Bokeh deployment cheat sheet

Dr. Rahul Remanan,
CEO, Moad Computer

What is Bokeh?
Bokeh is a Python library for interactive visualization. It uses the web browser to render the visualizations.

Why use Bokeh?
Bokeh's approach helps develop scalable, interactive data science tools with minimal graphic user interface (GUI) dependencies. It also fits with the 'browser is the computer' application development philosophy at Moad Computer. Since our goal at Moad Computer is to develop scalable data science applications within the comfort, safety and familiarity of a web browser, Bokeh fits very nicely into our development workflow.

What is this Bokeh cheat sheet for?
This cheat sheet is aimed at developers looking to deploy Bokeh either locally or in a cloud server.

What does this cheat sheet cover?
This cheat sheet covers three important platforms for enterprise customers. They are:

1.  AMD64/x86-64 Linux, the dominant enterprise server and cloud platform and also the easiest to deploy.
2.  AARCH64 Linux for ARM 64 bit CPU architecture. This CPU architecture is found in devices such as the Qualcomm internet of things (IoT)/ internet of machine learning things (IoMLT) CPUs and the Annapurna CPUs in Amazon cloud servers. This is a tricky deployment since each ARM64 CPU platform has its own requirements. But, the general principle described here applies to all of the ARM 64 CPU based devices or servers. This section is also relevant for developers interested in deploying Bokeh server from source.
3.  AMD64/x86-64 Windows 10 found in most workstations and enterprise devices.

For AMD64/x86-64 Linux users:

• Install using pip: python3 -m pip install fiona pandas scipy numpy geopandas bokeh

For AARCH64 Linux users:

• Install build tools and Cmake: apt install make clang llvm libc++-dev libc++abi-dev
• Install llmvmlite for python3: python3 -m pip install llvmlite
• Install libatlas-dev: apt install libatlas-base-dev
• Install AARCH64 linux dependencies: apt install libgdal-dev gdal-bin python3-gdal libtiff-dev libsqlite3-dev sqlite   
• Clone and prepare PROJ repository for Cmake: git clone https://github.com/OSGeo/PROJ; cd ./PROJ; ./autogen
• Configure PROJ: ./configure
• Make and install PROJ: make; make install
• Install pyproj: python3 -m pip install git+https://github.com/pyproj4/pyproj.git
• Install pandas: apt install python3-pandas
• Install app dependencies: python3 -m pip install fiona geopandas shapely bokeh --no-deps​

For Windows users (require Anaconda Python3):

• Create virtual Anaconda3 environment: conda create -n GeoPandas -c conda-forge python=3.7 geopandas --yes
  
• Activate virtual Anaconda3 environment: conda env activate GeoPandas
  
• Install git: conda install -n GeoPandas -c anaconda git
  
• Install matplolib: conda install -n GeoPandas -c conda-forge matplotlib descartes

• Install bokeh: conda install -n GeoPandas -c bokeh bokeh