Envyr is a tool to automagically package an application and run it in a sandboxed environment. This project is in active development and may break.
Ever wanted to run a script from a git repo without having to clone, install dependencies etc? Envyr does that for you. It can detect the language, install dependencies and run the project in a Sandboxed environment for you.
e.g
envyr run [email protected]:tchaudhry91/python-sample-script.git --autogen -- https://blog.tux-sudo.com > my_blog.htmlThis command will fetch the repo, build a sandbox (docker/podman supported at the moment), and run the script! Read more about the supported features below.
This project can be installed with Cargo. More pre-built packages will be available soon.
Envyr has built-in intelligence to run the following types of applications at the moment:
Envyr will automatically detect and run your python scripts.
Detection:
- If the project contains a .py file, it will be detected as a python script.
- If the project contains a requirements.txt file, it will be installed in the sandbox before execution.
- If a requirements.txt is not found, it will attempt to produce one using pipreqs.
- The entrypoint is detected via a
if __name__ == __main__or a shebang statements. Ties are broken via a priority and can be overridden with the-xflag.
Example:
- Here is envyr running a python script from a public repository.
$ envyr run --autogen [email protected]:sivel/speedtest-cli.git
Retrieving speedtest.net configuration...
Testing from xyz (xyz.xyz.xyz.xyz)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
^C
Cancelling...The first run will clone the repo and build the sandbox. Subsequent runs would be near instant.
Envyr will automatically detect and run your node.js scripts.
Detection:
- The project needs to contain a package.json.
- This is used to install the dependencies and figure out the entrypoint (
mainfrom package.json)
Detection:
- Based on Shebang.
- Pending: A way to detect dependencies. They can still be supplied manually while generating.
$ envyr -h
A tool to automagically create 'executable' packages for your scripts.
Usage: envyr [OPTIONS] <COMMAND>
Commands:
generate Generate the associated meta files. Overwrites if re-run.
alias Subcommands for aliases.
run Run the package with the given executor.
help Print this message or the help of the given subcommand(s)
Options:
-v, --verbose Emit Envyr logs to stdout. Useful for debugging. But may spoil pipes.
-h, --help Print help
-V, --version Print version
Running a Package
$ envyr run -h
Run the package with the given executor.
Usage: envyr run [OPTIONS] <PROJECT_ROOT> [-- <ARGS>...]
Arguments:
<PROJECT_ROOT> The location to the project. Accepts, local filesystem path/git repos.
[ARGS]...
Options:
-s, --sub-dir <SUB_DIR> relative sub-directory to the project_root, useful if you're working with monorepos.
-t, --tag <TAG> The tag of the package to run. Accepts git tags/commits. Defaults to latest. [default: latest]
--refresh refresh code cache before running.
--alias <ALIAS> Upon successful completion, record this run command as an alias. To allow usage of `envyr run <alias>` in the future.
-e, --executor <EXECUTOR> [default: docker] [possible values: docker, nix, native]
--autogen Attempt to automatically generate the package metadata before running. This overwrites existing metadata.
--fs-map [<FS_MAP>...] Mount the given directory as a volume. Format: host_dir:container_dir. Allows multiples. Only applicable on Docker Executor.
--port-map [<PORT_MAP>...] Map ports to host system, Format host_port:source_port. Allows multiples. Only applicable on Docker Executor.
--env-map [<ENV_MAP>...] Environment variables to pass through, leave value empty to pass through the value from the current environment. Format: 'key=value' or 'key' (passwthrough). Allows multiples.
-n, --name <NAME>
-i, --interpreter <INTERPRETER>
-x, --entrypoint <ENTRYPOINT>
-t, --type <PTYPE> [possible values: python, node, shell, other]
-h, --help Print help
Most cases should be covered by autodetection. Use the overrides if --autogen does not work.
Generating Package Metadata in Advance
Generate the associated meta files. Overwrites if re-run.
Usage: envyr generate [OPTIONS] <PROJECT_ROOT>
Arguments:
<PROJECT_ROOT> The location to the project. Accepts, local filesystem path/git repos.
Options:
-s, --sub-dir <SUB_DIR> relative sub-directory to the project_root, useful if you're working with monorepos.
-t, --tag <TAG> The tag of the package to run. Accepts git tags/commits. Defaults to latest. [default: latest]
--refresh refresh code cache before running.
-n, --name <NAME>
-i, --interpreter <INTERPRETER>
-x, --entrypoint <ENTRYPOINT>
-t, --type <PTYPE> [possible values: python, node, shell, other]
-h, --help Print help
The generate command is generally meant to be used by authors who can commit the .envyr folder generated by this command. This allows others to run this package with the optional (entrypoint/interpreter) overrides that the author desires by default.
Aliasing
You can generate aliases for common run commands to make them more ergonomic for regular use.
Pass the --alias flag to create a new alias on a successful run of a particular package.
$envyr run --alias sample --env-map=MYVAR --autogen [email protected]:tchaudhry91/python-sample-script.git -- https://blog.tux-sudo.com
This will store the above command as an alias called sample which can run as follows:
$envyr run sample
The args are also stored with the alias but can be overriden if required.
$envyr run sample -- https://test.com
Aliases can be managed using the following:
$envyr alias -h 2 ↵
Subcommands for aliases.
Usage: envyr alias <COMMAND>
Commands:
list List all aliases.
delete Delete an existing alias.
help Print this message or the help of the given subcommand(s)
Options:
-h, --help Print help
- Only Docker/Podman are available as the sandbox enviroments at the moment. Add nix/native options too.
- More Languages
- Bash Script Dependency Detection.
See the issue tracker/project board for more.