLocationInSentence: sentence.startOffsets[k],
+          syntaxData: {
+            tag: sentence.tags[k]
+          },
+          bioData: {
+            tag: ''
+          }
+        };
+
+        wordDataMap[[i,j,k].join('-')] = wordDataArray.length;
+        wordDataArray.push(wordDataObject);
+        return wordDataObject
+      });
+
+      // create POS links
+      sentence.graphs["stanford-collapsed"].edges.forEach(function(edge) {
+        syntaxDataArray.push({
+          destination: sentenceData[edge.destination],
+          label: edge.relation,
+          type: edge.relation,
+          source: sentenceData[edge.source]
+        })
+      });
+
+    })
+  }
+
+  // flatten data.mentions array
+  let printMention = function(mention, i) {
+    if (mention.arguments) {
+      for (var j in mention.arguments) {
+        mention.arguments[j] = mention.arguments[j].map(printMention)
+      }
+    }
+
+    switch (mention.type) {
+      case "CorefTextBoundMention":
+        // has text(s) only
+        let start = wordDataMap[[mention.document, mention.sentence, mention.tokenInterval.start].join('-')];
+        let end = wordDataMap[[mention.document, mention.sentence, mention.tokenInterval.end].join('-')];
+
+        var link = {
+          sourceId: null,
+          destinationId: null,
+          words: wordDataArray.slice(start, end),
+          label: mention.displayLabel,
+          id: mention.id,
+          charOffset: mention.characterStartOffset,
+          type: mention.type
+        };
+        mentionDataArray.push(link);
+        return link;
+      case "CorefRelationMention":
+        // has argument(s)
+        // hard-coded the property --- need better data to parse this correctly
+
+        let keys = Object.keys(mention.arguments);
+        if (keys.length != 2 || !mention.arguments.controlled || !mention.arguments.controller) {
+          console.log("bad data parse: check CorefRelationMention", mention.arguments);
+        }
+        var link = {
+          sourceId: mention.arguments.controller.map(arg => arg.id),
+          destinationId: [{
+            name: mention.displayLabel,
+            id: mention.arguments.controlled.map(arg => arg.id)
+          }],
+          label: mention.displayLabel,
+          id: mention.id,
+          charOffset: mention.characterStartOffset,
+          type: mention.type
+        };
+
+        mentionDataArray.push(link);
+        return link;
+      case "CorefEventMention":
+        // has a trigger & argument(s)
+
+        if (mention.trigger.type == "TextBoundMention") {
+          let start = wordDataMap[[mention.trigger.document, mention.trigger.sentence, mention.trigger.tokenInterval.start].join('-')];
+          let end = wordDataMap[[mention.trigger.document, mention.trigger.sentence, mention.trigger.tokenInterval.end].join('-')];
+
+          var link = {
+            sourceId: null,
+            destinationId: null,
+            words: wordDataArray.slice(start, end),
+            label: mention.displayLabel,
+            id: mention.trigger.id,
+            charOffset: mention.trigger.characterStartOffset,
+            type: mention.trigger.type
+          };
+          mentionDataArray.push(link);
+        }
+        var link = {
+          sourceId: [mention.trigger.id],
+          destinationId: Object.keys(mention.arguments).map(key => {
+
+            return {
+              name: key,
+              charOffset: mention.characterStartOffset,
+              id: mention.arguments[key].map(arg => arg.id)
+            }
+
+          }),
+          label: mention.displayLabel,
+          id: mention.id,
+          type: mention.type
+        };
+        mentionDataArray.push(link);
+        return link;
+      default:
+        console.log("invalid type", mention.type);
+        break;
+    }
+  }
+  data.mentions.forEach(printMention);
+
+  // done parsing into semi-flat datasets...
+
+  console.log('words',wordDataArray);
+  console.log('pos',syntaxDataArray);
+  console.log('events',mentionDataArray);
+}
+
+document.querySelector('input[type=file]').onchange = function() {
+  var fr = new FileReader();
+  fr.onload = function(e) {
+    var object = {};
+    try {
+      object = JSON.parse(e.target.result);
+    }
+    catch(e) {
+      console.log("error",e);
+    }
+    parseData(object);
+  }
+  fr.readAsText(this.files[0]);
+}