improvements

* chance to continue on mismatch url for redirectTab
* support empty field selectors
* add Extractor.results()
* add ExtractResult.walk(), ExtractResult.visit()
* add ! directive to click elements
* code optimize

readme.md

@@ -17,6 +17,8 @@ Extract current page
 ```js
 $('.item', ['a', 'a@href']);
 new Extractor().task('.item', ['a', 'a@href']).start();
+// fieldSelectors can be empty strings if items have no child to select
+new Extractor().task('.item a', ['', '@href']).start();
 ```
 
 > `$(...args)` is the short form of `new Extractor().task(...args).start();`, which is introduced later.
@@ -67,7 +69,9 @@ job = new Extractor().task('.search-list-item', ['a@href'], ["http://sample.com/
 job.stop();
 ```
 
-## Extract Attributes.
+> Next time you call `job.start();`, the job will continues from where it stopped.
+
+## Extract Attributes
 
 e.g.: link text and target (use 'selector@attribute')
 
@@ -75,6 +79,14 @@ e.g.: link text and target (use 'selector@attribute')
 new Extractor().task('.item', ['a', 'a@href']).start();
 ```
 
+## Click Selected Elements
+
+The following clicks selected links and extracts link `text` and `href`
+
+```js
+new Extractor().task('.item', ['!a', 'a@href']).start();
+```
+
 ## Advanced Usage
 
 ### Use Task Chain.
@@ -202,6 +214,26 @@ To stop watching, you can either `close current window`, or:
 e.stop();
 ```
 
+## Results Operation
+
+To get the results of a task:
+
+```js
+let results = job.results(0);
+```
+
+Visit URLs (if any) in the results one by one:
+
+```js
+results.visit();
+```
+
+Walk through all results one by one:
+
+```js
+results.walk((row,col,value)=>{console.log(value)});
+```
+
 ## Developpment
 
 Clone this project and execute: