Merge pull request #26 from keyoke/keyoke/features

remove punc

overview.md

@@ -2,13 +2,21 @@
 This extension provides the ability to block duplicate work item creation, similarity between work items is currently determined based on [Dice's Coefficient](http://en.wikipedia.org/wiki/S%C3%B8rensen%E2%80%93Dice_coefficient).
 
 Checks are automatically performed on work items of the *same type* and on the following fields :
+
 - Title
 - Description
 
+We first check if a work item with a similar **Title** exists, if no similar work item exists we then perform the same checks based on the **Description** field. 
+
+Before performing our similarity check we normalize our text:
+
+1. Removing all HTML tags.
+2. Removing the following punctuation **!"#$%&'()\*+,-./:;?@[\\]^_`{|}~**.
+
 Similarity is established based on an index 0.0 - 1.0 :
 - 0.0 being least similar
 - 1.0 being most similar.
 
-By default this extension leverages a similarity index of 0.8.
+By default this extension leverages a similarity index of 0.8, in the future threshold will be configurable.
 
 This extension can be leveraged in combination with the [Find similar workitems](https://marketplace.visualstudio.com/items?itemName=tschmiedlechner.find-similar-workitems) extension to establish which work items are similar to the current item.

src/block-duplicate-observer.ts

@@ -104,9 +104,14 @@ class duplicateObserver implements IWorkItemNotificationListener {
         }
     }
 
+    // Remove things we dont want to compare on and ensure comparison based on lower case strings
     private normalizeString(orignial_text: string): string {
         if (orignial_text)
-            return striptags(orignial_text).trim().toLowerCase();
+            return striptags(orignial_text)
+                    .replace(/[.,\/#!$%\^&\*;:{}=\-_`~()]/g,"") // !"#$%&'()*+,-./:;?@[\]^_`{|}~ 
+                    .replace(/\s{2,}/g," ")
+                    .trim()
+                    .toLowerCase();
         else
             return "";
     }