conda-lock

Conda lock is a lightweight library that can be used to generate fully reproducible lock files for conda environments.

It does this by performing a conda solve for each platform you desire a lockfile for.

This also has the added benefit of acting as an external pre-solve for conda as the lockfiles it generates results in the conda solver not being invoked when installing the packages from the generated lockfile.

why?

Conda environment.yml files are very useful for defining desired environments but there are times when we want to be able to EXACTLY reproduce an environment by just installing and downloading the packages needed.

This is particularly handy in the context of a gitops style setup where you use conda to provision environments in various places.

installation

pip install conda-lock
conda install -c conda-forge conda-lock

Basic usage

# generate the lockfiles
conda-lock -f environment.yml -p osx-64 -p linux-64

# create an environment from the lockfile
conda-lock install [-p {prefix}|-n {name}] conda-linux-64.lock

# alternatively, use conda command directly
conda create -n my-locked-env --file conda-linux-64.lock

Advance usage

File naming

By default conda-lock will name files as "conda-{platform}.lock".

If you want to override that call conda-lock as follows.

conda-lock --filename-template "specific-{platform}.conda.lock"

Compound specification

Conda-lock will build a spec list from several files if requested.

conda-lock -f base.yml -f specific.yml -p linux-64 --filename-format "specific-{platform}.lock"

In this case all dependencies are combined, and the first non-empty value for channels is used as the final specification.

This works for all supported file types.

channel overrides

You can override the channels that are used by conda-lock in case you need to override the ones specified in an environment.yml

conda-lock -c conda-forge -p linux-64

--dev-dependencies/--no-dev-dependencies

By default conda-lock will include dev dependencies in the specification of the lock (if the files that the lock is being built from support them). This can be disabled easily

conda-lock --no-dev-dependencies -f ./recipe/meta.yaml

--strip-auth, --auth and --auth-file

By default conda-lock will leave basic auth credentials for private conda channels in the lock file. If you wish to strip authentication from the file, provide the --strip-auth argument.

conda-lock --strip-auth -f environment.yml

In order to conda-lock install a lock file with its basic auth credentials stripped, you will need to create an authentication file in .json format like this:

{
  "domain": "username:password",
  // ...
}

You can provide the authentication either as string through --auth or as a filepath through --auth-file.

conda-lock install --auth-file auth.json conda-linux-64.lock

Supported file sources

Conda lock supports more than just environment.yml specifications!

Additionally conda-lock supports meta.yaml (conda-build) and pyproject.toml ( flit and poetry based). These do come with some gotchas but are generally good enough for the 90% use-case.

meta.yaml

Conda-lock will attempt to make an educated guess at the desired environment spec in a meta.yaml. This is not guaranteed to work for complex recipes with many selectors and outputs. For multi-output recipes, conda-lock will fuse all the dependencies together. If that doesn't work for your case fall back to specifying the specification as an environment.yml

Since a meta.yaml doesn't contain channel information we make use of the following extra key to retrieve channels

# meta.yaml

extra:
  channels:
    - conda-forge
    - defaults

pyproject.toml configuration

Since pyproject.toml files are commonly used by python packages it can be desirable to create a lock file directly from those dependencies to single-source a package's dependencies. This makes use of some conda-forge infrastructure (pypi-mapping) to do a lookup of the PyPI package name to a corresponding conda package name (e.g. docker -> docker-py). In cases where there exists no lookup for the package it assumes that the PyPI name, and the conda name are the same.

channels

# pyproject.toml

[tool.conda-lock]
channels = [
    'conda-forge', 'defaults'
]

extra dependencies

Since in a pyproject.toml all the definitions are python dependencies if you need to specify some non-python dependencies as well this can be accomplished by adding the following sections to the pyproject.toml

# pyproject.toml

[tool.conda-lock.dependencies]
sqlite = ">=3.34"

Dockerfile example

In order to use conda-lock in a docker-style context you want to add the lockfile to the docker container. In order to refresh the lock file just run conda-lock again.

Given aa file tree like

  Dockerfile
  environment.yaml
* conda-linux-64.lock

You want a dockerfile that is structured something similar to this

# Dockerfile

# Build container
FROM continuumio/miniconda:latest as conda

ADD conda-linux-64.lock /locks/conda-linux-64.lock
RUN conda create -p /opt/env --copy --file /locks/conda-linux-64.lock

# Primary container

FROM gcr.io/distroless/base-debian10

COPY --from=conda /opt/env /opt/env

Release files for conda-lock 0.10.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for conda-lock 0.10.0
File	Size	Uploaded
conda_lock-0.10.0.tar.gz	61.5 kB	Jun 9, 2021	Details

Built distribution (wheel)

Table of built distributions (wheels) for conda-lock 0.10.0
File	Interpreter	ABI	Platform	Reset
conda_lock-0.10.0-py3-none-any.whl 19.4 kB Jun 9, 2021	Python 3	none	any	Details

Total release size: 80.9 kB

Release files / conda_lock-0.10.0.tar.gz

Download URL	conda_lock-0.10.0.tar.gz
Size	61.5 kB
Tags	Source
SHA-256 checksum How to use checksums	`cc947c9fe98d67eb6cb0a77190515f552933c3c7658c2837e117ca792a5feec3`
BLAKE2b-256 checksum How to use checksums	`131e854cbebc4e97029c5e4010f4edd4405552ddf7bec80091fb60a0aa33f64e`
Upload date	Jun 9, 2021
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/3.4.1 importlib_metadata/4.5.0 pkginfo/1.7.0 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.61.0 CPython/3.9.5`

Release files / conda_lock-0.10.0-py3-none-any.whl

Download URL	conda_lock-0.10.0-py3-none-any.whl
Size	19.4 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`4a8c07ff82e554420d8e01843fdefea110732335b8ca26fe54682ccaf0c2e840`
BLAKE2b-256 checksum How to use checksums	`d998bf8545fe5172d762d7760d1f751cfc2528457485bed4cd98032f8416176e`
Upload date	Jun 9, 2021
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/3.4.1 importlib_metadata/4.5.0 pkginfo/1.7.0 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.61.0 CPython/3.9.5`