β Star us on GitHub β it motivates us a lot! π
Scramjet Transform Hub is a cool way to run programs on your computer or in the cloud. It's easy to use, and it supports lots of different types of programs.
The Hub is the heart of Scramjet Cloud Platform, a serverless data processing system that allows you to easily deploy, run and interconnect programs that process repetitive data tasks in long-running processes. STH can be run just as well on a Raspberry Pi as it can on a massive 128 core Epyc bare metal server. It installs in one simple command and deploys your app to processes, Docker containers or Kubernetes clusters just as easily... Oh, and the apps are incredibly simple to write too, here's a slack posting sample:
export (input: Readable) => async function*() {
for (const post of input) {
yield await slack.note(`New post titled "${post.title}" from ${post.author}`);
}
}
It currently supports Node.js and Python based apps, but Java and other languages are coming.
Get Scramjet Transform Hub straight from NPM
If you don't have a server ready or want to run something from the cloud itself, do check out our website and try the hosted version of Scramjet Cloud Platform.
- Introduction π€
- Usage π‘
- The basics π‘
- Development instructions π·
- Sample usage π
- Troubleshooting π₯
- Start your 30 day trial π₯³
- License and contributions π
- Help wanted π
- Donation πΈ
This is a development repo for Scramjet Transform Hub, a container supervisor that allows deployment, execution and monitoring of any application based on a simple interface.
This repository contains the source code for Scramjet Transform Hub.
If you want to read more about the usage please go to Quick Start repo.
You may also be interested in our Scramjet Cloud Platform that is based on Transform-Hub.
What's more, this repository is intended for developers who would like to:
- participate in the Scramjet Transform Hub development community
- register feature requests, issues and PRs for this product
- build Transform Hub from source
- dive into code and see how it works
For developers who would like to simply use Scramjet Transform Hub to run data processing applications, we recommend following resources:
Repository with Introductory README and templates
NPM page for main scramjet/sth package
NPM page for scramjet/cli package
The easiest way to start Transform Hub is to run the follow (you'll need node.js and npm installed):
npx @scramjet/sth --runtime-adapter=<docker|process|kubernetes>
Scramjet Transform Hub is a platform that helps you run programs on local and remote machines. You can start programs in the background or connect to them and see their output. You can also pipe your data to the program even if it is running on a different machine. You can install this platform in different places like AWS, Google Cloud, Azure or even your own computer.
There's no limit to what you can use it for. Do you want a stock checker? A chatbot? Maybe you'd like to automate your home? Retrieve sensor data? Maybe you have a lot of data and want to transfer and wrangle it? You have a database of cities and you'd like to enrich your data? You do machine learning and you want to train your set while the data is fetched in real-time? Or perhaps you want to use it for something else and ask us if that's a good use? Ask us via email π§ or hop on our Scramjet Discord and we'll be happy to talk with you.
NOTE: This is the STH development repo and in order to use it, you need to have Unix/Linux based operating system, for example Ubuntu or Windows with WSL2. We also use docker and node.js v16.x. We're working on development guides for Mac and Windows. π
Scramjet Transform Hub allows you to deploy and execute programs that you build and develop. As mentioned above, you can run any program you like, but you need to know a couple of important things:
- The program should consist of a function or an array of functions, such a program is called a Transform Sequence.
- The Sequence will be executed within a separate docker instance (π we're working on other execution environment integrations - help will be appreciated π¦Ύ).
- The Sequence function will receive a stream as input in the first argument - you can send the data to it via the command
si instance input
. - If your Sequence contains more than one function, then the output from the first function is passed to the next one. Also, the first function in Sequence receives the input from API.
- The functions can also be Generators and iterate over input data in
node
andpython
versions. This is the simplest way to process data, but it won't leverage asynchronous processing. Check out our frameworks to run your transforms faster Scramjet Framework - The last (or the only) function in Sequence can return a
Promise
or aStream
- based on this, STH will know when processing is done. - Once the returned
Promise
orfuture
is resolved, or theStream
is ended, STH will gracefully stop the Sequence and remove its container. - You can communicate with the server via API, command line client
si
which we wrote for your convenience. - The Sequence is called with an AppContext as
this
orself
, a class that allows you to communicate back from the Sequence: send logs, provide health info, send and receive events from the API or CLI. - You can run your Sequence multiple times with different arguments (like for instance currency tickers with different symbols or sensor data readers for each sensor)
- The program does not leave your server and doesn't use any external systems. It runs on the server you install the host on.
- Currently STH supports node.js and Python runners only, we're working on bringing you runners for other languages, with Java and C++ as the first ones.
We hope this helps you understand the basics of STH and get started with it. If you have any questions or need help, please don't hesitate to contact us!
Some important links π:
- All the documentation can be found on our documentation portal docs.scramjet.org.
- Here you can find the definition of the Transform Sequence AppContext
- You can see the Scramjet Transform Hub API docs here
- You can see the CLI documentation here, but
si help
should also be quite effective. - Don't forget to β this repo if you like it,
subscribe
to releases and keep visiting us for new versions and updates. - You can open an issue - file a bug report or a feature request here
This section contains more detailed installation descriptions, it is intended for developers who would like to contribute or build Transform Hub from source. If you wish to fire up STH quickly and without unnecessary installations, go to Quick Start installation section, which focuses on starting STH and deploying a sample application.
In this section we will also show you some useful commands, tools and functionalities that you can use to develop your own programs.
If you want to help out, we're happy to accept your pull requests. Please follow the below information to start development.
Reminder: This is a development guide. In order to use it, you need to have a Unix/Linux based os (e.g. Ubuntu), docker and node.js v16.x installed. We're working on development guides for Mac and Windows. π
There are several installations you need to perform to get STH up and running, and even more to start developing with us. You may already have some of these below installed, but we will show you how to install them anyway.
- nvm
- node.js
- yarn
- typescript
- ts-node
- docker
- pip
- wget
- jq
Our hub is based on node.js, so you need to install node.js and npm, which is the default package manager for node.js. To check if you already have Node.js(v16.xx.x) and npm installed, run the following commands in your console:
node -v
npm -v
If none of the above commands is found, you must proceed with the installation.
It is recommended to use a Node version manager like nvm to install Node.js and npm. Do it by running the install script using either of the commands bellow.
Note: Make sure you replace v0.39.0 with the latest version of nvm. The actual version of nvm to install is here.
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.0/install.sh | bash
wget -qO- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.0/install.sh | bash
After executing the above command you should see:
which means that you need to run those two more commands to finish the nvm installation:
export NVM_DIR="$HOME/.nvm"
[ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh"
[ -s "$NVM_DIR/bash_completion" ] && \. "$NVM_DIR/bash_completion"
Export lines from above will be saved in your ~/.bashrc file so that you won't need to run it with every session start.
Now you are ready to install node.js, simply type in your console:
nvm install --lts # command will install latest LTS Version of Node.js
π‘ Note: The project is working on Node Long Term Support (LTS) Version, which contains Node Package Manager (NPM) in
^8.15.0
version. NodeJS in version^17.XX.X
will install NPM in version^8.1.2
and we don't use it right now π.
π€ For more info visit the node.js official webpage.
Now you can check installed versions of node.js and npm, run the following commands in your console:
$ node -v
v16.13.0
$ npm -v
8.15.0
OK! It looks like you have successfully installed node.js and npm.
There are two more installations you need to perform, run the following commands in your console one after another:
npm install -g yarn # yarn is a package manager
The same as before the installations can be confirmed by checking the installed versions:
$ yarn -v
1.22.17
OK! The installation was successful. π π
STH can run python packages, and for that it needs to have python
and pip
. These should be already installed on your system - check with:
python3 --version
pip --version
If they are not present, refer to the official installation guide for Python and Pip.
Additionally you have to have wget
and jq
installed:
# Debian based distributions:
apt install wget jq
# Mac
brew install wget jq
If you don't want to use Docker, please skip this step!
We also work with Docker, but this is optional. Running STH is possible without Docker in the background. If you want to use Docker, you can install it by running the following commands in your console:
On Mac u can use Docker-Desktop
# Install docker and docker-compose
sudo apt install -y docker.io docker-compose
# Verify installations by getting versions
sudo docker version
docker-compose version
# Add you user to the docker group
sudo gpasswd -a $USER docker
There are two ways to install STH, either by getting the package from npmjs or by cloning repo.
First, let's install current STH from the npm registry -> @scramjet/sth:
npm i -g @scramjet/sth # installs the package globally
When you start STH from npm, no package installations or builds are required. You can start STH right away, with one of those two commands: scramjet-transform-hub
or sth
. This was quick, wasn't it?:D
Second way is designed for developers who want to contribute and not just use STH.
Let's clone the transform-hub repo. After that, some more installations will be required. Please copy the following commands to the terminal:
git clone https://github.com/scramjetorg/transform-hub.git && \
cd transform-hub && \
yarn install && yarn build:all && npm i -g ./dist/cli
Depending on your machine this may take some time, so it is a perfect time for another hot beverage β or walk πΆπΌββοΈ or joggling π€ΉββοΈ or push-ups maybe..? no? Then simply wait π§.
In the meantime let me describe for you what is happening in the command you have just pasted into the console:
git clone https://github.com/scramjetorg/transform-hub.git
is cloning STH repository.cd transform-hub
is changing the directory to STH repository.yarn install
is installing all the dependencies of STH.yarn build:all
is building all STH packages, this script includes three other building scripts (yarn build:packages && yarn build:refapps && yarn build:docker).npm i -g ./dist/cli
is installing STH CLI as a global command.
When both the package installation and build are complete, STH should be ready to run.
STH can be started in multiple ways. The default way is to run it with Docker and on localhost and port number 8000. Choose one of the following:
If installed by cloning the repo:
yarn start # Starts Hub after it's been built using script
yarn start:dev # Starts Hub in development mode
If installed as npm package:
scramjet-transform-hub # Starts Hub from installed npm package (@scramjet/sth)
sth # alias for scramjet-transform-hub
You can also start STH without Docker, use the same commands as above but add --no-docker
flag:
yarn start --no-docker
yarn start:dev --no-docker
sth --no-docker
There is a wide range of options that you can start STH with. Please add --help
or -h
flag to list all the options:
yarn start --help
yarn start:dev --help
sth --help
$ sth --help
Usage: sth [options]
Options:
-L, --log-level <level> Specify log level (default: "debug")
-S, --socket-path <socket> CSI socket location
-P, --port <port> API port
-H, --hostname <IP> API IP
-E, --identify-existing Index existing volumes as sequences (default: false)
-C, --cpm-url <host:ip>
--id <id>
--runner-image <image name> Image used by runner
--runner-max-mem <mb> Maximum mem used by runner
--prerunner-image <image name> Image used by prerunner
--prerunner-max-mem <mb> Maximum mem used by prerunner
--expose-host-ip <ip> Host IP address that the Runner containers port is mapped to.
--runtime-adapter <type> Run all the instances on the given type of container or on host machine. Possible value: `process`, `docker`, `kubernetes`. Keep in mind that using `process` is UNSAFE FOR RUNNING ARBITRARY CODE.
--sequences-root Only works with --runtime-adapter=process option. Where should ProcessSequenceAdapter save new sequences
-h, --help display help for command
CLI installation was already done at the end of the Installation section, just before starting STH. But it is worth mentioning, that there are two ways to install the CLI:
- from the source code. In the root folder, after building, run the commands:
npm i -g ./dist/cli # install CLI globally from the source folder
si --help # show CLI commands
- from npm:
npm i -g @scramjet/cli # install CLI globally from npm
si --help # show CLI commands
Running si --help
command will confirm that the installation went properly and also to see the list of available commands:
$ si --help
Current profile: default
Usage: si [options] [command]
This is a Scramjet Command Line Interface to communicate with Transform Hub and Cloud Platform.
Options:
-v, --version Display current CLI version
--config <name> Set global configuration profile
--config-path <path> Set global configuration from file
-h, --help display help for command
Commands:
hub Allows to run programs in different data centers, computers or devices in local network
config|c Config contains default Scramjet Transform Hub (STH) and Scramjet Cloud Platform (SCP) settings
sequence|seq Operations on a Sequence package, consisting of one or more functions executed one after another
instance|inst [command] Operations on the running Sequence
topic Manage data flow through topics operations
completion completion operations
util|u Various utilities
To find out more about CLI, please check out our docs at https://hub.scramjet.org/docs/cli
Read more about Scramjet at https://scramjet.org/ π
You can also run CLI commands from the source code using yarn start:dev:cli
instead of si
, in this case no installation is needed, eg.:
$ yarn start:dev:cli --version
yarn run v1.22.17
$ ts-node packages/cli/src/bin/index.ts --version
CLI version: 0.28.1
We will use CLI later on to execute the Sequence. If you would like to read more about our command line interface jump right away to CLI docs in this repo or explore CLI Reference section on our official website www.scramjet.org.
All the packages in the project need to be installed and built before they can be used.
Install will create a node_modules
folder in the root of the project and in each package (dependencies will be installed).
Build will create a dist
folder in the root of the project and in each package. The folder will contain the files with the code compiled into JavaScript.
This is how to perform a clean install and build of all the packages:
yarn install:clean # this command will perform 'yarn clean && yarn clean:modules && yarn install' at once
yarn build:packages # optionally 'build:all' if you want all dockerfiles.
π‘ HINT: For more scripts please see
"scripts: {}"
in main package.json.
During development, some artifacts may be leftover in docker. See how to clean them:
docker ps # list containers
docker volume prune -f # remove all unused local volumes
docker system prune --all -f # remove all unused images not just dangling ones
docker stop $(docker ps -a -q) # stops all running containers
*(
-f
) - doesn't prompt confirmation
Build from current source:
cd ./packages/sth/
yarn build:docker
cd ./packages/sth/
docker-compose up
# or run in detached mode
docker-compose up -d
docker-compose logs -f
To run Hub without docker-compose:
docker run \
-ti \
--name scramjet-hub \
--init \
-p 8000:8000 \
-v /tmp/:/tmp/ \
-v /var/run/docker.sock:/var/run/docker.sock \
scramjetorg/sth:$(jq -r .version package.json)
We use our own scripts to control our monorepo. Here's a couple of helpful commands, which might be useful during development:
./scripts/run-script.js [script]
# Run an npm script in each package that contains that script.
./scripts/run-script.js --scope @scramjet/<package_name> <script-name>
# Run script only in one package
./scripts/run-script.js --workspace packages <script-name>
# Run script in all packages in workspace
./scripts/run-script.js --help
# Check this out for additional information
Completion script depends on bash-completion so make sure it's already installed by running type _init_completion
.
Below command installs completion script in ~/.bashrc.
si completion install
For immediate effect make sure to run source ~/.bashrc
. Also you can manage completion script by yourself. Running si completion bash
prints script to the terminal.
There are two important branches: main
(representing stable version of the
project) and devel
(where most of the work happens; this is the default
branch).
- New features, and especially breaking changes, should go to the
devel
branch. - Bugfixes, especially ones that should be published in a bugfix release,
should be merged into the
main
branch.- Start such bugfix branches from
main
, and make PRs tomain
. - After a bugfix is merged,
main
should be merged intodevel
to get the bugfix there as well.
- Start such bugfix branches from
The purpose of having two branches is to keep important bugfixes separate from new features, so that a bugfix version (a release without new features, just bugfixes) can be published immediately if needed.
Further actions will work only if you have properly configured your environment, installed hub and built all the packages. By this time you should already have all those things done by going through the Installation section.
π‘ HINT: The following instructions apply to the state of the repository from the
release/0.28.0
.
- Run STH:
yarn start # if cloned STH repo
sth # if installed STH as npm package
Open a new terminal window inside transform-hub root folder and do the following:
- Send the hello-world Sequence package to the STH and start it:
si seq send packages/hello-world.tar.gz
si seq start - # '-' means last uploaded Sequence
si seq start <id> # or you can specify id parameter instead
Alternatively, you can use deploy which does 3in1 - it packs (if neccessary), sends and starts the Sequence:
si seq deploy packages/hello-world.tar.gz
You should see on the STH terminal that it received the Sequence.
Let's connect to the output stream. Run this command:
si inst output - # '-' means last uploaded Sequence
si inst output <id> # or you can specify id parameter instead
Note that we need an instance ID, not the Sequence. If the output from previous command have been cleared on your terminal, you can still list all your Sequences and instances by running the following:
- List Sequences:
si seq list
- List instances:
si inst list
After using instance's ID with si inst output
command, you should see "Hello World!" on your terminal window.
See more about streams and curl commands =>
To check out more of our ready-to-go samples, please go to our Samples repo on GitHub, or to our Samples Hub.
We have also prepared a template for you to use. You can use it as a base for your own samples π sample template. For this moment we support two variants of template in two programming languages:
There is also C++ template that is still in development stage:
Error: connect ENOENT /var/run/docker.sock
During sending the Sequence compressed package to the host you may come across this error:
-
Error: connect ENOENT /var/run/docker.sock
2022-01-13T11:54:26.948Z info (object:Host) New sequence incoming... 2022-01-13T11:54:26.948Z log (object:DockerSequenceAdapter) DockerSequenceAdapter init. 2022-01-13T11:54:26.948Z log (object:DockerodeDockerHelper) Checking image scramjetorg/pre-runner:0.14.0 2022-01-13T11:54:26.949Z log (object:DockerodeDockerHelper) Start pulling image scramjetorg/pre-runner:0.14.0 2022-01-13T11:54:26.949Z debug (object:Host) Error: connect ENOENT /var/run/docker.sock at PipeConnectWrap.afterConnect [as oncomplete] (node:net:1161:16) 2022-01-13T11:54:26.950Z debug (object:Host) Request date: 2022-01-13T11:54:26.948Z, method: POST, url: /api/v1/sequence, status: 422
To solve this issue you need to install docker and docker-compose. You can install them using the following command below or refer to official Docker instructions.
```bash
sudo apt install -y docker.io docker-compose
```
Error: connect EACCES /var/run/docker.sock
During sending the Sequence compressed package to the host you may come across this error:
-
Error: connect EACCES /var/run/docker.sock
2022-01-13T11:58:00.368Z info (object:Host) New sequence incoming... 2022-01-13T11:58:00.369Z log (object:DockerSequenceAdapter) DockerSequenceAdapter init. 2022-01-13T11:58:00.370Z log (object:DockerodeDockerHelper) Checking image scramjetorg/pre-runner:0.14.0 2022-01-13T11:58:00.375Z log (object:DockerodeDockerHelper) Start pulling image scramjetorg/pre-runner:0.14.0 2022-01-13T11:58:00.377Z debug (object:Host) Error: connect EACCES /var/run/docker.sock at PipeConnectWrap.afterConnect [as oncomplete] (node:net:1161:16) 2022-01-13T11:58:00.379Z debug (object:Host) Request date: 2022-01-13T11:58:00.373Z, method: POST, url: /api/v1/sequence, status: 422
To solve this issue you need to add the current user to the docker group, please use this command:
sudo gpasswd -a $USER docker
If the problem persists, restart your PC.
Errors related to build packages
If something goes wrong during building packages, any errors occur, please try to run clean build, which will remove all node_modules
and dist
directories, after that you try to install and build them again.
yarn clean && yarn clean:modules && yarn install && yarn build:all
Errors related to reference-apps package
Every reference-app package before we run it, needs to:
- have
node_modules
installed (yarn install
) - have
dist
directory created and tar.gz package created (yarn build:refapps
)
Remember to install dependencies and build your sample package before compressing it.
If you create your sample in packages/reference-apps
folder, you can use the following command to build it:
yarn build:refapps
It will build all the packages in the packages/reference-apps
folder.
π‘ HINT: Have a look at the root
package.json
, there is thescripts
section, which contains the list of all the scripts you can run with yarn. You may find them useful.
Log an issue/bug every time you encounter a problem or find a bug. Maybe you will also find that some feature is missing?
This project is dual licensed under the AGPL-3.0 and MIT licenses. Parts of the project that are linked with your programs are MIT licensed, the rest is AGPL.
We accept valid contributions and we will be publishing a more specific project road map so contributors can propose features and also help us implement them. We kindly ask you that contributed commits are Signed-Off git commit --sign-off
.
We provide support for contributions via test cases. If you expect a certain type of workflow to be officially supported, please specify and implement a test case in Gherkin
format in bdd
directory.
More info about bdd tests can be found here.
The project need's your help! There's lots of work to do and we have a lot of plans. If you want to help and be part of the Scramjet team, please reach out to us, on discord or email us: opensource@scramjet.org.
Do you like this project? It helped you to reduce time spent on delivering your solution? You are welcome to buy us a coffee β
-
There's also a Paypal donation link if you prefer that: