Merge pull request #67 from LuoXueling/dev-xlluo

Update documentation

README.md

@@ -58,6 +58,8 @@ The package stands on the shoulder of the giants:
 
 ## Installation/Usage
 
+A full documentation is available [here](https://tabular-ensemble.readthedocs.io/en/latest/index.html). For a quick start:
+
 1. `tabular_ensemble` can be installed using pypi by running the following command:
 
 ```shell
@@ -89,7 +91,7 @@ python main.py --base sample --epoch 10
 ```
 where `--base` refers to the configuration file, and additional arguments (such as `--epoch` here) refer to those in `config/default.py`.
 
-See the documentation pages for details.
+See the [documentation pages](https://tabular-ensemble.readthedocs.io/en/latest/index.html) for details.
 
 ## Citation
 

docs/source/examples/get_started.rst

@@ -6,6 +6,7 @@ Get Started
     :maxdepth: 2
 
     get_started/installation
+    get_started/quick_start
     get_started/run_sample
     get_started/new_dataset
     get_started/inference

docs/source/examples/get_started/installation.md

@@ -2,7 +2,17 @@
 
 ## Ubuntu
 
-Before installing `tabular-ensemble`, `PyTorch` should be installed first following its [documentation](https://pytorch.org/get-started/locally/).
+Before installing `tabular-ensemble`, `torch>=1.12.0` should be installed first following its [documentation](https://pytorch.org/get-started/locally/).
+
+### From PyPI
+
+```shell
+pip install tabensemb
+```
+
+Use `pip install tabensemb[test]` instead if you want to run unit tests.
+
+### From source
 
 Clone the repository from GitHub
 
@@ -17,10 +27,13 @@ Then install the package
 pip install -e .
 ```
 
-To test the functionality of the package
+Use `pip install -e .[test]` instead if you want to run unit tests.
+
+### Unit test
+
+To run unit tests:
 
 ```shell
-pip install -e .[test]
 cd test
 pytest .
 ```

docs/source/examples/get_started/quick_start.md

@@ -0,0 +1,23 @@
+# Quick start
+
+Place your `.csv` or `.xlsx` file in a `data` subfolder (e.g., `data/sample.csv`), and generate a configuration file in a `configs` subfolder (e.g., `configs/sample.py`), containing the following content
+```python
+cfg = {
+    "database": "sample",
+    "continuous_feature_names": ["cont_0", "cont_1", "cont_2", "cont_3", "cont_4"],
+    "categorical_feature_names": ["cat_0", "cat_1", "cat_2"],
+    "label_name": ["target"],
+}
+```
+
+Run the experiment using the configuration and the data using
+```python
+python main.py --base sample --epoch 10
+```
+where `--base` refers to the configuration file, and additional arguments (such as `--epoch` here) refer to those in `config/default.py`.
+
+Here is a typical usage of the repository
+
+```python
+python main.py --base sample --epoch 200 --batch_size 128 --bayes_opt --split_ratio 0.6 0.2 0.2
+```