Skip to content

Latest commit

 

History

History
219 lines (150 loc) · 10.3 KB

README.md

File metadata and controls

219 lines (150 loc) · 10.3 KB

Test and Deploy

mermaid-workflow

Runnable user centric workflow described by mermaid diagrams

Context

Mermaid diagrams is a tool to create complex diagrams in simple markdown. Its main purpose is to help documentation catch-up with development.

Problem statement

I would like to describe how work flows in a simple diagram.

This could be for:

  • My own simple to do list of work.
  • A repeatable pattern / process of work for use in an organisation.
  • To describe how services to end users work and how work flows between consumers of the service and the organisations which are required to fulfil the service e.g. Service design

Constraints and Criteria

I would like to be able to describe my workflow as simply as possible in a mermaid diagram. I want the diagram to be "test-able / run-able" with as little (ideally no) effort as possible.

Live Demo - Swagger UI

Here is a Swagger UI implentation of the OpenAPI rest interface project in mermaid-worflow-api

To use it you can provide a context, chart, and mappings input.

Chart is a mermaid chart (or a URL to a mermaid chart) and mapping is the action-mapping file (or a URL to it) which shows how to map nodes and routes to actions.

An example chart is here https://raw.githubusercontent.com/jonnymoo/mermaid-workflow/main/mermaid-processor/test/resources/lhasa/flowchart.mermaid

An example mapping is here https://raw.githubusercontent.com/jonnymoo/mermaid-workflow/main/mermaid-processor/test/resources/lhasa/mappings.json

These demonstrate the Lhasa example (see below for more details)

To try it out you need to set the field svg-input to be a string containing an svg test file. The workflow will then convert this to SVG and nagivate down the flow chart. Tip - you need to json stringify the input, you can use an online utility to do this e.g. https://jsonformatter.org/json-stringify-online

Here is an example input.

{
  "context":  {"fields": {"svg-input":"<svg xmlns=\"http://www.w3.org/2000/svg\" width=\"600\" height=\"600\">\r\n  <circle cx=\"150\" cy=\"150\" r=\"65\" fill=\"blue\" />\r\n  <circle cx=\"350\" cy=\"250\" r=\"70\" fill=\"blue\" />\r\n  <circle cx=\"250\" cy=\"450\" r=\"25\" fill=\"blue\" />\r\n</svg>"}},
  "chart": "https://raw.githubusercontent.com/jonnymoo/mermaid-workflow/main/mermaid-processor/test/resources/lhasa/flowchart.mermaid",
  "mappings": "https://raw.githubusercontent.com/jonnymoo/mermaid-workflow/main/mermaid-processor/test/resources/lhasa/mappings.json"
}

Expand /chart/process and choose "Try it out".

Paste the input into the request body, choose "Execute" and examine the response body. You will see a number of fields, the one of interest is the score. Also of note is the path taken, and the audit which will show you how the input was processed through the chart.

Enjoy.

If you want to try other tests, have a look at the Lhasa examples below and pick other SVG inputs to try from the test input resources. The number at the end of the file name denotes what the expected score is (apart from random_svg_661-II.svg which gives a score of one back - I think this is just a mistake)

Quick Start Guide for mermaid-processing

Prerequisites

1. Java

Ensure you have Java installed. You can check by running:

java -version

If not installed, download and install the OpenJDK from AdoptOpenJDK's official website.

2. Clojure

Install Clojure using the instructions provided on the official Clojure site.

3. Leiningen

This is a build tool for Clojure. Install it by following the instructions on the official Leiningen website.

Setting Up the Project

1. Clone the Repository

git clone https://github.com/jonnymoo/mermaid-workflow.git
cd mermaid-workflow
cd mermaid-processor

2. Fetch Dependencies

lein deps

3. Run the Tests

lein test

This will execute all the tests in the project. Ensure all tests pass to verify that everything is set up correctly.

Exploring Further: Understanding the Lhasa Test

The Lhasa test, located in /mermaid_processor/test/mermaid_processor/lhasa_test.clj, is a test suite designed by Lhasa here https://github.com/jmaes12345/lhasa-kata

It has been used to show how to provide custom functionality and run an automatic from a mermaid flowchart.

The Mermaid Flowchart

The test begins by reading a Mermaid flowchart from the file resources/lhasa/flowchart.mermaid. This flowchart describes the expected behavior and conditions that SVG files should meet.

(def mermaid 
  (slurp 
   (io/resource "resources/lhasa/flowchart.mermaid")))

(def lhasa-chart 
  (parse/parse-mermaid mermaid))

View the flowchart.mermaid file

Action Map

The action-map is a collection of regex patterns paired with corresponding actions. Each regex pattern is designed to match specific conditions or queries from the Mermaid flowchart. When a match is found, the associated action is executed.

Single File Test

The process-lhasa-example-test function is an example of how to process a single SVG file against the Mermaid flowchart. It uses the behaviors derived from the action map and the parsed Mermaid chart to make decisions according to contents of the SVG file and generate a score.

Processing All SVG Files

The process-lhasa-all-examples-test function is designed to process all SVG files located in the resources/lhasa/inputs/ directory. For each SVG file, the test determines the expected score based on the file's naming convention (e.g., -I.svg indicates a score of 1). It then processes the SVG file using the behaviors and the Mermaid chart and compares the expected score with the actual score.

API Documentation

API Documentation

Note - to build the documentation

cd mermaid-processor
lein codox

Conclusion

For a deeper dive, explore the Lhasa test file in the repository

Solution concepts

journey
    title A simple workflow
    section Implement Simplest Workflow
      Create an API: 5: Me
      Create a Work Tray: 3: Me
    section Persistence
      Create backing stores: 5: Me
    section User Testing
      Observation: 5: Janey
Loading

Bugs and Issues

...

Design guidelines

Use TDD to implement a modular / composable design that promotes simplicity over complexity

Order function parameters by specificity (general to specific), e.g. my-function [context behavior item]

Aim towards a DDD like Ubiquitous Language. This means entities should use real life terms from the business / user domain. This language should be refleced in the code. And on language, a decision has been made to not use british english, but stick with american english, e.g. behavior, not behaviour. This is simply to make it consistent with other code bases.

Dont throw an exception unless it really is an exception. As a rule dont handle exceptions in business logic code. Let the top layer handle it and report it appropriately. The exception to this rule is if you want to decorate the exception with more information and then rethrow it.

Keep any messy code as small as possible and as close to the entry point to the application (the composition root). Messy code is code with is non-deterministric, input driven, needs to access external systems. Do all this first and pass it into your domain logic. This helps you organise you systems with the appropriate level and direction of interconnectivity and lower complexity.

C4Container
    title Container diagram for Workflow Automation System

    Person_Ext(user, "User", "Uses the WorkflowUI to interact with the system.")

    System_Boundary(workflowSys, "Workflow System") {
        Container(webUI, "WorkflowUI", "Web Application", "Provides a reference front-end for the workflow.")
        Container(webAPI, "WorkflowAPI", "Web API", "Handles the business logic and data processing for workflows.")
        ContainerDb(clojars, "Clojars", "Artifact Repository", "Stores the Clojure libraries used in the system.")
        Container(mermaidLib, "mermaid-processing", "Clojure Library", "Processes and renders diagrams.")
    }

    Rel(user, webUI, "Uses", "HTTP/HTTPS")
    Rel(webUI, webAPI, "Sends requests to", "HTTP/HTTPS")
    Rel(webAPI, clojars, "Fetches libraries from")
    Rel(webAPI, mermaidLib, "Uses for diagram processing")

    UpdateElementStyle(user, $fontColor="blue")
    UpdateRelStyle(webUI, webAPI, $textColor="green", $lineColor="green")
    UpdateRelStyle(webAPI, clojars, $textColor="purple", $lineColor="purple")
    UpdateRelStyle(webAPI, mermaidLib, $textColor="orange", $lineColor="orange")
Loading
C4Component
    title Component diagram for mermaid-processing library

    Boundary(mermaidLib, "mermaid-processing") {
        Component(context, "Context", "Stores the state of the running workflow, current node, and other state required by conditions or actions.")
        Component(behavior, "Behavior", "Specifies actions and conditions. Can modify the Context.")
        Component(chart, "Chart", "Represents the parsed Mermaid diagram.")
    }

    Rel(behavior, context, "Modifies")
    Rel(behavior, chart, "Reads state from")
    Rel(chart, context, "Updates state in")
Loading

Azure setup - Mermaid CMS

For development if you want to run a sqlite database you will need to create an empty database in WAL mode. Also you will need to create the media folder manually.

Once you are able to run umbraco you will need to create a login and usync will handle the rest.

mkdir -p /home/site/wwwroot/media 
mkdir -p /home/site/wwwroot/umbraco/Data
sqlite3 /tmp/Umbraco.sqlite.db 'PRAGMA journal_mode=WAL;'
mv /tmp/Umbraco.sqlite.db* /home/site/wwwroot/umbraco/Data