回答

流程就是通过http的调用prompt接口提交生成图片的请求

1、提交生成图片请求接口

http://127.0.0.1:8188/

POST /prompt

prompt参数

{
  "client_id": "533ef3a3-39c0-4e39-9ced-37d290f371f8",
  "prompt": {
    "3": {
      "inputs": {
        "seed": 764714814161513,
        "steps": 26,
        "cfg": 5,
        "sampler_name": "dpmpp_3m_sde_gpu",
        "scheduler": "karras",
        "denoise": 1,
        "model": [
          "40",
          0
        ],
        "positive": [
          "49",
          0
        ],
        "negative": [
          "6",
          0
        ],
        "latent_image": [
          "5",
          0
        ]
      },
      "class_type": "KSampler"
    },
    "5": {
      "inputs": {
        "width": 1024,
        "height": 768,
        "batch_size": 1
      },
      "class_type": "EmptyLatentImage"
    },
    "6": {
      "inputs": {
        "text": "",
        "clip": [
          "40",
          1
        ]
      },
      "class_type": "CLIPTextEncode"
    },
    "8": {
      "inputs": {
        "samples": [
          "3",
          0
        ],
        "vae": [
          "40",
          2
        ]
      },
      "class_type": "VAEDecode"
    },
    "9": {
      "inputs": {
        "filename_prefix": "ComfyUI",
        "images": [
          "8",
          0
        ]
      },
      "class_type": "SaveImage"
    },
    "13": {
      "inputs": {
        "clip_vision": [
          "39",
          0
        ],
        "image": [
          "34",
          0
        ]
      },
      "class_type": "CLIPVisionEncode"
    },
    "19": {
      "inputs": {
        "strength": 1,
        "noise_augmentation": 0,
        "conditioning": [
          "42",
          0
        ],
        "clip_vision_output": [
          "13",
          0
        ]
      },
      "class_type": "unCLIPConditioning"
    },
    "34": {
      "inputs": {
        "image": "clipspace/clipspace-mask-1645940.7000000002.png [input]",
        "choose file to upload": "image"
      },
      "class_type": "LoadImage"
    },
    "36": {
      "inputs": {
        "clip_vision": [
          "39",
          0
        ],
        "image": [
          "38",
          0
        ]
      },
      "class_type": "CLIPVisionEncode"
    },
    "37": {
      "inputs": {
        "strength": 0.75,
        "noise_augmentation": 0,
        "conditioning": [
          "19",
          0
        ],
        "clip_vision_output": [
          "36",
          0
        ]
      },
      "class_type": "unCLIPConditioning"
    },
    "38": {
      "inputs": {
        "image": "beijing1 (2).webp",
        "choose file to upload": "image"
      },
      "class_type": "LoadImage"
    },
    "39": {
      "inputs": {
        "clip_name": "clip_vision_g.safetensors"
      },
      "class_type": "CLIPVisionLoader"
    },
    "40": {
      "inputs": {
        "ckpt_name": "sd_xl_base_1.0.safetensors"
      },
      "class_type": "CheckpointLoaderSimple"
    },
    "42": {
      "inputs": {
        "conditioning": [
          "6",
          0
        ]
      },
      "class_type": "ConditioningZeroOut"
    },
    "43": {
      "inputs": {
        "safe": "enable"
      },
      "class_type": "HEDPreprocessor"
    },
    "44": {
      "inputs": {
        "safe": "enable",
        "image": [
          "34",
          0
        ]
      },
      "class_type": "HEDPreprocessor"
    },
    "45": {
      "inputs": {
        "images": [
          "44",
          0
        ]
      },
      "class_type": "PreviewImage"
    },
    "46": {
      "inputs": {
        "control_net_name": "control-lora-depth-rank256.safetensors"
      },
      "class_type": "ControlNetLoader"
    },
    "47": {
      "inputs": {
        "image": [
          "34",
          0
        ]
      },
      "class_type": "ScribblePreprocessor"
    },
    "48": {
      "inputs": {
        "images": [
          "47",
          0
        ]
      },
      "class_type": "PreviewImage"
    },
    "49": {
      "inputs": {
        "strength": 0.5,
        "conditioning": [
          "37",
          0
        ],
        "control_net": [
          "46",
          0
        ],
        "image": [
          "47",
          0
        ]
      },
      "class_type": "ControlNetApply"
    }
  }
}

client_id就是你自己传给comfyui的一个参数，这个在后来的comfyui通知你生成图片的进度的时候有用

2、实时获取任务进度

websocket：/ws?client_id=第一步的client_id

当获取绘图结束的通知消息的时候，消息如下：

{
    "type":"executing",
    "data":{
        "node":null,
        "prompt_id":"37099310-a790-44f4-8d13-41111232"
    }
}

3、调用GET /history/{prompt_id}获取输出的图片信息

{
    "37099310-a790-44f4-8d13-41111232": {
    	略。。。。。。。。。。
        "outputs": {
            "18": {
                "images": [
                    {
                        "filename": "ComfyUI_temp_slqio_00001_.png",
                        "subfolder": "",
                        "type": "temp"
                    },
                    {
                        "filename": "ComfyUI_temp_slqio_00002_.png",
                        "subfolder": "",
                        "type": "temp"
                    },
                    {
                        "filename": "ComfyUI_temp_slqio_00003_.png",
                        "subfolder": "",
                        "type": "temp"
                    },
                    {
                        "filename": "ComfyUI_temp_slqio_00004_.png",
                        "subfolder": "",
                        "type": "temp"
                    }
                ]
            },
           
        }
    }
}

1其中filename就是生成图片的名称

4、显示图片

调用这个url http://127.0.0.1:8188/view?filename=ComfyUI_00705_.png&type=output

就能显示这个图片了