Conda (Python)

A Conda feed in ProGet acts as a private Conda package repository. Packages can be used directly by the Conda client and can also be browsed directly on the ProGet website.

Prerequisite Configuration

Although ProGet was tested with older versions, we recommend using Conda 4.10.0 or higher. You can use conda --version to see which version you have installed.

Conda Client Configuration

Before installing packages from ProGet, you should add ProGet to your list of Conda channels by using the conda config command.

conda config --add channels http://«proget-server»/conda/«feed-name»

If you want to install packages only from ProGet, then you'll want to remove other channels.

$ conda config --show channels
channels:
 - defaults
 http://«proget-server»/conda/«feed-name»

$ conda config --remove channels defaults

Note that these sources are stored in your ~/.condarc file.

Authenticated Feeds & Channels

If you've configured your feed to require authentication, you'll need to edit your ~/.condarc file to include basic authentication credentials in the channel URL.

These credentials can be your ProGet username and password, or a ProGet API key (that uses "api" as the username). For example:

• http://«username»:«password»@«proget-server»/conda/«feed-name»
• http://api:«api-key»@«proget-server»/conda/«feed-name»

You can also override the channel in the conda install and conda search command and include authentication in the URL. This can be done using the --override-channels parameter and formating the URL to include the username and password --channel http://«username»:«password»@«proget-server»/conda/«feed-name». For example:

$ conda install «package-name»=«package-version» --channel http://«username»:«password»@«proget-server»/conda/«feed-name» --override-channels

Testing Client Configuration

You can use the conda search command to make sure that your client is properly connected to ProGet.

$ conda search «search-string»

This will search all sources configured in your ~/.condarc file.

Installing Packages

You can use Conda packages from ProGet just like you would from anaconda.org. Packages are generally installed using the conda install command:

$ conda install «package-name»=«package-version»

Creating Packages

There are no special considerations when creating Conda packages for use in ProGet. ProGet supports all types of Conda packages, including .tar.bz2 files and the newer .conda format files.

Publishing Packages

You can push a Conda package to Proget by either uploading it through the ProGet site, or uploading a package to http://«proget-server»/conda/«feed-name» using an HTTP PUT or POST request. For example, to upload a package using curl:

$ curl --host http://«proget-server»/conda/«feed-name» --user «user»:«password» --upload-file «package-name-version».conda

Using pgscan With Conda and ProGet SCA

pgscan is a free and open-source tool designed to run immediately after building your application, generally as part of the Continuous Integration (CI) process. Starting in 1.5.6, pgscan added support to scan a Conda environment file.

To create a Conda environment file, you will need to run either:

conda list --explicit > requirments.txt

or

conda list --export > requirments.txt

This will list create a text file with a list of depedencies used by your Conda library or application. Once this hfile has been generated, you will need to run the pgscan identify passing your requirements.txt to the --input command.

For example, to run pgscan for your Conda application, you would run:

conda list --explicit > requirments.txt
pgscan identify
 --api-key=mySecretKey1000
 --proget-url=https://proget.corp.local/
 --input=requirements.txt
 --type=Conda
 --project-name="My Application"
 --version=5.00.21