cloudsh

A Python CLI that wraps common Linux commands for both local and cloud files using cloudpathlib.

Cloud Storage Provider dependencies:

Installation

pip install -U cloudsh

# Install for different cloud storage providers
pip install -U cloudsh[gcs]  # Google Cloud Storage
pip install -U cloudsh[aws]  # Amazon S3
pip install -U cloudsh[azure]  # Azure Blob Storage

# Install for all cloud storage providers
pip install -U cloudsh[all]

Usage

cloudsh provides common Linux commands that work with both local and cloud files. Currently supported commands include:

• cat: Concatenate and print files
• cp: Copy files and directories
• head: Output the first part of files
• ls: List directory contents
• mkdir: Make directories
• mv: Move files and directories
• rm: Remove files and directories
• tail: Output the last part of files
• touch: Create empty files

And two additional commands:

• complete: Generate shell completion scripts
• sink: Redirect output to a file

Authentication

See: https://cloudpathlib.drivendata.org/stable/authentication/ for details on how to authenticate with cloud storage providers.

The commands works on local files as the GNU/Linux commands do

$ cloudsh ls /tmp
$ cloudsh cp /tmp/file.txt /tmp/file2.txt

The commands works on cloud files

$ cloudsh ls gs://my-bucket
$ cloudsh touch gs://my-bucket/file.txt

The commands works between local and cloud files

$ cloudsh cp /tmp/file.txt gs://my-bucket/file.txt
$ cloudsh mv gs://my-bucket/file.txt /tmp/file.txt

The sink command redirects output to a file

# It is easy to redirect output to a local file
$ echo "Hello, World!" > /tmp/hello.txt
# But it is not so easy to redirect output to a cloud file, so we use `sink`
$ echo "Hello, World!" | cloudsh sink gs://my-bucket/hello.txt
# Append to a cloud file
$ echo "Hello, World!" | cloudsh sink -a gs://my-bucket/hello.txt

Drop-in Replacement for GNU/Linux Commands

Since the commands work on local files as well, you can make aliases to use cloudsh as a drop-in replacement for the GNU/Linux commands.

alias cat='cloudsh cat'
alias cp='cloudsh cp'
alias head='cloudsh head'
alias ls='cloudsh ls'
alias mkdir='cloudsh mkdir'
alias mv='cloudsh mv'
alias rm='cloudsh rm'
alias tail='cloudsh tail'
alias touch='cloudsh touch'

What if I want to use the original GNU/Linux commands?

# alias ls='cloudsh ls'
ls -- -l  # actually executes `/usr/bin/ls -l`

Shell Completion

Generating Shell Completion Scripts

cloudsh provides shell completion support, including the subcommands, options and both local and cloud paths, for bash, zsh and fish. To enable it:

# For bash
mkdir -p ~/.local/share/bash-completion
cloudsh complete --shell bash > ~/.local/share/bash-completion/cloudsh
activate-global-python-argcomplete --user
# Restart your shell

# For zsh
# Create the completions directory
mkdir -p ~/.zsh/completions

# Generate the Zsh script
cloudsh complete --shell zsh > ~/.zsh/completions/_cloudsh

# Update ~/.zshrc
echo 'fpath=(~/.zsh/completions $fpath)' >> ~/.zshrc
echo 'autoload -Uz compinit && compinit' >> ~/.zshrc

# Restart your shell

# For fish
cloudsh complete --shell fish > ~/.config/fish/completions/cloudsh.fish

Because of the latency when completing cloud paths, a message 'fetching ...' will be shown when completing cloud paths. export CLOUDSH_COMPLETE_NO_FETCHING_INDICATOR=1 to disable it.

Using a caching file for the completion to avoid latency when completing cloud paths

# Only cache the paths at depth 2 in the bucket
cloudsh complete --update-cache --depth 2 gs://my-bucket

Note

Remember to update the cache when the bucket structure changes. You can set up a cron job to update the cache periodically.

Tip

For the first time you are using cached cloud path completion, a warning message will be shown to remind you that you are using a cached completion. The warning will only be shown when <tmpdir>/cloudsh_caching_warned does not exist, which will be created after the first warning. To disable the warning permanently, try export CLOUDSH_COMPLETE_CACHING_WARN=1.