代码日常上传归档

config_test.py

@@ -0,0 +1,12 @@
+import pickle
+
+# 假设你的 pickle 文件名为 'data.pkl'
+file_path = '/home/xiazj/ai-station-code/tmp/sam/7afa76b5-6730-434f-9470-651b4bcb7e8f/model_params.pickle'
+
+# 读取 pickle 文件
+with open(file_path, 'rb') as file:  # 以二进制模式打开文件
+    data,config = pickle.load(file)  # 反序列化数据
+
+# 输出读取的数据
+print(data)
+print(config)

dimaoshibie/predict.py

@@ -18,5 +18,5 @@ def predict(img):
     r_image.show()
     
 if __name__ == "__main__":
-    path = "D:\\project\\ai-station\\tmp\\dimaoshibie\\crop_9_14.tif"
+    path = "/home/xiazj/ai-station-code/tmp/dimaoshibie/d94afe94-2fae-4ce5-9ee4-94ac1d699337/dimao2.jpg"
     predict(path)

dimaoshibie/segformer.py

@@ -22,7 +22,7 @@ class SegFormer_Segmentation(object):
         "num_classes": 10 + 1,  # 类别数（包括背景）
         "phi": "b0",  # 模型规模（b0-b5）
         "input_shape": [512, 512],  # 输入图像尺寸
-        "mix_type": 0,  # 可视化方式：0-混合原图，1-仅分割图，2-仅目标区域
+        "mix_type": 1,  # 可视化方式：0-混合原图，1-仅分割图，2-仅目标区域
         "cuda": True,  # 是否使用GPU
     }
 

main.py

@@ -1,15 +0,0 @@
-import sys
-
-from fastapi import FastAPI
-from fastapi.staticfiles import StaticFiles
-version = f"{sys.version_info.major}.{sys.version_info.minor}"
-
-app = FastAPI()
-
-# 将 /root/app 目录挂载为静态文件
-app.mount("/files", StaticFiles(directory="/root/app"), name="files")
-
-@app.get("/")
-async def read_root():
-    message = f"Hello world! From FastAPI running on Uvicorn with Gunicorn. Using Python {version}"
-    return {"message": message}

run.py

@@ -108,7 +108,7 @@ async def read_root():
 # 获取数据界面资源信息
 @app.get("/ai-station-api/index/show")
 async def get_source_index_info():
-    sql = "SELECT id,application_name, describe_data, img_url FROM app_shouye"
+    sql = "SELECT id,application_name, describe_data, img_url,url FROM app_shouye"
     data = data_util.fetch_data(sql)
     if data is None:
         return {
@@ -679,10 +679,13 @@ async def upload_image(file: UploadFile = File(...),type: str = Form(...), ):
     file_location = os.path.join(upload_dir , file.filename)
     with open(file_location, "wb") as buffer:
         shutil.copyfileobj(file.file, buffer)
+    encoded_string = sam_deal.load_tmp_image(file_location)
     file_location = model_deal.get_pic_url(file_location)
     return {"success":True,
             "msg":"获取信息成功",
-            "data":{"location": file_location}}
+            "data":{"location": file_location},
+            "image":JSONResponse(content={"image_data": encoded_string})
+            }
 
 
 """
@@ -698,10 +701,14 @@ async def dmsb_image_analysis(path:str = None):
                 "data":None}
     else:
         path = result['reason']
+        encoded_string = sam_deal.load_tmp_image(path[0])
         path= model_deal.get_pic_url(path[0]) # 是一个列表
         return {"success":True,
                 "msg":"获取信息成功",
-                "data":{"result": path}}
+                "data":{"result": path},
+                "image":JSONResponse(content={"image_data": encoded_string})
+                }
+
 
 
 """
@@ -710,19 +717,24 @@ async def dmsb_image_analysis(path:str = None):
 @app.get("/ai-station-api/dmsb_image/calculate")
 async def dmsb_image_calculate(path:str = None, scale:float = 1.92*1.92):
     path = model_deal.get_pic_path(path)
-    result = model_deal.dimaoshibie_area(path,scale,param.dmsb_colors)
+    # result = model_deal.dimaoshibie_area(path,scale,param.dmsb_colors)
-    logger.info(result)
+    # logger.info(result)
-    if result['status'] == True:
+    # if result['status'] == True:
-        res = result['reason']
+    #     res = result['reason']
-        translated_dict = {param.dmsb_type[key]: value for key, value in res.items()}
+    #     translated_dict = {param.dmsb_type[key]: value for key, value in res.items()}
-        # json_data =  json.dumps(translated_dict)
+    file_directory = os.path.dirname(path)
-        return {"success":True,
+    output_file_path = os.path.join(file_directory,'result.txt')
-                "msg":"获取信息成功",
+    # 初始化一个空字典
-                "data":{"result": translated_dict}}
+    total_piex = {}
-    else:
+    with open(output_file_path, 'r') as file:
-        return {"success":False,
+        for line in file:
-                "msg":result['reason'],
+            # 去掉行首尾的空白字符，并分割键值
-                "data":None}
+            key, value = line.strip().split(': ')
+            # 将值转换为整数并存入字典
+            total_piex[key] = int(value) * scale
+    return {"success":True,
+            "msg":"获取信息成功",
+            "data":{"result": total_piex}}
 
 
 """
@@ -760,17 +772,19 @@ async def download_zip(path:str = None):
 async def roof_image_analysis(path:str = None):
     path = model_deal.get_pic_path(path)
     result = model_deal.roof_pic(roof_model,path,param.wdpv_palette)
-    
     if result['status'] == False:
         return {"success":False,
                 "msg":result['reason'],
                 "data":None}
     else:
         path = result['reason']
+        encoded_string = sam_deal.load_tmp_image(path[0])
         path= model_deal.get_pic_url(path[0])
         return {"success":True,
                 "msg":"获取信息成功",
-                "data":{"result": path}}
+                "data":{"result": path},
+                "image":JSONResponse(content={"image_data": encoded_string})
+                }
 
 
 """
@@ -812,10 +826,12 @@ async def pv_image_analysis(path:str = None):
                 "data":None}
     else:
         path = result['reason']
+        encoded_string = sam_deal.load_tmp_image(path[0])
         path= model_deal.get_pic_url(path[0])
         return {"success":True,
                 "msg":"获取信息成功",
-                "data":{"result": path}}
+                "data":{"result": path},
+                "image":JSONResponse(content={"image_data": encoded_string})}
 
 """
 光伏面积 图 - 分割结果计算 ,图片demo的是z18
@@ -856,10 +872,13 @@ async def roofpv_image_analysis(path:str = None):
     else:
         file_list = result['reason']
         final_path = prepare_data.merge_binary(file_list)
+        encoded_string = sam_deal.load_tmp_image(final_path)
         final_path = model_deal.get_pic_url(final_path) 
         return {"success":True,
                 "msg":"获取信息成功",
-                "data":{"result": final_path}}
+                "data":{"result": final_path},
+                "image":JSONResponse(content={"image_data": encoded_string})
+}
 
 
 """
@@ -995,7 +1014,8 @@ async def get_ch4_predict(path:str=None,start_time:str=None, end_time:str = None
     else:
         return {"success":True,
             "msg":"获取信息成功",
-            "data":data['reason']}
+            "data":data['reason'],
+            'time':end_time}
 
 #========================================光伏预测==========================================================
 """
@@ -1108,6 +1128,7 @@ async def get_pvelectri_predict(path:str=None,start_time:str=None, end_time:str
     else:
         return {"success":True,
             "msg":"获取信息成功",
+            'time':end_time,
             "data":data['reason']}
 
 
@@ -1215,6 +1236,7 @@ csv上传
 @app.get("/ai-station-api/wind_electric_data/predict")
 async def get_wind_electri_predict(path:str=None,start_time:str=None, end_time:str = None,is_show:bool=True):
     data = model_deal.start_wind_electric_predict_endpoint(windfd_model,path,start_time,end_time,is_show)
+    # logger.info(data)
     if data['status'] ==False:
         return {
             "success":False,
@@ -1224,6 +1246,7 @@ async def get_wind_electri_predict(path:str=None,start_time:str=None, end_time:s
     else:
         return {"success":True,
             "msg":"获取信息成功",
+            'time':end_time,
             "data":data['reason']}
 
 
@@ -1305,12 +1328,11 @@ async def upload_sam_image(file: UploadFile = File(...),type: str = Form(...), )
 
 要求：针对返回current_index,将列表默认成选择current_index对应的分类
 """
-@app.get("/ai-station-api/sam_class/create")
+@app.post("/ai-station-api/sam_class/create")
-async def sam_class_set(
+async def sam_class_set(item:post_model.samItem):    
-    class_name: str = None,
+    class_name =  item.class_name
-    color: Optional[List[int]] = Query(None, description="list of RGB color"),
+    color= item.color
-    path: str = None
+    path = item.path
-):    
     loaded_data,api_config = sam_deal.load_model(path)
     result = sam_deal.add_class(loaded_data,class_name,color)
     if result['status'] == True:
@@ -1335,6 +1357,7 @@ async def sam_class_set(
             "data":{"class_name_list": loaded_data['class_names'],
                     "current_index": loaded_data['class_index'],
                     "class_dict":loaded_data['class_colors'],
+                    "color":color,
                     }}
 
 
@@ -1344,12 +1367,11 @@ async def sam_class_set(
 current_index : 下拉列表中的分类索引
 rgb_color : 
 """
-@app.get("/ai-station-api/sam_color/select")
+@app.post("/ai-station-api/sam_color/select")
-async def set_sam_color(
+async def set_sam_color(item:post_model.samItem2):
-    current_index: int = None,
+    current_index = item.current_index
-    rgb_color: List[int] = Query(None, description="list of RGB color"),
+    rgb_color = item.rgb_color
-    path: str = None
+    path = item.path
-    ):
     loaded_data,api_config = sam_deal.load_model(path)
     r, g, b = [int(c) for c in rgb_color]
     bgr_color = (b, g, r)
@@ -1362,12 +1384,14 @@ async def set_sam_color(
         return {
             "success":True,
             "msg":"",
+            "color":rgb_color,
             "image": JSONResponse(content={"image_data": encoded_string}),
         }
     else:
         return {
         "success":False,
         "msg":img['reason'],
+        "color":rgb_color,
         "image":None
     }
 
@@ -1375,12 +1399,12 @@ async def set_sam_color(
 """
 选择分类
 """
-@app.get("/ai-station-api/sam_classs/change")
+@app.get("/ai-station-api/sam_class/change")
 async def on_class_selected(class_index : int = None,path: str = None):
     # 加载配置内容
     loaded_data,api_config = sam_deal.load_model(path)
     result, api_config = sam_deal.set_current_class(loaded_data, api_config, class_index, color=None)
-    # loaded_data['class_index'] = class_index
+    loaded_data['class_index']=class_index
     sam_deal.save_model(loaded_data,api_config,path)
     if result:
         img = sam_deal.refresh_image(loaded_data,api_config,path)
@@ -1388,7 +1412,7 @@ async def on_class_selected(class_index : int = None,path: str = None):
             encoded_string = sam_deal.load_tmp_image(img['reason'])
             return {
                 "success":True,
-                "msg":"",
+                "msg":"更改成功",
                 "image":JSONResponse(content={"image_data": encoded_string})
             }
         else:
@@ -1411,6 +1435,12 @@ async def on_class_selected(class_index : int = None,path: str = None):
 @app.get("/ai-station-api/sam_class/delete")
 async def sam_remove_class(path:str=None,select_index:int=None):
     loaded_data,api_config = sam_deal.load_model(path)
+    if select_index == -1:
+        return {
+        "success":False,
+        "msg":"没有选定分类，请选定当前分类后再删除",
+        "image":None
+    }
     class_name = loaded_data['class_names'][select_index]
     loaded_data,api_config = sam_deal.remove_class(loaded_data,api_config,class_name)
     sam_deal.save_model(loaded_data,api_config,path)
@@ -1419,7 +1449,10 @@ async def sam_remove_class(path:str=None,select_index:int=None):
         encoded_string = sam_deal.load_tmp_image(img['reason'])
         return {
             "success":True,
-            "msg":"",
+            "msg":"删除成功",
+            "class_name_list": loaded_data['class_names'],
+            "current_index": loaded_data['class_index'],
+            "class_dict":loaded_data['class_colors'],
             "image":JSONResponse(content={"image_data": encoded_string})
         }
     else:
@@ -1634,6 +1667,7 @@ async def sam_predict_mask(path:str=None):
 async def sam_reset_annotation(path:str=None):
     loaded_data,api_config = sam_deal.load_model(path)
     loaded_data,api_config = sam_deal.reset_annotation_all(loaded_data,api_config)
+    sam_deal.save_model(loaded_data,api_config,path)
     img = sam_deal.refresh_image(loaded_data,api_config,path)
     if img['status'] == True:
         encoded_string = sam_deal.load_tmp_image(img['reason'])
@@ -1682,8 +1716,6 @@ async def sam_add_to_class(path:str=None,class_index:int=None):
                 "image":None
                 }
 
-
-
 """
 保存结果
 """

test.py

@@ -0,0 +1,52 @@
+# import sys
+
+# from fastapi import FastAPI
+# from fastapi.staticfiles import StaticFiles
+# version = f"{sys.version_info.major}.{sys.version_info.minor}"
+
+# app = FastAPI()
+
+# # 将 /root/app 目录挂载为静态文件
+# app.mount("/files", StaticFiles(directory="/root/app"), name="files")
+
+# @app.get("/")
+# async def read_root():
+#     message = f"Hello world! From FastAPI running on Uvicorn with Gunicorn. Using Python {version}"
+#     return {"message": message}
+
+
+# from PIL import Image
+# import pandas as pd
+
+# # 读取图片
+# image_path = '/home/xiazj/ai-station-code/tmp/dimaoshibie/d94afe94-2fae-4ce5-9ee4-94ac1d699337/merge_binary_dimao2.jpg'  # 替换成你的图片路径
+# image = Image.open(image_path)
+
+# # 获取图片的RGB数据
+# pixels = list(image.getdata())
+
+# # 创建一个DataFrame来统计每种颜色的出现次数
+# color_counts = pd.Series(pixels).value_counts()
+
+# # 将颜色统计结果转换为DataFrame
+# color_summary = color_counts.reset_index()
+# color_summary.columns = ['RGB', 'Count']
+
+# # 打印结果
+# print(color_summary)
+
+# # 可选：将结果保存到CSV文件
+# color_summary.to_csv('color_summary.csv', index=False)
+
+import pickle
+
+# 指定文件路径
+file_path = '/home/xiazj/ai-station-code/tmp/sam/c768f709-b123-48fb-b98b-addf0bbb8a04/model_params.pickle'
+
+# 读取 pickle 文件
+with open(file_path, 'rb') as file:
+    data,api = pickle.load(file)
+
+# 打印数据
+print(data)
+print(api)

work_util/model_deal.py

@@ -82,13 +82,42 @@ def dimaoshibie_pic(model,path,count=None, name_classes=None):
         output_floder_path = os.path.join(file_directory,'binary')
         # 创建输出文件夹（如果不存在）
         os.makedirs(output_floder_path, exist_ok=True)
-        
+        total_piex = {
+            "_background_" : 0,           # Background (黑色)
+            "Cropland" : 0,     # Cropland (淡黄色)
+            "Forest": 0,        # Forest (深绿色)
+            "Grass": 0,     # Grass (浅绿色)
+            "Shrub": 0,      # Shrub (浅蓝绿色)
+            "Wetland": 0,     # Wetland (浅蓝色)
+            "Water": 0,       # Water (深蓝色)
+            "Tundra": 0,       # Tundra (土黄色)
+            "Impervious surface": 0,       # Impervious surface (红色)
+            "Bareland": 0,     # Bareland (灰色)
+            "Ice/snow": 0      # Ice/snow (浅天蓝色)
+        }
         for ii, line in enumerate(lines):
             _imagepath = os.path.join(folder_path, line)
             r_image, count_dict, classes_nums = dimaoshibie_cal(model,_imagepath,count, name_classes)
             final_path = os.path.join(output_floder_path,line)  
             # 保存融合后的图像
             r_image.save(final_path)  # 替换为你想保存的路径
+            for key in count_dict:
+                if key in total_piex:
+                    total_piex[key] += count_dict[key]
+        
+        # 将像素分类，存储
+
+        # 创建一个新的字典，用于存储中文键
+        translated_total_piex = {}
+        for key, value in total_piex.items():
+            if key in param.dmsb_type:
+                translated_total_piex[param.dmsb_type[key]] = value
+        
+        output_file_path = os.path.join(file_directory,'result.txt')
+        # 写入字典到文本文件
+        with open(output_file_path, 'w') as file:
+            for key, value in translated_total_piex.items():
+                file.write(f"{key}: {value}\n")
 
         # 图片合并
         status = prepare_data.merge_pic_binary(path,['binary'])
@@ -115,6 +144,9 @@ def dimaoshibie_pic(model,path,count=None, name_classes=None):
             "msg": "临时文件删除失败",
             "data": None
         } 
+
+
+
         # 返回图片地址
         return {"status":True, "reason":target_path}
 
@@ -131,10 +163,10 @@ def dimaoshibie_area(path,scale,colors=None):
     file_path = os.path.join(file_directory,"merge_binary_" + file_name)
     if not os.path.exists(file_path):
         return {"status": False, "reason": "没有找到对应文件，请先进行图像分割", "result":""}
-    
     image = Image.open(path)
     try:
         # 如果文件存在，读取图片
+        # logger.info(file_path)
         image = Image.open(file_path)
         image = image.convert('RGB')
         color_count = defaultdict(int)
@@ -144,8 +176,11 @@ def dimaoshibie_area(path,scale,colors=None):
                 pixel_color = image.getpixel((x, y))
                 if pixel_color in colors:
                     color_count[pixel_color] += 1
+                else:
+                    color_count[pixel_color] += 1
         for k,v in color_count.items():
             color_count[k] = v * scale
+        # logger.info(color_count)
         result = {color: color_count[color] for color in colors}
         return {"status": True, "reason": result}
     except Exception as e:
@@ -367,6 +402,74 @@ def roofpv_pic(net_roof,net_pv,path,palette):
 
 
 # 甲烷预测
+# def start_predict_endpoint(ch4_model_flow,ch4_model_gas,data_path, start_index, end_index, type, is_show):
+#     try:
+#         data = pd.read_csv(data_path)
+#         data['date_time'] = pd.to_datetime(data['date_time'])
+#         max_date = data['date_time'].max()
+#         min_date = data['date_time'].min()
+#         start_index =  pd.to_datetime(start_index)
+#         end_index = pd.to_datetime(end_index)
+#         if max_date < end_index :
+#             return {"reason": "结束日期填写错误，超过上传数据最大日期","status": False}
+#         if min_date > start_index :
+#             return {"reason": "开始日期填写错误，小于上传数据最小日期","status": False}
+#         end_index_dt = pd.to_datetime(end_index)
+#         end_index_plus_one_hour = end_index_dt + pd.Timedelta(hours=1)
+#         filtered_data = data[(data['date_time'] >= start_index) & (data['date_time'] <= end_index)]
+#         if is_show:
+#             if  len(filtered_data)  < 96:
+#                 return {"reason": "日期填写错误，截取步长应该超过96个步长","status": False}
+#             if  max_date < end_index_plus_one_hour:
+#                 return {"reason": "选择显示真实值，需要保留最终后四节点作为展示信息，请调整结束日期","status": False}
+#         else:
+#             if len(filtered_data) < 96:
+#                 return {"reason": "上传文件中有效信息长度应大于96","status": False}
+        
+#         train_data = prepare_data.get_pred_data(data,start_index,end_index)
+#         del train_data['date_time']
+#         train_data = np.array(train_data.values)
+#         train_data = xgb.DMatrix(train_data)
+#         target = None
+#         if type == 1: # 流量
+#             target = "Nm3d-1-ch4"
+#             result = ch4_model_flow.predict(train_data)
+#         else:         # 气相
+#             target = "S_gas_ch4"
+#             result = ch4_model_gas.predict(train_data)
+#         if is_show:
+#             history = data[(data['date_time'] >= start_index) & (data['date_time'] <= end_index_plus_one_hour)]
+#             cols = ['date_time']
+#             cols.append(target)
+#             history = history[cols]
+#             history.rename(columns={target: 'true_data'}, inplace=True)
+#             history['pred_data'] = 0
+#             total_rows = len(history)
+#             history.reset_index(drop=True, inplace=True)
+#             history.loc[total_rows - 4:total_rows - 1, 'pred_data'] = result[0]
+#         else: 
+#             history = data[(data['date_time'] >= start_index) & (data['date_time'] <= end_index)]
+#             history.reset_index(drop=True, inplace=True)
+#             cols = ['date_time']
+#             cols.append(target)
+#             history = history[cols]
+#             history.rename(columns={target: 'true_data'}, inplace=True)
+#             history['pred_data'] = 0
+#             last_date = history['date_time'].iloc[-1]
+#             # 创建新的日期和对应的值
+#             new_dates = [last_date + pd.Timedelta(minutes=15 * (i + 1)) for i in range(4)]
+#             new_data = pd.DataFrame({
+#                 'date_time': new_dates,
+#                 'true_data':[0,0,0,0],
+#                 'pred_data': result[0]
+#             })
+#             history = pd.concat([history, new_data], ignore_index=True)
+            
+#         return {"status": True,"reason":history.to_dict(orient='records')}
+#     except Exception as e:
+#         return{"reason": str(e),"status":False}
+    
+
 def start_predict_endpoint(ch4_model_flow,ch4_model_gas,data_path, start_index, end_index, type, is_show):
     try:
         data = pd.read_csv(data_path)
@@ -403,40 +506,43 @@ def start_predict_endpoint(ch4_model_flow,ch4_model_gas,data_path, start_index,
             target = "S_gas_ch4"
             result = ch4_model_gas.predict(train_data)
         if is_show:
+            """两字典，历史+真实，预测"""
             history = data[(data['date_time'] >= start_index) & (data['date_time'] <= end_index_plus_one_hour)]
             cols = ['date_time']
             cols.append(target)
             history = history[cols]
-            history.rename(columns={target: 'true_data'}, inplace=True)
+            # history.rename(columns={target: 'true_data'}, inplace=True)
-            history['pred_data'] = 0
+            last_date = history['date_time'].iloc[-5]
-            total_rows = len(history)
+            new_dates = [last_date + pd.Timedelta(minutes=15 * (i + 1)) for i in range(4)]
-            history.reset_index(drop=True, inplace=True)
+            new_data = pd.DataFrame({
-            history.loc[total_rows - 4:total_rows - 1, 'pred_data'] = result[0]
+                'date_time': new_dates,
+                target: result[0]
+            })
+
         else: 
             history = data[(data['date_time'] >= start_index) & (data['date_time'] <= end_index)]
             history.reset_index(drop=True, inplace=True)
             cols = ['date_time']
             cols.append(target)
             history = history[cols]
-            history.rename(columns={target: 'true_data'}, inplace=True)
+            
-            history['pred_data'] = 0
             last_date = history['date_time'].iloc[-1]
             # 创建新的日期和对应的值
             new_dates = [last_date + pd.Timedelta(minutes=15 * (i + 1)) for i in range(4)]
             new_data = pd.DataFrame({
                 'date_time': new_dates,
-                'true_data':[0,0,0,0],
+                target: result[0]
-                'pred_data': result[0]
             })
-            history = pd.concat([history, new_data], ignore_index=True)
             
-        return {"status": True,"reason":history.to_dict(orient='records')}
+            
+        return {"status": True,"reason":[history.to_dict(orient='records'),new_data.to_dict(orient='records')]}
     except Exception as e:
         return{"reason": str(e),"status":False}
 
 
 
 # 光伏出力预测 - 预测长度96
+
 def start_pvelectric_predict_endpoint(pvfd_model,data_path, start_index, end_index, is_show):
     try:
         data = pd.read_csv(data_path)
@@ -471,32 +577,91 @@ def start_pvelectric_predict_endpoint(pvfd_model,data_path, start_index, end_ind
             cols = ['date_time']
             cols.append(target)
             history = history[cols]
-            history.rename(columns={target: 'true_data'}, inplace=True)
+            # history.rename(columns={target: 'true_data'}, inplace=True)
-            history['pred_data'] = 0
+            # history['pred_data'] = 0
-            total_rows = len(history)
+            last_date = history['date_time'].iloc[-97]
-            history.reset_index(drop=True, inplace=True)
+            new_dates = [last_date + pd.Timedelta(minutes=15 * (i + 1)) for i in range(96)]
-            history.loc[total_rows - 96:total_rows - 1, 'pred_data'] = predictions_value
+            new_data = pd.DataFrame({
+                'date_time': new_dates,
+                target: predictions_value
+            })
         else: 
             history = data[(data['date_time'] >= start_index) & (data['date_time'] <= end_index)]
             history.reset_index(drop=True, inplace=True)
             cols = ['date_time']
             cols.append(target)
             history = history[cols]
-            history.rename(columns={target: 'true_data'}, inplace=True)
-            history['pred_data'] = 0
             last_date = history['date_time'].iloc[-1]
             # 创建新的日期和对应的值
             new_dates = [last_date + pd.Timedelta(minutes=15 * (i + 1)) for i in range(96)]
             new_data = pd.DataFrame({
                 'date_time': new_dates,
-                'true_data':[0]*96,
+                target: predictions_value
-                'pred_data': predictions_value
             })
-            history = pd.concat([history, new_data], ignore_index=True)
             
-        return {"status": True,"reason":history.to_dict(orient='records')}
+        return {"status": True,"reason":[history.to_dict(orient='records'),new_data.to_dict(orient='records')]}
     except Exception as e:
         return{"reason": str(e),"status":False}
+# def start_pvelectric_predict_endpoint(pvfd_model,data_path, start_index, end_index, is_show):
+#     try:
+#         data = pd.read_csv(data_path)
+#         data['date_time'] = pd.to_datetime(data['date_time'])
+#         max_date = data['date_time'].max()
+#         min_date = data['date_time'].min()
+#         start_index =  pd.to_datetime(start_index)
+#         end_index = pd.to_datetime(end_index)
+#         if max_date < end_index :
+#             return {"reason": "结束日期填写错误，超过上传数据最大日期","status": False}
+#         if min_date > start_index :
+#             return {"reason": "开始日期填写错误，小于上传数据最小日期","status": False}
+#         end_index_dt = pd.to_datetime(end_index)
+#         end_index_plus_one_day = end_index_dt + pd.Timedelta(hours=24)
+#         filtered_data = data[(data['date_time'] >= start_index) & (data['date_time'] <= end_index)]
+#         if is_show:
+#             if  len(filtered_data)  < 192:
+#                 return {"reason": "日期填写错误，截取步长应该超过192个步长","status": False}
+#             if  max_date < end_index_plus_one_day:
+#                 return {"reason": "选择显示真实值，需要保留最终一天数据96点作为展示信息，请调整结束日期","status": False}
+#         else:
+#             if len(filtered_data) < 192:
+#                 return {"reason": "上传文件中有效信息长度应大于192","status": False}
+        
+#         predictions = pvfd_model.run_inference(filtered_data)
+#         if predictions['status'] == True:
+#             predictions_value = np.array(predictions['reason']).flatten()
+#             predictions_value = [max(0, x) for x in predictions_value]
+#         target = "power"
+#         if is_show:
+#             history = data[(data['date_time'] >= start_index) & (data['date_time'] <= end_index_plus_one_day)]
+#             cols = ['date_time']
+#             cols.append(target)
+#             history = history[cols]
+#             history.rename(columns={target: 'true_data'}, inplace=True)
+#             history['pred_data'] = 0
+#             total_rows = len(history)
+#             history.reset_index(drop=True, inplace=True)
+#             history.loc[total_rows - 96:total_rows - 1, 'pred_data'] = predictions_value
+#         else: 
+#             history = data[(data['date_time'] >= start_index) & (data['date_time'] <= end_index)]
+#             history.reset_index(drop=True, inplace=True)
+#             cols = ['date_time']
+#             cols.append(target)
+#             history = history[cols]
+#             history.rename(columns={target: 'true_data'}, inplace=True)
+#             history['pred_data'] = 0
+#             last_date = history['date_time'].iloc[-1]
+#             # 创建新的日期和对应的值
+#             new_dates = [last_date + pd.Timedelta(minutes=15 * (i + 1)) for i in range(96)]
+#             new_data = pd.DataFrame({
+#                 'date_time': new_dates,
+#                 'true_data':[0]*96,
+#                 'pred_data': predictions_value
+#             })
+#             history = pd.concat([history, new_data], ignore_index=True)
+            
+#         return {"status": True,"reason":history.to_dict(orient='records')}
+#     except Exception as e:
+#         return{"reason": str(e),"status":False}
 
 
 
@@ -535,32 +700,92 @@ def start_wind_electric_predict_endpoint(windfd_model,data_path, start_index, en
         cols = ['date']
         cols.append(target)
         history = history[cols]
-        history.rename(columns={target: 'true_data'}, inplace=True)
+        # history.rename(columns={target: 'true_data'}, inplace=True)
-        history['pred_data'] = 0
+        last_date = history['date'].iloc[-13]
-        total_rows = len(history)
+        new_dates = [last_date + pd.Timedelta(minutes=15 * (i + 1)) for i in range(12)]
-        history.reset_index(drop=True, inplace=True)
+        new_data = pd.DataFrame({
-        history.loc[total_rows - 12:total_rows - 1, 'pred_data'] = predictions_value
+            'date': new_dates,
+            target: predictions_value
+        })
     else: 
         history = data[(data['date'] >= start_index) & (data['date'] <= end_index)]
         history.reset_index(drop=True, inplace=True)
         cols = ['date']
         cols.append(target)
         history = history[cols]
-        history.rename(columns={target: 'true_data'}, inplace=True)
-        history['pred_data'] = 0
         last_date = history['date'].iloc[-1]
         # 创建新的日期和对应的值
         new_dates = [last_date + pd.Timedelta(minutes=15 * (i + 1)) for i in range(12)]
         new_data = pd.DataFrame({
             'date': new_dates,
-            'true_data':[0]*12,
+            target: predictions_value
-            'pred_data': predictions_value
         })
-        history = pd.concat([history, new_data], ignore_index=True)
+    history.rename(columns={target: 'power'}, inplace=True)
-        
+    new_data.rename(columns={target: 'power'}, inplace=True)
-    return {"status": True,"reason":history.to_dict(orient='records')}
+    return {"status": True,"reason":[history.to_dict(orient='records'),new_data.to_dict(orient='records')]}
     # except Exception as e:
     #     return{"reason": str(e),"status":False}
+# def start_wind_electric_predict_endpoint(windfd_model,data_path, start_index, end_index, is_show):
+#     # try:
+#     data = pd.read_csv(data_path)
+#     data['date'] = pd.to_datetime(data['date'])
+#     max_date = data['date'].max()
+#     min_date = data['date'].min()
+#     start_index =  pd.to_datetime(start_index)
+#     end_index = pd.to_datetime(end_index)
+#     if max_date < end_index :
+#         return {"reason": "结束日期填写错误，超过上传数据最大日期","status": False}
+#     if min_date > start_index :
+#         return {"reason": "开始日期填写错误，小于上传数据最小日期","status": False}
+#     end_index_dt = pd.to_datetime(end_index)
+#     end_index_plus_one_day = end_index_dt + pd.Timedelta(hours=3)
+#     filtered_data = data[(data['date'] >= start_index) & (data['date'] <= end_index)]
+#     if is_show:
+#         if  len(filtered_data)  < 192:
+#             return {"reason": "日期填写错误，截取步长应该超过192个步长","status": False}
+#         if  max_date < end_index_plus_one_day:
+#             return {"reason": "选择显示真实值，需要保留最终一天数据12个点作为展示信息，请调整结束日期","status": False}
+#     else:
+#         if len(filtered_data) < 192:
+#             return {"reason": "上传文件中有效信息长度应大于192","status": False}
+    
+#     predictions = windfd_model.run_inference(filtered_data)
+#     if predictions['status'] == True:
+#         predictions_value = np.array(predictions['reason']).flatten()
+#         predictions_value = [max(0, x) for x in predictions_value]
+#     target = "Power"
+#     if is_show:
+#         history = data[(data['date'] >= start_index) & (data['date'] <= end_index_plus_one_day)]
+#         cols = ['date']
+#         cols.append(target)
+#         history = history[cols]
+#         history.rename(columns={target: 'true_data'}, inplace=True)
+#         history['pred_data'] = 0
+#         total_rows = len(history)
+#         history.reset_index(drop=True, inplace=True)
+#         history.loc[total_rows - 12:total_rows - 1, 'pred_data'] = predictions_value
+#     else: 
+#         history = data[(data['date'] >= start_index) & (data['date'] <= end_index)]
+#         history.reset_index(drop=True, inplace=True)
+#         cols = ['date']
+#         cols.append(target)
+#         history = history[cols]
+#         history.rename(columns={target: 'true_data'}, inplace=True)
+#         history['pred_data'] = 0
+#         last_date = history['date'].iloc[-1]
+#         # 创建新的日期和对应的值
+#         new_dates = [last_date + pd.Timedelta(minutes=15 * (i + 1)) for i in range(12)]
+#         new_data = pd.DataFrame({
+#             'date': new_dates,
+#             'true_data':[0]*12,
+#             'pred_data': predictions_value
+#         })
+#         history = pd.concat([history, new_data], ignore_index=True)
+        
+#     return {"status": True,"reason":history.to_dict(orient='records')}
+#     # except Exception as e:
+#     #     return{"reason": str(e),"status":False}
+
 
 
 def pred_single_tar(test_content):

work_util/params.py

@@ -46,21 +46,34 @@ class ModelParams():
         (211, 242, 255)      # Ice/snow (浅天蓝色)
     ]
     
-
     dmsb_type = {
-        (0, 0, 0) : "背景",           # Background (黑色)
+        "_background_" : "背景",           # Background (黑色)
-        (252, 250, 205) : "农田",     # Cropland (淡黄色)
+        "Cropland" : "农田",     # Cropland (淡黄色)
-        (0, 123, 79): "森林",        # Forest (深绿色)
+        "Forest": "森林",        # Forest (深绿色)
-        (157, 221, 106): "草地",     # Grass (浅绿色)
+        "Grass": "草地",     # Grass (浅绿色)
-        (77, 208, 159): "灌木",      # Shrub (浅蓝绿色)
+        "Shrub": "灌木",      # Shrub (浅蓝绿色)
-        (111, 208, 242): "湿地",     # Wetland (浅蓝色)
+        "Wetland": "湿地",     # Wetland (浅蓝色)
-        (10, 78, 151): "水体",       # Water (深蓝色)
+        "Water": "水体",       # Water (深蓝色)
-        (92, 106, 55): "苔原",       # Tundra (土黄色)
+        "Tundra": "苔原",       # Tundra (土黄色)
-        (155, 36, 22): "建筑",       # Impervious surface (红色)
+        "Impervious surface": "建筑",       # Impervious surface (红色)
-        (205, 205, 205): "裸地",     # Bareland (灰色)
+        "Bareland": "裸地",     # Bareland (灰色)
-        (211, 242, 255): "冰雪"      # Ice/snow (浅天蓝色)
+        "Ice/snow": "冰雪"      # Ice/snow (浅天蓝色)
     }
 
+    # dmsb_type = {
+    #     (0, 0, 0) : "背景",           # Background (黑色)
+    #     (252, 250, 205) : "农田",     # Cropland (淡黄色)
+    #     (0, 123, 79): "森林",        # Forest (深绿色)
+    #     (157, 221, 106): "草地",     # Grass (浅绿色)
+    #     (77, 208, 159): "灌木",      # Shrub (浅蓝绿色)
+    #     (111, 208, 242): "湿地",     # Wetland (浅蓝色)
+    #     (10, 78, 151): "水体",       # Water (深蓝色)
+    #     (92, 106, 55): "苔原",       # Tundra (土黄色)
+    #     (155, 36, 22): "建筑",       # Impervious surface (红色)
+    #     (205, 205, 205): "裸地",     # Bareland (灰色)
+    #     (211, 242, 255): "冰雪"      # Ice/snow (浅天蓝色)
+    # }
+
 
     # 前缀 wdpv 屋顶光伏
     wdpv_palette = [0, 0, 0, 255, 0, 0, 0, 255, 0]

work_util/post_model.py

@@ -1,6 +1,6 @@
 from pydantic import BaseModel
 from typing import Optional
-
+from typing import List
 
 class Zongbiaomianji(BaseModel):
     A: float
@@ -63,3 +63,15 @@ class FormData(BaseModel):
     Rt_min: Optional[float] = None
     Rt_max: Optional[float] = None
     Rt_step: Optional[float] = None
+
+
+class samItem(BaseModel):
+    class_name: str
+    color: Optional[List[int]] = None
+    path: str
+
+
+class samItem2(BaseModel):
+    current_index:int
+    rgb_color: List[int] 
+    path: str

work_util/sam_deal.py

@@ -57,9 +57,11 @@ def reset_annotation(loaded_data):
 def remove_class(data,api,class_name):
     if class_name in api['class_annotations']:
         del api['class_annotations'][class_name]
-    
+        del api['class_colors'][class_name]
     if class_name == data['class_names'][data['class_index']]:
-        data['class_index'] = 0
+        # data['class_index'] = 0
+        data['class_index'] = -1
+        api['current_class'] = None
     data['class_names'].remove(class_name)
     if class_name in data['class_colors']:
         del data['class_colors'][class_name]
@@ -236,7 +238,6 @@ def add_to_class(api,class_name):
 
 
 def get_image_display(data,api):
-
     if data['image'] is None:
         logger.info("get_display_image: No image loaded")
         return {"status":False, "reason":"获取图像：没有图像加载"}
@@ -245,10 +246,11 @@ def get_image_display(data,api):
         for point, label in zip(data['input_point'], data['input_label']):
             color = (0, 255, 0) if label == 1 else (0, 0, 255)
             cv2.circle(display_image, tuple(point), 5, color, -1)
-        if data['selected_mask'] is not None:
+        if data['class_index'] != -1:
-            class_name = data['class_names'][data['class_index']]
+            if data['selected_mask'] is not None:
-            color = data['class_colors'].get(class_name, (0, 0, 128))
+                class_name = data['class_names'][data['class_index']]
-            display_image = sam_apply_mask(display_image, data['selected_mask'], color)
+                color = data['class_colors'].get(class_name, (0, 0, 128))
+                display_image = sam_apply_mask(display_image, data['selected_mask'], color)
         logger.info(f"get_display_image: Returning image with shape {display_image.shape}")
         img = display_image
         if not isinstance(img, np.ndarray) or img.size == 0: