Skip to content

Manage your Bee node and interact with the Swarm network via the CLI

License

Notifications You must be signed in to change notification settings

ethersphere/swarm-cli

Repository files navigation

Swarm-CLI

Node.js tests standard-readme compliant js-standard-style

Stay up to date with changes by joining the official Discord and by keeping an eye on the releases tab.

Table of Contents

Demo

Purchasing a Postage Stamp

Swarm CLI Stamp Buy Command

Uploading a File

Swarm CLI Upload Command

Creating an Identity

Swarm CLI Identity Create Command

Uploading to a Feed

Swarm CLI Feed Upload Command

Description

Manage your Bee node and interact with the Swarm network via the CLI

The goal of this project is to handle most of the Swarm operations through CLI at some point in the future.

For the currently supported operations, see the Commands section.

Installation

Requires Node.js 18 or higher. It is recommended to use NVM to manage Node.js versions.

Warning! Installing Node.js via package managers (e.g. apt, yum, brew) may not work as expected, as they may install an older version of Node.js.

From npm

Install globally:

npm install --global @ethersphere/swarm-cli

From source

See the Development section.

Usage

The general usage is to provide a command, optionally a subcommand, then arguments and options.

swarm-cli command [subcommand] <arguments> [options]

Running a command with the --help option prints out the usage of a command.

Commands

Running swarm-cli without arguments prints the available commands:

Swarm CLI 2.11.0 - Manage your Bee node and interact with the Swarm network via the CLI

█ Usage:

swarm-cli COMMAND [OPTIONS]

█ Available Groups:

pinning    Pin, unpin and check pinned chunks
identity   Import, export and manage keypairs, identities
feed       Upload, update and view feeds
cheque     Deposit, withdraw and manage cheques
stamp      Buy, list and show postage stamps
pss        Send, receive, or subscribe to PSS messages
manifest   Operate on manifests
utility    Utility commands for managing wallets

Run 'swarm-cli GROUP --help' to see available commands in a group

█ Available Commands:

upload      Upload file to Swarm
download    Download arbitrary Swarm hash
hash        Print the Swarm hash of a file
status      Check Bee status
addresses   Display the addresses of the Bee node
stake       Manages nodes stake

Run 'swarm-cli COMMAND --help' for more information on a command

Example usage

Let's say we want to upload our website to Swarm and update a feed to point to the newest version. For updating a feed we would need to sign it with an Ethereum key, so first we need to create one with the identity create command:

swarm-cli identity create

This command will ask for a password. After that a new identity is created (named main). Now we can use this identity to sign updates. It's also possible to import and export Ethereum JSON V3 format identities that works with other apps (e.g. wallets).

Another requirement for uploading to the Swarm network is a valid postage batch, also called a postage stamp or simply a stamp. Stamps need to be purchased with xBZZ tokens. We can use the stamp buy command to take care of this step. The --amount and --depth options alter the capacity of the postage stamp. For example, running stamp buy --amount 1 --depth 20 will get back with a Stamp ID after a while. We will be using that with the --stamp option in commands which upload files, or write feeds.

For uploading to a feed we can use the feed upload command. It expects the path of the folder (or file) we want to upload and as options it expects identity to be provided along with the password that belongs to it, as well as the earlier mentioned stamp.

swarm-cli feed upload path-to-be-uploaded --identity my-identity --password my-secret-password --stamp stamp-id

In this example we are uploading the content of the dist folder. If the uploading was successful the last printed line will contain a Feed Manifest URL. This URL can be opened in the browser. If the uploaded folder contains an index.html file then it will be automatically displayed when visiting the URL.

This URL will stay the same when we upload an updated version of the website. Because of this we can also put this URL into a reverse proxy configuration or use the reference (the hex string after the /bzz/) in an ENS record. There is more information about that in the Bee documentation. The uploaded content can be found on the link in the line starting with URL. This will change every time the content is modified.

Usability Features

Uploading Files, Folders, Websites, and Arbitrary Data from stdin

Files

Use swarm-cli to upload a single file:

swarm-cli upload README.md

The command above will print a /bzz URL that may be opened in the browser. If the browser is able to handle the file format then the file is displayed, otherwise it will be offered to be downloaded.

Folders and Websites

swarm-cli also supports uploading folders with the same upload command:

swarm-cli upload build/

This also yields a /bzz URL. If there is an index.html present in the root of the folder, --index-document will be automatically applied by swarm-cli. This option sets which file the browser should open for an empty path. You may also freely set --index-document during upload to change this.

Standard Input

You can pipe data from other commands to swarm-cli using the --stdin option.

curl -L https://picsum.photos/200 | swarm-cli --stdin --stamp [...]

Unlike other upload methods, this results in a /bytes URL, which cannot be displayed by browsers normally. You can still share your hash and others can download it. However, with the --name option, you can give your arbitrary data a file name, and swarm-cli will attempt to determine the suitable content type for your data. Given it is successful, swarm-cli will print a /bzz URL instead of the /bytes URL, which is good to be displayed in browsers. Example:

curl -L https://picsum.photos/200 | swarm-cli --stdin --stamp [...] --name random.jpg

There is also a --content-type option if you want to adjust it manually:

curl -L https://picsum.photos/200 | swarm-cli --stdin --stamp [...] --name random --content-type image/jpeg

Please note that stdin is reserved for the data you are uploading, so interactive features are disabled during this time. Because of that, --stamp must be passed beforehand. You may create an alias for grabbing the ID of the least used postage stamp:

alias st='swarm-cli stamp list --least-used --limit 1 --hide-usage --quiet'

Leveraging the alias above, you can use a shortcut for uploading from stdin:

curl -L https://picsum.photos/200 | swarm-cli --stdin --stamp $(st)

Custom HTTP Headers

Similarly to curl, you may use the --header or -H option to specify as many additional headers as you want, which will be sent with all requests:

swarm-cli upload README.md -H "Authorization: [...]" -H "X-Custom-Header: Your Value"

Autocomplete

swarm-cli has support for autocomplete in bash, zsh and fish. This turns on <tab><tab> suggestions which can complete commands, paths and options for you.

To enable it, you need to install it once via two options:

  • Running swarm-cli --generate-completion and following the instructions there
  • Running swarm-cli --install-completion which automatically appends the completion script to your configuration file
Shell Completion System Configuration Path
bash compdef & compadd OR complete & COMPREPLY $HOME/.bashrc & $HOME/.bash_profile
zsh compdef & compadd OR complete & COMPREPLY $HOME/.zshrc
fish complete $HOME/.config/fish/config.fish

Warning! If you start a subshell (e.g. running bash from zsh), your SHELL env variable would still be the old value! The generation and completion script cannot detect your shell accurately in that case, so please set SHELL manually. It is generally advised to run --generate-completion first to ensure the shell and the paths are properly detected.

Example:

$ SHELL=zsh
$ swarm-cli --generate-completion
Your shell is: zsh
Found configuration file path: /Users/Swarm/.zshrc

Append the completion script below to your configuration file to enable autocomplete.
You need to source your configuration, or restart your shell, to load the changes.

<script>

Numerical Separator and Units

As most of the units are specified in wei and PLUR - the smallest denominations of currencies - they are a bit difficult to write out.

To aid this, you may use underscores (_) and K, M, B and T units to make your numbers more comprehensible.

Example:

swarm-cli stamp buy --amount 10M --depth 17 --gas-price 10_000_000_000_000

You may combine the two: 100_000T.

Stamp Picker

Unless you are running in --quiet mode, some options are not hard-required.

Look for hints in the --help sections. Take the upload command for example:

█ Required Options:

   --stamp   ID of the postage stamp to use  [required when quiet][string]

That means, you don't have to provide the postage stamp ID beforehand. Simply running swarm-cli upload <path> will prompt you with an interactive stamp picker:

? Please select a stamp for this action.

  Stamp ID                                                         Utilization
 (Use arrow keys)
❯ b9d5bb548c2c209cb99cbb27b0bef59b8f0cd3558363e307f45177b5a64ad0c8 (1)

Identity Picker

Similarly to Stamp Picker, when an identity is not provided, an interactive picker will be prompted.

Take the command feed upload for example. Albeit it takes quite a lot of options, you can run it with typing as little as feed upload <path>.

swarm-cli will take you through some prompts to interactively specify all required options, such as identity, password of the identity, and the mandatory stamp.

Passing identities is also tolerant to errors, so if you provide one which does not exist, the output will tell you and you can correct it:

The provided identity does not exist. Please select one that exists.
? Please select an identity for this action (Use arrow keys)
❯ main

Human Readable Topics

You may need to pass topics on multiple occasions - for example, when uploading to feeds.

Topics are 32-byte long identifiers, so you need 64 characters to write them out in hexadecimal string format.

You can do that with the --topic or -t option, or alternatively take a shortcut and use a human readable string which will be hashed by swarm-cli for your convenience. It is available via the --topic-string or -T option.

Example:

swarm-cli feed upload [...] -T "Awesome Swarm Website"

This is also indicated in the --help section:

-t --topic         32-byte long identifier in hexadecimal format    [hex-string][default all zeroes]
-T --topic-string  Construct the topic from human readable strings                          [string]

Only one is required: [topic] or [topic-string]

Manifest address scheme

The manifest commands enable low-level operation on manifests. These always require a root manifest reference (hash) argument as the input. Some commands, however, work with subparts of the manifest. A few examples are: downloading only a folder from a manifest, listing files only under a specific path in a manifest, and adding files or folders not to the root of the manifest, but under some path.

These can be achieved by using the bzz://<hash>/<path> scheme in the <address> argument as follows:

List entries under the /command/pss prefix in manifest 1512546a3f4d0fea9f35fa1177486bdfe2bc2536917ad5012ee749604a7b425f

swarm-cli manifest list bzz://1512546a3f4d0fea9f35fa1177486bdfe2bc2536917ad5012ee749604a7b425f/command/pss

Download README.md from manifest 1512546a3f4d0fea9f35fa1177486bdfe2bc2536917ad5012ee749604a7b425f

swarm-cli manifest download bzz://1512546a3f4d0fea9f35fa1177486bdfe2bc2536917ad5012ee749604a7b425f/README.md

Note: The bzz:// protocol can be omitted.

Automating tasks with Swarm-CLI

Running swarm-cli with the flag --quiet (or -q for short) disables all interactive features, and makes commands print information in an easily parsable format. The exit code also indicates whether running the command was successful or not. These may be useful for automating tasks both in CI environments and in your terminal too.

Below you will find a few snippets to give an idea how it can be used to compose tasks.

Connectivity

Exit if not all status checks succeed:

swarm-cli status -q || exit 1

Check Bee API connection, compatibility does not matter:

swarm-cli status -q | head -n 1 | grep "^OK"

Postage Stamps

Grab the first postage stamp:

swarm-cli stamp list --limit 1 --quiet --hide-usage

Grab the least used postage stamp:

swarm-cli stamp list --limit 1 --quiet --hide-usage --least-used

List all postage stamps with zero utilization:

swarm-cli stamp list --max-usage 0 --quiet --hide-usage

Sort postage stamps based on utilization (least utilized comes first):

swarm-cli stamp list --least-used --quiet

Uploading

Upload a file with the least utilized postage stamp (that has at most 50% usage):

STAMP=$(swarm-cli stamp list --max-usage 50 --least-used --limit 1 --quiet --hide-usage)
swarm-cli upload -q README.md --stamp $STAMP

Config

The configuration file is placed in a hidden folder named swarm-cli.

In case of Unix-based systems this config path will be: $HOME/.swarm-cli

On Windows systems: $HOME\AppData\swarm-cli

The configuration file is saved with 600 file permission.

On first run, this configuration will be generated with default values, that you are able to change on your demand under the before mentioned path.

Assignment priority

It is possible to set value of particular parameters in different ways.

The assignment priority of how option gets its value in question is the following:

  1. passed CLI option value (with e.g. --option-example-1)
  2. env variable for that option in form of either OPTION_EXAMPLE_1 or SWARM_CLI_OPTION_EXAMPLE_1 (if it is available)
  3. CLI configuration value of that option (if it is available)
  4. option's default fallback value (or it is required to define by #1)

System environment

With specific system environment variables you can alter the behaviour of the CLI

  • BEE_API_URL - API URL of Bee client
  • SWARM_CLI_CONFIG_FOLDER - full path to a configuration folder
  • SWARM_CLI_CONFIG_FILE - configuration file name, defaults to config.json

Development

After the project has been cloned, the dependencies must be installed. Run the following in the project folder:

 $ npm install

Then you need to compile the TypeScript code:

 $ npm run build

To make the local swarm-cli files in the dist/ directory available as a global package:

 $ npm link

If all went well you should be able to run swarm-cli.

If npm link fails, or you don't want to install anything, then you can use node dist/src/index.js to run swarm-cli from the checked out directory.

Contribute

There are some ways you can make this module better:

  • Consult our open issues and take on one of them
  • Help our tests reach 100% coverage!

Maintainers

See what "Maintainer" means here.

License

BSD-3-Clause