UniTao

Main Repo Moved To

https://github.com/TuringCompute/UniTAO

NOTE from Author (Yi Huo):

Apparently, This opensource project is terminated by salesforce.

All further development will be fully based on my own requirement now.

Thus, I decided to fork this project to my own Organization (TuringCompute) for further development

https://github.com/TuringCompute

The new repo will be a fork to this repo.

https://github.com/TuringCompute/UniTAO

all future development will happens to that repo first.

I will try to sync all future development back here, as long as I have the permission to do so.

Description

UniTAO was originally created in 2022 by Shai Herzog & Yi Huo as an Universal No-Coding Heterogeneous Infrastructure Maintenance & Inventory system that is holistically driven by open/community-developed semantic models/schemas.

It is designed to be self-contained and self-documenting to serve as foundation for human-free automation of heterogeneous computing infrastructure.

The model/schema uses the standard JSON schema format as basis, with additional enhancements. Those can be found in the documentation.

schema of schema is defined in path:

lib/Schema/data/schema.json

Demo

Docker

We can build DataService and Inventory Service into docker container images. so we can run these in a docker environment for demo.

to build all images, run one of the following command based on your target environment.

# Powershell script for Windows environment
./docker/buildAll.ps1

# Bash script for Mac or Linux
./docker/buildAll.sh

Environment

after all docker images build successfully. we can bring a set of docker instances up as demo environment.

Data Service 01
Data Service 01 Admin
Data Service 02
Data Service 02 Admin
Inventory Service
Inventory Service Admin
Database (DynamoDB / MongoDB)
Database Admin (UI interface)

pick a folder and run the command docker compose up -d

# demo environment with DynamoDB
cd ./docker-compose/2data1inv

# demo environment with MongoDB
cd ./docker-compose/2data1invMongo

Run

to run demo:

bring up demo environment. wait all instance running stable after multiple restart for data setup
login to demo folder

./demo

choose a demo folder and run demo script depending on demo environment.

# Windows environment:
demo.ps1

# Mac/Linux environment
demo.sh

Components

Data Layer

Supported Databases: DynamoDb, MongoDb
plugin-able data layer to support multiple types of database
Language: GoLang
sub folder: ./data

REST API Service

Language: GOLANG
Workspace: all project folders are included in go.work
install golang on MacOS

brew install go

follow src/README.md to init and run the Inventory Service

JSON Schema Extensions

the JSON schema of schema to define data format is the key feature of this Data Service. it give enough flexibility to define JSON data in order to support any coding requirement.

in order to achieve certain feature of Data Service, we have extended JSON schema as following:

contentMediaType

the original meaning of contentMediaType is define how to parse value of the attached attribute.

Example:

the following code means value of testAttr is a json string that we can use json library to parse it

{
    "name": "testAttr",
    "type": "string",
    'contentMediaType': 'json'
}

in our schema, we introduced

contentMediaType = [inventory/{dataType}]

The following attribute definition means, value of testId is the id of type test from inventory service

{
    "name": "testId",
    "type": "string",
    'contentMediaType': 'inventory/test'
}

Reason：

in JSON schema, there is already a key $ref that can reference remote schema.

but after close look, we found $ref is just a simple include method. it does not include the meaning of parsing and validation.

so we decide conentMediaType are closer to what we really means here.

that is:

we only specify how to find the data that referenced by the value, but we don't really tell you how to use it.
this will seperate the logic of parsing and using of the schema

indexTemplate

this attribute is for DataService automation in order to auto fill in reference value in registry attribute of other record.

Example

what we want to achieve here is to automatically fill m01 into machines attribute from listA of type machineList when add record machine id=m01

{
    "machine": [
        {
            "id": "m01"
            "listId": "listA"
        }
    ],
    "machineList": [
        {
            "id": "listA",
            "machines": [
                "m01"
            ]
        }
    ]
}

salesforce/UniTAO