MidjourneyAPI

MidJourneyAPI is as its name suggests, an API for MidJourney as of now (14 june 23)
MidJourney has no API for public use.

This project is a personal project to make a tool that makes it quick and easy to use MidJourney in the terminal.

How it works

Your machine will create a session with Discord. MidjourneyAPI will then send messages to the Midjouney bot on behalf of you. Download the generated images and crop them into individual images. (Midjourney makes a 2x2 image from one prompt)

TODO:

• Figure out what the api session is
• Make documentation for the UML
• Say what secrets needs to be present
• Make the includeURL automatically detect the links ???
• Description on how MidJourneyAPI and MidJourney work
• Design choce
  • abstract class for interface use https://stackoverflow.com/questions/31457855/cant-instantiate-abstract-class-with-abstract-methods
  • Make the relations correct with the amount relation as well in the UML

Prerequisites

• websockets library (pip3 install websockets)
• Your keys (found in the 'Keys' folder)
  • API_channel_id.key
  • API_guild_id.key
  • API_session_id.key
  • DiscordServerID.key
  • authToken.key
• A Discord Server that you can manage with Midjourney's bot joined

Development Design

classDiagram
    class Config{
        <<singleton>>
        - dict : Keys
        - dict : Global
        - dict : Env
        # ProductionType : chosenProd
        - String : session_id 
        - openFile(String : path) : String
        - parseStrJSON(String : str) : dict
        - loadKey(dict : srcDict, dict : dict : dict)
        - loadAllKeys(srcDict : dict) 
        + Key(String : keyName) : object
        + Global(String : keyName) : object
        + Env()
    }

    class NetworkManager{
        + WebsocketConnection wsConnection
        + POST(String : url, dict : headers, dict : data) : Response
        + MJsplit(String : discordImgUrl) : dict
        + MJdownloadPictures(dict : pictureUrls, String : downloadPath)
        + connectToWs(String : url)
        + generateSnowflake() : Int$
    }

    class IPacketAssembler{
        <<interface>>
        + createApplicationCommand() : dict
        + createPrelim() : dict
        + createDataPacket(String : prompt, String : includeImgURL) : dict
        + AssemblePacket(String : prompt, String : includeImgURL)
    }

    class PacketManager{
        AssembleImagine(String : prompt, List~String~ : includeImgUrl) : dict
    }
    class ImaginePacket {
        - uri_validator(String url) : bool
    }

    class InfoPacket {
    }

    class WebsocketConnection {
        + String reconnectURL
        + Int heartbeatInterval 
        + String session_id
        + int lastSeq
        + Connection cConnection
        + connect(String : url) : ClientConnection
        + maintainHeartbeat(int : heartbeatInterval)
    }

    class Opcodes{

    }
    class MidjourneyAPI {
        + Imagine(String : prompt, String : downloadPath)
    }
    class TerminalController {
        - Queue requestQueue 
        - String TerminalIcon
        + getInput(String : inputPromt) : String
    }

    PacketManager *-- MidjourneyAPI
    NetworkManager *-- PacketManager
    NetworkManager --* WebsocketConnection

    TerminalController  -->  MidjourneyAPI
    PacketManager --* ImaginePacket
    PacketManager --* InfoPacket
    InfoPacket  ..> IPacketAssembler 
    ImaginePacket ..> IPacketAssembler

Loading

this is an example of the request sent when you are connected to discord and whant to send an imagine command.
You will find the

• guild_id : the first number after the 'channels/' in the URI
• channel_id: : the last number in the URI.
• API_session_id : just create the file and leave it empty
• authToken : Fetched when a session with Discord is opened.

An example of Discords API payload, which is integrated in the terminal of your choice on your machine:

"payload_json":{
    "type": 2,
    "application_id": "936929561302675456",
    "guild_id": "",
    "channel_id": "",
    "session_id": "",
    "data": {
        "version": "1118961510123847772",
        "id": "938956540159881230",
        "name": "imagine",
        "type": 1,
        "options": [
            {
                "type": 3,
                "name": "prompt",
                "value": "minions walking in a line --q 0.25"
            }
        ],
        "application_command": {
            "id": "938956540159881230",
            "application_id": "936929561302675456",
            "version": "1118961510123847772",
            "default_member_permissions": null,
            "type": 1,
            "nsfw": false,
            "name": "imagine",
            "description": "Create images with Midjourney",
            "dm_permission": true,
            "contexts": [
                0,
                1,
                2
            ],
            "options": [
                {
                    "type": 3,
                    "name": "prompt",
                    "description": "The prompt to imagine",
                    "required": true
                }
            ]
        },
        "attachments": []
    },       
    "nonce": "1119669035626135552"
}

Developed by Phillip Lundin