pat-tree - npm Package Compare versions

Comparing version 1.0.3 to 1.0.4

index.js

@@ -112,7 +112,13 @@ var Tree = require("./lib/BTree");
 
-	segmentDoc: function(doc, SLPs) {
+	segmentDoc: function(doc, asArray) {
+		var SLPs = this.SLPs;
 		SLPs.sort(function(item1, item2) {
 			return item2.sistring.length - item1.sistring.length;
 		})
-		var result = "";
+		var result;
+		if(asArray) {
+			result = [];
+		} else {
+			result = "";
+		}
 		var paper = [];
@@ -150,6 +156,14 @@ for(var i = 0; i < doc.length; i++) {
 			if(paper[i].marked) {
-				result += " " + paper[i].keyword;
+				if(asArray && paper[i]) {
+					result.push(paper[i].keyword);
+				} else {
+					result += " " + paper[i].keyword;					
+				}
 				i += paper[i].keyword.length - 1;
-			} else {
-				result += " " + paper[i].char;
+			} else if(!/\s/.test(paper[i].char)){
+				if(asArray) {
+					result.push(paper[i].char);
+				} else {
+					result += " " + paper[i].char;					
+				}
 			}
@@ -243,2 +257,3 @@ }
 		}
+		this.SLPs = result;
 		return result;
@@ -565,1 +580,2 @@ },
 
+

package.json

 {
   "name": "pat-tree",
-  "version": "1.0.3",
+  "version": "1.0.4",
   "description": "PAT tree construction for Chinese documents, keyword extraction and text segmentation",
@@ -5,0 +5,0 @@ "main": "index.js",

README.md

@@ -68,5 +68,13 @@ pat-tree
 ```javascript
-var result = tree.segmentDoc(doc, SLPs); 
+var PATtree = require("pat-tree");
+var tree = new PATtree();
+
+//...
+
+tree.extractSLP(10, 0.5); 
+var result = tree.segmentDoc(doc, asArray); 
 ```
 
+you shold do `extractSLP` before doing text segmentation with `segmentDoc`.
+
 `doc` is the document to be segmented, data type: string.
@@ -76,4 +84,6 @@
 
-`result` is the result of document segmentation, data type: string.
+`result` is the result of document segmentation as an string of terms seperated by whitespaces, 
+or an array of terms if `asArray` is set to true.
 
+
 ### Convert to JSON
@@ -278,3 +288,3 @@
 
-
+* 1.0.4 `segmentDoc` no need to pass in SLPs, and enable to return array of terms.
 * 1.0.3 Minor change in module Node.js
@@ -281,0 +291,0 @@ * 1.0.2 Gaurantee SLP sorting order when `segmentDoc()`

No alert changes

Improved metrics

Total package byte prevSize

37278

increased by1.75%

Lines of code

973

increased by1.57%

Number of lines in readme file

308

increased by3.36%

No dependency changes