[doc] Include dask examples into doc. (#7530)

demo/dask/README.md

@@ -1,6 +0,0 @@
-Dask
-====
-
-This directory contains some demonstrations for using `dask` with `XGBoost`.
-For an overview, see
-https://xgboost.readthedocs.io/en/latest/tutorials/dask.html .

demo/dask/README.rst

@@ -0,0 +1,5 @@
+XGBoost Dask Feature Walkthrough
+================================
+
+This directory contains some demonstrations for using `dask` with `XGBoost`.  For an
+overview, see :doc:`/tutorials/dask`

demo/dask/cpu_survival.py

@@ -1,3 +1,9 @@
+"""
+Example of training survival model with Dask on CPU
+===================================================
+
+"""
+
 import xgboost as xgb
 import os
 from xgboost.dask import DaskDMatrix

demo/dask/cpu_training.py

@@ -1,3 +1,8 @@
+"""
+Example of training with Dask on CPU
+====================================
+
+"""
 import xgboost as xgb
 from xgboost.dask import DaskDMatrix
 from dask.distributed import Client

demo/dask/dask_callbacks.py

@@ -1,4 +1,7 @@
-"""Example of using callbacks in Dask"""
+"""
+Example of using callbacks with Dask
+====================================
+"""
 import numpy as np
 import xgboost as xgb
 from xgboost.dask import DaskDMatrix

demo/dask/gpu_training.py

@@ -1,3 +1,7 @@
+"""
+Example of training with Dask on GPU
+====================================
+"""
 from dask_cuda import LocalCUDACluster
 from dask.distributed import Client
 from dask import array as da

demo/dask/sklearn_cpu_training.py

@@ -1,6 +1,7 @@
-'''Dask interface demo:
-
-Use scikit-learn regressor interface with CPU histogram tree method.'''
+"""
+Use scikit-learn regressor interface with CPU histogram tree method
+===================================================================
+"""
 from dask.distributed import Client
 from dask.distributed import LocalCluster
 from dask import array as da
@@ -16,7 +17,7 @@ def main(client):
     y = da.random.random(m, partition_size)
 
     regressor = xgboost.dask.DaskXGBRegressor(verbosity=1, n_estimators=2)
-    regressor.set_params(tree_method='hist')
+    regressor.set_params(tree_method="hist")
     # assigning client here is optional
     regressor.client = client
 
@@ -26,13 +27,13 @@ def main(client):
     bst = regressor.get_booster()
     history = regressor.evals_result()
 
-    print('Evaluation history:', history)
+    print("Evaluation history:", history)
     # returned prediction is always a dask array.
     assert isinstance(prediction, da.Array)
-    return bst                  # returning the trained model
+    return bst  # returning the trained model
 
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     # or use other clusters for scaling
     with LocalCluster(n_workers=4, threads_per_worker=1) as cluster:
         with Client(cluster) as client:

demo/dask/sklearn_gpu_training.py

@@ -1,6 +1,7 @@
-'''Dask interface demo:
-
-Use scikit-learn regressor interface with GPU histogram tree method.'''
+"""
+Use scikit-learn regressor interface with GPU histogram tree method
+===================================================================
+"""
 
 from dask.distributed import Client
 # It's recommended to use dask_cuda for GPU assignment