doc

README.md

@@ -4,19 +4,30 @@
 ![demo](https://raw.githubusercontent.com/lsgrep/mldocs/master/assets/mldocs.gif)
 
 ## How does it work
-- Workflow will fetch the doc data from Github, then cache the data for a few days
+- `mldocs` fetches the doc data from Github(`data/ml.json`), then cache the data for a few days
+- The first query will be slow then it will be pretty fast afterwards
+- The plan is to update the `ml.json` periodically, so you won't have to update the workflow manually
 
-## Clear Cache
+## Clear the Cache
+To force update the local cache
 - `ml workflow:delcache`
 
 ## Commonly used keywords
+I've also merged my commonly used links(`data/base.json`) into the `ml.json` as well. 
+If you want to add your favorite website or link please submit a PR(just edit the `base.json`).
+
+
+| `keyword`   | `link`  | `description`  | 
+|---|---|---|
+| `colab` | `http://colab.research.google.com/`  |   |
+| `kaggle` | `https://www.kaggle.com/` |    | 
 
 
 ## Supported Versions
 - `ml.json` will be periodically updated to the latest stable versions of TensorFlow & PyTorch
 
 ## Keyword Descriptions
-- Not supported yet, but will add them in the future. You don't have to update the workflow 
+- Not supported yet, but will add them in the future.
 
 ## TODO
 - [ ] add Github Actions to generate `ml.json`