feat（新功能）: 模特腿部编辑功能

fix（修复bug）:
docs（文档变更）:
refactor（重构）:
test(增加测试):

app/api/api_mannequins_edit.py

@@ -0,0 +1,40 @@
+import json
+import logging
+
+from fastapi import APIRouter, HTTPException
+
+from app.schemas.mannequin_edit import MannequinModel
+from app.schemas.response_template import ResponseModel
+from app.service.mannequins_edit.service import MannequinEditService
+
+router = APIRouter()
+logger = logging.getLogger()
+
+
+@router.post("/mannequins_edit")
+def mannequins_edit(request_data: MannequinModel):
+    """
+        模特腿长调整
+        创建一个具有以下参数的请求体:
+        - **mannequins**: mannequins url等信息
+        - **scale**: 大腿小腿比例
+        - **bucket_name**: bucket name
+        - **mannequin_name**: 模特名称
+
+        示例参数：
+        - **{
+          "mannequins": "aida-sys-image/models/male/dc36ce58-46c3-4b6f-8787-5ca7d6fc26e6.png",
+          "scale": [0.75, 0.75],
+          "bucket_name": "test",
+          "mannequin_name": "mannequin_name"
+        }**
+    """
+    try:
+        logger.info(f"mannequins_edit request item is : @@@@@@:{json.dumps(request_data.dict())}")
+        service = MannequinEditService(request_data)
+        data = service()
+        logger.info(f"mannequins_edit response @@@@@@:{json.dumps(data)}")
+    except Exception as e:
+        logger.warning(f"mannequins_edit Run Exception @@@@@@:{e}")
+        raise HTTPException(status_code=404, detail=str(e))
+    return ResponseModel(data=data)

app/api/api_route.py

@@ -8,9 +8,10 @@ from app.api import api_design
 from app.api import api_design_pre_processing
 from app.api import api_generate_image
 from app.api import api_image2sketch
+from app.api import api_mannequins_edit
 from app.api import api_prompt_generation
-from app.api import api_super_resolution
 from app.api import api_recommendation
+from app.api import api_super_resolution
 from app.api import api_test
 
 router = APIRouter()
@@ -28,3 +29,4 @@ router.include_router(api_brighten.router, tags=['api_brighten'], prefix="/api")
 router.include_router(api_query_image.router, tags=['api_query_image'], prefix="/api")
 router.include_router(api_brand_dna.router, tags=['api_brand_dna'], prefix="/api")
 router.include_router(api_recommendation.router, tags=['api_recommendation'], prefix="/api")
+router.include_router(api_mannequins_edit.router, tags=['api_mannequins_edit'], prefix="/api")

app/schemas/mannequin_edit.py

@@ -0,0 +1,8 @@
+from pydantic import BaseModel
+
+
+class MannequinModel(BaseModel):
+    mannequins: str
+    scale: list[float, float]
+    bucket_name: str
+    mannequin_name: str

app/service/mannequins_edit/service.py

@@ -0,0 +1,101 @@
+import cv2
+import mediapipe as mp
+import numpy as np
+from minio import Minio
+
+from app.core.config import MINIO_URL, MINIO_ACCESS, MINIO_SECRET, MINIO_SECURE
+from app.schemas.mannequin_edit import MannequinModel
+from app.service.utils.new_oss_client import oss_get_image, oss_upload_image
+
+minio_client = Minio(MINIO_URL, access_key=MINIO_ACCESS, secret_key=MINIO_SECRET, secure=MINIO_SECURE)
+
+
+class MannequinEditService():
+    def __init__(self, request_data):
+        self.scale = request_data.scale
+        self.image = oss_get_image(oss_client=minio_client, bucket=request_data.mannequins.split('/')[0], object_name=request_data.mannequins[request_data.mannequins.find('/') + 1:], data_type="cv2")
+        self.mannequin_name = request_data.mannequin_name
+        self.bucket_name = request_data.bucket_name
+        if self.image.shape[2] == 4:
+            self.bgr = self.image[:, :, :3]
+            self.alpha = self.image[:, :, 3]
+            self.bgr = cv2.bitwise_and(self.bgr, self.bgr, mask=cv2.normalize(self.alpha, None, 0, 1, cv2.NORM_MINMAX))
+            self.h, self.w, _ = self.bgr.shape
+        else:
+            self.bgr = self.image
+            self.h, self.w, _ = self.bgr.shape
+            self.alpha = None
+
+    def __call__(self, *args, **kwargs):
+        leg_top, leg_bottom = self.attitude_detection()
+        if leg_top and leg_bottom:
+            new_mannequin = self.resize_leg(leg_top, leg_bottom)
+            _, encoded_image = cv2.imencode('.png', new_mannequin)
+            image_bytes = encoded_image.tobytes()
+            req = oss_upload_image(oss_client=minio_client, bucket=self.bucket_name, object_name=f"{self.mannequin_name}.png", image_bytes=image_bytes)
+            return req.bucket_name + "/" + req.object_name
+        else:
+            return "No leg detected"
+
+    def attitude_detection(self):
+        mp_pose = mp.solutions.pose
+        pose = mp_pose.Pose()
+
+        # 将 BGR 图像转换为 RGB 格式
+        image_rgb = cv2.cvtColor(self.bgr, cv2.COLOR_BGR2RGB)
+        leg_top, leg_bottom = None, None
+        # 进行姿态检测
+        results = pose.process(image_rgb)
+        if results.pose_landmarks:
+            # 获取腿部关键点
+            landmarks = results.pose_landmarks.landmark
+
+            # 找到腿部上边界和下边界
+            leg_top = int(landmarks[mp_pose.PoseLandmark.LEFT_HIP].y * self.h)
+            leg_bottom = int(max(landmarks[mp_pose.PoseLandmark.LEFT_ANKLE].y,
+                                 landmarks[mp_pose.PoseLandmark.RIGHT_ANKLE].y) * self.h)
+
+        return leg_top, leg_bottom
+
+    def resize_leg(self, leg_top, leg_bottom):
+        # 上半身
+        top_part_bgr = self.bgr[:leg_top, :]
+        top_part_bgr_alpha = self.alpha[:leg_top, :]
+
+        # 小腿
+        part_thigh = self.bgr[leg_top:leg_bottom, :]
+        part_thigh_alpha = self.alpha[leg_top:leg_bottom, :]
+
+        # 大腿
+        part_calf = self.bgr[leg_bottom:, :]
+        part_calf_alpha = self.alpha[leg_bottom:, :]
+
+        new_thigh_height = int((leg_bottom - leg_top) * self.scale[0])
+        new_calf_height = int((self.h - leg_bottom) * self.scale[1])
+
+        resized_thigh = cv2.resize(part_thigh, (self.w, new_thigh_height), interpolation=cv2.INTER_LINEAR)
+        resized_thigh_alpha = cv2.resize(part_thigh_alpha, (self.w, new_thigh_height), interpolation=cv2.INTER_LINEAR)
+        resized_calf = cv2.resize(part_calf, (self.w, new_calf_height), interpolation=cv2.INTER_LINEAR)
+        resized_calf_alpha = cv2.resize(part_calf_alpha, (self.w, new_calf_height), interpolation=cv2.INTER_LINEAR)
+
+        new_bgr = np.vstack((top_part_bgr, resized_thigh, resized_calf))
+        new_bgr_alpha = np.vstack((top_part_bgr_alpha, resized_thigh_alpha, resized_calf_alpha))
+
+        if self.alpha is not None:
+            # 拼接 alpha 通道
+            # 合并 BGR 通道和 alpha 通道
+            new_image = np.dstack((new_bgr, new_bgr_alpha))
+        else:
+            new_image = new_bgr
+        return new_image
+
+
+if __name__ == '__main__':
+    request_data = MannequinModel(
+        mannequins="aida-sys-image/models/male/dc36ce58-46c3-4b6f-8787-5ca7d6fc26e6.png",
+        scale=[0.75, 0.75],
+        bucket_name="test",
+        mannequin_name="mannequin_name"
+    )
+    service = MannequinEditService(request_data)
+    print(service())