installation

# Install conda (miniconda is lighter) for python 3.6+.

https://docs.conda.io/projects/conda/en/latest/user-guide/install/


# Go to the folder you manage your projects (e.g. .../repos)

cd .../repos


# Get the *sensingbee* source code from git

$ git clone https://github.com/adaj/sensingbee.git


# Also, you'll need the *geohunter* package
#  to manage auxiliary data and variables from OpenStreetMap

$ git clone https://github.com/adaj/geohunter.git


# Create an environment to install the requirements there.

$ conda create -n senv
$ conda activate senv
$ conda init

# To exit your conda environment, "conda deactivate"


# Install sensingbee requirements.

$ conda install pandas==0.25.1
$ conda install scikit-learn==0.21.3
$ conda install geopandas==0.3.0


# Install pip in your env, we need some packages that are easily 
#  installed from pip.

$ conda install pip
$ conda init

# (Optional) If you want to use jupyter notebooks on your project,
#  you'll need to install it inside your conda environment.

$ conda install jupyter


# Create a folder for your project, let's say "~/sensingbee/project1"
#  and create a configuration file called "sensingbee.conf" there
#  with the path to the sensingbee source code. Please change 
#  "~/sensingbee" to the place where you placed the folder you got.
#  The FULL_PATH here refers to the absolute path (e.g. got from pwd)

$ mkdir ~/sensingbee/project1
$ echo "param,val" > FULL_PATH/repos/sensingbee/project1/sensingbee.conf
$ echo "SOURCE_PATH,FULL_PATH/repos/sensingbee" >> FULL_PATH/repos/sensingbee/project1/sensingbee.conf
$ echo "GEOHUNTER_PATH,FULL_PATH/repos/geohunter" >> FULL_PATH/repos/sensingbee/project1/sensingbee.conf


# Test *senginbee* installation

$ python3
>>> import pandas as pd
>>> conf = pd.read_csv('sensingbee.conf', index_col='param')
>>> import sys
>>> sys.path.append(conf.loc['GEOHUNTER_PATH','val'])
>>> sys.path.append(conf.loc['SOURCE_PATH','val'])
>>> import sensingbee
>>> import sensingbee.data_preparation
>>> import sensingbee.feature_engineering
>>> import sensingbee.ml_modeling