Conda - Env and Package manager for Python¶

What is conda?

• Conda is a powerful environment and package manager.
  • What are packages?
  • PyPi
• Create multiple environments with different versions of Python with different sets of libraries
• easily switch between environments
• http://anaconda.org/ provides free hosting for Python libraries
• conda helps to easily search for, install libraries and also to keep them up to date

Learn resources:

• getting started: http://conda.pydata.org/docs/test-drive.html#managing-environments
• reference doc: http://conda.pydata.org/docs/

List all environments¶

conda env list

Run the above command in terminal

Create a new environment¶

conda create --name <name_of_env> package=version

conda create --name devsummit_precon python=3.6

conda create --name data_sci_proj scipy pandas

Check if new folder was created in C:\anaconda3\envs

Activating environemnts¶

activate pro_devsummit_handson

search packages¶

conda search <package name>
conda search --channel <channel name> <package name>

conda search beautifulsoup

get info about a package¶

conda info <package name> ## gives you all dependencies
conda info --channel esri arcgis

Install packages¶

conda install <package name>

conda install pandas
conda install requests

conda instal --channel <channel_name> <package name>
conda install --channel esri arcgis
conda install -c esri arcgis

List all packages in current active environment¶

conda list
conda list nb*   # wild card search

Remove a package¶

conda remove <package name>

Remove an environment¶

conda env remove --name <name of env>
conda env remove --name devsummit_precon

Advanced steps¶

Clone an existing environment¶

For instance, to clone the root environment:

conda create --name clone_of_root --clone root

But you would most likely want to clone the ArcGIS Pro's Python environment so you can experiment without messing up the default one

conda create --name propy_clone --clone arcgispro-py3

Starting Jupyter notebook in a different directory¶

For the most part you can cd to desired directory after activating your environment and start the notebook. But if you cannot do it:

jupyter-notebook --notebook-dir=D:\code\my_notebooks

Exporting an environment to a file¶

It is useful to export your environment to a environment.yml file and include it in your repository. This allows replicating the project in a different computer by you or someone else.

conda env export --name your_env_name --file your_env_name.yml

This creates the .yml file with version number and source repo for all your packages. Thus, this file is specific for a particular OS and Python build version.

To create an agnostic file, use

conda env export --name your-env-name | cut -f 1 -d '=' > your-env-name.yml

This strips out the version numbers, enabling you to upgrade to newer versions of the packages.

Install from an environment file¶

To create a new environment from this file, run

conda env create -f your_env_file.yml