Chenhui/setup scripts (#178)

* move ray to linux setup script

* remove duplicated azureml-sdk to avoid errors

* add ray to ci yaml files

* update azureml-sdk

* update manual setup instructions

* minor change

Former-commit-id: 88f099759b

docs/SETUP.md

@@ -83,6 +83,11 @@ pip install -e fclib
 
 The library is installed in developer mode with the `-e` flag. This means that all changes made to the library locally, are immediately available.
 
+If you work with Linux or MacOS, you could also install Ray for parallel model training:
+```
+pip install ray>=0.8.2 
+``` 
+
 #### Jupyter kernel
 In order to run the example notebooks, make sure to run the notebooks in the conda environment we previously set up, `forecasting_env`. To register the conda environment in Jupyter, please run:
 

fclib/requirements.txt

@@ -6,5 +6,5 @@ requests>=2.23.0
 tqdm>=4.43.0
 matplotlib>=3.1.2
 tensorflow==2.0
-azureml-sdk[explain,automl]==1.0.85
-gitpython>=3.0.8
+gitpython>=3.0.8
+azureml-sdk==1.0.85

tests/ci/cpu_integration_tests_linux.yml

@@ -41,6 +41,7 @@ jobs:
       yes | conda env create -n forecast_cpu -f tools/environment.yml
       eval "$(conda shell.bash hook)" && conda activate forecast_cpu
       pip install -e fclib
+      pip install ray>=0.8.2
       echo "Conda env installed."
     displayName: 'Creating conda environment with dependencies'
 

tests/ci/cpu_unit_tests_linux.yml

@@ -41,6 +41,7 @@ jobs:
       yes | conda env create -n forecast_cpu -f tools/environment.yml
       eval "$(conda shell.bash hook)" && conda activate forecast_cpu
       pip install -e fclib
+      pip install ray>=0.8.2
       echo "Conda env installed."
     displayName: 'Creating Conda Environment with dependencies'
 

tools/environment.yml

@@ -38,7 +38,6 @@ dependencies:
     - tensorflow==2.0
     - tensorboard==2.1.0
     - nteract-scrapbook==0.3.1
-    - azureml-sdk[explain,automl]==1.0.85
     - statsmodels>=0.11.1
     - pmdarima>=1.1.1
-    - ray>=0.8.2
+    - azureml-sdk[explain,automl]==1.0.85

tools/environment_setup.sh

@@ -14,5 +14,8 @@ eval "$(conda shell.bash hook)" && conda activate forecasting_env
 # Install forecasting utility library
 pip install -e fclib
 
+# Install ray (available only on Linux and MacOS)
+pip install ray>=0.8.2 
+
 # Register conda environment in Jupyter
 python -m ipykernel install --user --name forecasting_env