davisAPI/davisAPI.py

@@ -1,11 +1,10 @@
-from PyWeather.weather.stations.davis import VantagePro
 import logging
+import time
+
 import mariadb
 import serial.tools.list_ports
-import gc
-import time
-from pprint import pprint
 
+from PyWeather.weather.stations.davis import VantagePro
 
 logging.basicConfig(filename="Stations.log",
                     format='%(asctime)s %(message)s',
@@ -13,37 +12,10 @@ logging.basicConfig(filename="Stations.log",
 logger = logging.getLogger('davis_api')
 logger.setLevel(logging.DEBUG)
 
-
-def write_data(device, station, send=True):
-    try:
-        #device.parse()
-        data = device.fields
-        print(data)
-        if len(data) < 1:
-            return
-        else:
-            print(data)
-        fields = ['BarTrend', 'CRC', 'DateStamp', 'DewPoint', 'HeatIndex', 'ETDay', 'HeatIndex',
-                  'HumIn', 'HumOut', 'Pressure', 'RainDay', 'RainMonth', 'RainRate', 'RainStorm',
-                  'RainYear', 'SunRise', 'SunSet', 'TempIn', 'TempOut', 'WindDir', 'WindSpeed',
-                  'WindSpeed10Min']
-
-        if send:
-            placeholders = ', '.join(['%s'] * len(fields))
-            field_names = ', '.join(fields)
-            sql = f"INSERT INTO weather_data ({field_names}) VALUES ({placeholders})"
-            values = [data[field] for field in fields]
-            cursor.execute(sql, values)
-            conn.commit()
-        else:
-            pprint(data)
-
-        del data
-        del fields
-        gc.collect()
-    except Exception as e:
-        logger.error(str(e))
-        raise e
+console_handler = logging.StreamHandler()
+console_handler.setLevel(logging.DEBUG)
+console_handler.setFormatter(logging.Formatter('%(asctime)s %(message)s'))
+logger.addHandler(console_handler)
 
 try:
     conn = mariadb.connect(
@@ -57,7 +29,7 @@ try:
 except mariadb.Error as e:
     logger.error('DB_ERR: ' + str(e))
     raise e
-
+while True:
     try:
         ports = serial.tools.list_ports.comports()
         available_ports = {}
@@ -67,13 +39,15 @@ try:
                 available_ports[port.name] = port.vid
 
         devices = [VantagePro(port) for port in available_ports.keys()]
-    print(available_ports)
         while True:
-        for i in range(len(devices)):
-            print(devices[i].fields)
-            #write_data(devices[i], 'st' + str(available_ports[list(available_ports.keys())[i]]), True)
-        time.sleep(1)
+            for i in range(1):
+                if len(devices) != 0:
+                    logger.info(devices)
+                else:
+                    raise Exception('Can`t connect to device')
+                time.sleep(60)
     except Exception as e:
-    logger.error('Device_error: ' + str(e))
-    raise e
+        logger.error('Device_error' + str(e))
+        time.sleep(60)
 
+# todo переписать под influx, для линухи приколы сделать

davisAPI/prediction.py

@@ -0,0 +1,200 @@
+from datetime import datetime
+from pathlib import Path
+
+import metpy.calc
+import numpy as np
+import requests
+import torch
+import xarray as xr
+from aurora import AuroraSmall, Batch, Metadata
+from metpy.units import units
+
+
+def get_download_paths(date):
+    """Создает список путей для загрузки данных."""
+    download_path = Path("~/downloads/hres_0.1").expanduser()
+    downloads = {}
+    var_nums = {
+        "2t": "167", "10u": "165", "10v": "166", "msl": "151", "t": "130",
+        "u": "131", "v": "132", "q": "133", "z": "129", "slt": "043", "lsm": "172",
+    }
+    for v in ["2t", "10u", "10v", "msl", "z", "slt", "lsm"]:
+        downloads[download_path / date.strftime(f"surf_{v}_%Y-%m-%d.grib")] = (
+            f"https://data.rda.ucar.edu/ds113.1/"
+            f"ec.oper.an.sfc/{date.year}{date.month:02d}/ec.oper.an.sfc.128_{var_nums[v]}_{v}."
+            f"regn1280sc.{date.year}{date.month:02d}{date.day:02d}.grb"
+        )
+    for v in ["z", "t", "u", "v", "q"]:
+        for hour in [0, 6, 12, 18]:
+            prefix = "uv" if v in {"u", "v"} else "sc"
+            downloads[download_path / date.strftime(f"atmos_{v}_%Y-%m-%d_{hour:02d}.grib")] = (
+                f"https://data.rda.ucar.edu/ds113.1/"
+                f"ec.oper.an.pl/{date.year}{date.month:02d}/ec.oper.an.pl.128_{var_nums[v]}_{v}."
+                f"regn1280{prefix}.{date.year}{date.month:02d}{date.day:02d}{hour:02d}.grb"
+            )
+    return downloads, download_path
+
+
+def download_data(downloads):
+    """Скачивает файлы, если они отсутствуют в целевой директории."""
+    for target, source in downloads.items():
+        if not target.exists():
+            print(f"Downloading {source}")
+            target.parent.mkdir(parents=True, exist_ok=True)
+            response = requests.get(source)
+            response.raise_for_status()
+            with open(target, "wb") as f:
+                f.write(response.content)
+            print("Downloads finished!")
+
+
+def load_surf(v, v_in_file, download_path, date):
+    """Загружает переменные поверхностного уровня или статические переменные."""
+    ds = xr.open_dataset(download_path / date.strftime(f"surf_{v}_%Y-%m-%d.grib"), engine="cfgrib")
+    data = ds[v_in_file].values[:2]
+    data = data[None]
+    return torch.from_numpy(data)
+
+
+def load_atmos(v, download_path, date, levels):
+    """Загружает атмосферные переменные для заданных уровней давления."""
+    ds_00 = xr.open_dataset(
+        download_path / date.strftime(f"atmos_{v}_%Y-%m-%d_00.grib"), engine="cfgrib"
+    )
+    ds_06 = xr.open_dataset(
+        download_path / date.strftime(f"atmos_{v}_%Y-%m-%d_06.grib"), engine="cfgrib"
+    )
+    ds_00 = ds_00[v].sel(isobaricInhPa=list(levels))
+    ds_06 = ds_06[v].sel(isobaricInhPa=list(levels))
+    data = np.stack((ds_00.values, ds_06.values), axis=0)
+    data = data[None]
+    return torch.from_numpy(data)
+
+
+def create_batch(date, levels, downloads, download_path):
+    """Создает объект Batch с данными для модели."""
+    ds = xr.open_dataset(next(iter(downloads.keys())), engine="cfgrib")
+    batch = Batch(
+        surf_vars={
+            "2t": load_surf("2t", "t2m", download_path, date),
+            "10u": load_surf("10u", "u10", download_path, date),
+            "10v": load_surf("10v", "v10", download_path, date),
+            "msl": load_surf("msl", "msl", download_path, date),
+        },
+        static_vars={
+            "z": load_surf("z", "z", download_path, date)[0, 0],
+            "slt": load_surf("slt", "slt", download_path, date)[0, 0],
+            "lsm": load_surf("lsm", "lsm", download_path, date)[0, 0],
+        },
+        atmos_vars={
+            "t": load_atmos("t", download_path, date, levels),
+            "u": load_atmos("u", download_path, date, levels),
+            "v": load_atmos("v", download_path, date, levels),
+            "q": load_atmos("q", download_path, date, levels),
+            "z": load_atmos("z", download_path, date, levels),
+        },
+        metadata=Metadata(
+            lat=torch.from_numpy(ds.latitude.values),
+            lon=torch.from_numpy(ds.longitude.values),
+            time=(date.replace(hour=6),),
+            atmos_levels=levels,
+        ),
+    )
+    return batch.regrid(res=0.1)
+
+
+def create_batch_random(levels: tuple[int], date: tuple):
+    """Создает объект Batch с рандомными данными для модели."""
+    return Batch(
+        surf_vars={k: torch.randn(1, 2, 17, 32) for k in ("2t", "10u", "10v", "msl")},
+        static_vars={k: torch.randn(17, 32) for k in ("lsm", "z", "slt")},
+        atmos_vars={k: torch.randn(1, 2, 4, 17, 32) for k in ("z", "u", "v", "t", "q")},
+        metadata=Metadata(
+            lat=torch.linspace(90, -90, 17),
+            lon=torch.linspace(0, 360, 32 + 1)[:-1],
+            time=date,
+            atmos_levels=levels,
+        ),
+    )
+
+
+def run_model(batch):
+    """Инициализирует модель AuroraSmall и выполняет предсказание."""
+    model = AuroraSmall()
+    model.load_checkpoint("microsoft/aurora", "aurora-0.25-small-pretrained.ckpt")
+    model.eval()
+    model = model.to("cpu")
+    with torch.inference_mode():
+        prediction = model.forward(batch)
+    return prediction
+
+
+def get_wind_speed_and_direction(prediction, batch: Batch, lat: float, lon: float):
+    target_lat = lat
+    target_lon = lon
+
+    lat_idx = torch.abs(batch.metadata.lat - target_lat).argmin()
+    lon_idx = torch.abs(batch.metadata.lon - target_lon).argmin()
+
+    u_values = prediction.atmos_vars["u"][:, :, :, lat_idx, lon_idx]
+    v_values = prediction.atmos_vars["v"][:, :, :, lat_idx, lon_idx]
+    wind_speeds=[]
+    wind_directions=[]
+    for i in range(u_values.numel()):
+        u_scalar = u_values.view(-1)[i].item()  # Разворачиваем тензор в одномерный и берем элемент
+        v_scalar = v_values.view(-1)[i].item()
+
+        print("u value:", u_scalar)
+        print("v value:", v_scalar)
+
+        u_with_units = u_scalar * units("m/s")
+        v_with_units = v_scalar * units("m/s")
+
+        # Рассчитайте направление и скорость ветра
+        wind_dir = metpy.calc.wind_direction(u_with_units, v_with_units)
+        wind_speed = metpy.calc.wind_speed(u_with_units, v_with_units)
+
+        wind_speeds.append(wind_speed.magnitude.item())
+        wind_directions.append(wind_dir.magnitude.item())
+
+    return wind_speeds,wind_directions
+
+
+def wind_direction_to_text(wind_dir_deg):
+    directions = [
+        "север", "северо-восток", "восток", "юго-восток",
+        "юг", "юго-запад", "запад", "северо-запад"
+    ]
+    idx = int((wind_dir_deg + 22.5) // 45) % 8
+    return directions[idx]
+
+
+def get_weather_predict(
+        dates: tuple[datetime],
+        latitude: float,
+        longitude: float,
+):
+    levels = (100,)
+    batch_actual = create_batch_random(levels, dates)
+    prediction_actual = run_model(batch_actual)
+    return get_wind_speed_and_direction(prediction_actual, batch_actual, latitude, longitude)
+
+
+def main():
+    levels = (100,)
+
+    date1 = datetime(2024, 11, 27, 12)
+    date2 = datetime(2024, 11, 28, 12)
+    date_tuple = (date1, date2,)
+    # downloads, download_path = get_download_paths(date)
+    # download_data(downloads)  # Скачиваем данные, если их нет
+    # batch_actual = create_batch(date, levels, downloads, download_path)
+    batch_actual = create_batch_random(levels, date_tuple)
+    prediction_actual = run_model(batch_actual)
+    wind_speed_and_direction = get_wind_speed_and_direction(prediction_actual, batch_actual, 50, 20)
+    return wind_speed_and_direction
+
+
+if __name__ == "__main__":
+    main()
+    print("Prediction completed!")

server/src/routers/floris_router.py

@@ -1,3 +1,4 @@
+import datetime
 import math
 import sys
 from http import HTTPStatus
@@ -18,6 +19,8 @@ sys.path.append(str(Path(__file__).parent.parent.parent))
 from floris_module.src import FlorisULSTU
 from floris_module.src.OpenMeteoClient import OpenMeteoClient
 
+from utils import prediction as weather_prediction
+
 FLORIS_IMAGES_PATH = Path(__file__).parent.parent.parent / "public" / "floris"
 
 router = APIRouter(
@@ -25,6 +28,11 @@ router = APIRouter(
     tags=["Floris Api"],
 )
 
+def daterange(start_date: datetime.date, end_date: datetime.date):
+    days = int((end_date - start_date).days)
+    for n in range(days + 1):
+        yield datetime.datetime.combine(start_date + datetime.timedelta(n), datetime.datetime.min.time())
+
 
 @router.get("/get_windmill_data")
 async def get_windmill_data(
@@ -40,7 +48,14 @@ async def get_windmill_data(
 
     client = OpenMeteoClient()
 
+    if data.date_start >= datetime.date.today():
+        dates = tuple(daterange(data.date_start, data.date_end))
+        climate_info = weather_prediction.get_weather_predict(
+            dates, 54.35119762746125, 48.389356992149345
+        )
+    else:
         climate_info = client.get_weather_info(data.date_start, data.date_end)
+
     wind_speeds = climate_info[0]
     wind_directions = climate_info[1]
 
@@ -114,13 +129,20 @@ async def get_windmill_data_by_wind_park(
 
     turbine_yaw_angles = [turbine.angle for turbine in turbines]
 
-    weather_info_list = [
-        client.get_weather_info(start_date=data.date_start, end_date=data.date_end, latitude=lat, longitude=lon)
-        for lat, lon in turbine_coordinates
-    ]
+    
 
     park_centerX, park_centerY = get_absolute_coordinates(park.CenterLatitude, park.CenterLongitude)
 
+    if data.date_start >= datetime.date.today():
+        dates = tuple(daterange(data.date_start, data.date_end))
+        weather_info_list = weather_prediction.get_weather_predict(
+            dates, park.CenterLatitude, park.CenterLongitude
+        )
+    else:
+        weather_info_list = client.get_weather_info(start_date=data.date_start, end_date=data.date_end,
+                                                    latitude=park.CenterLatitude, longitude=park.CenterLongitude)
+
+
     turbineX = [
         park_centerX + turbine.x_offset
         for turbine in turbines
@@ -131,8 +153,8 @@ async def get_windmill_data_by_wind_park(
         for turbine in turbines
     ]
 
-    wind_speeds = np.array([item[0][0] for item in weather_info_list])
-    wind_directions = np.array([item[1][0] for item in weather_info_list])
+    wind_speeds = np.array(weather_info_list[0])
+    wind_directions = np.array(weather_info_list[1])
 
     print(wind_directions)
 

server/src/utils/prediction.py

@@ -0,0 +1,200 @@
+from datetime import datetime
+from pathlib import Path
+
+import metpy.calc
+import numpy as np
+import requests
+import torch
+import xarray as xr
+from aurora import AuroraSmall, Batch, Metadata
+from metpy.units import units
+
+
+def get_download_paths(date):
+    """Создает список путей для загрузки данных."""
+    download_path = Path("~/downloads/hres_0.1").expanduser()
+    downloads = {}
+    var_nums = {
+        "2t": "167", "10u": "165", "10v": "166", "msl": "151", "t": "130",
+        "u": "131", "v": "132", "q": "133", "z": "129", "slt": "043", "lsm": "172",
+    }
+    for v in ["2t", "10u", "10v", "msl", "z", "slt", "lsm"]:
+        downloads[download_path / date.strftime(f"surf_{v}_%Y-%m-%d.grib")] = (
+            f"https://data.rda.ucar.edu/ds113.1/"
+            f"ec.oper.an.sfc/{date.year}{date.month:02d}/ec.oper.an.sfc.128_{var_nums[v]}_{v}."
+            f"regn1280sc.{date.year}{date.month:02d}{date.day:02d}.grb"
+        )
+    for v in ["z", "t", "u", "v", "q"]:
+        for hour in [0, 6, 12, 18]:
+            prefix = "uv" if v in {"u", "v"} else "sc"
+            downloads[download_path / date.strftime(f"atmos_{v}_%Y-%m-%d_{hour:02d}.grib")] = (
+                f"https://data.rda.ucar.edu/ds113.1/"
+                f"ec.oper.an.pl/{date.year}{date.month:02d}/ec.oper.an.pl.128_{var_nums[v]}_{v}."
+                f"regn1280{prefix}.{date.year}{date.month:02d}{date.day:02d}{hour:02d}.grb"
+            )
+    return downloads, download_path
+
+
+def download_data(downloads):
+    """Скачивает файлы, если они отсутствуют в целевой директории."""
+    for target, source in downloads.items():
+        if not target.exists():
+            print(f"Downloading {source}")
+            target.parent.mkdir(parents=True, exist_ok=True)
+            response = requests.get(source)
+            response.raise_for_status()
+            with open(target, "wb") as f:
+                f.write(response.content)
+            print("Downloads finished!")
+
+
+def load_surf(v, v_in_file, download_path, date):
+    """Загружает переменные поверхностного уровня или статические переменные."""
+    ds = xr.open_dataset(download_path / date.strftime(f"surf_{v}_%Y-%m-%d.grib"), engine="cfgrib")
+    data = ds[v_in_file].values[:2]
+    data = data[None]
+    return torch.from_numpy(data)
+
+
+def load_atmos(v, download_path, date, levels):
+    """Загружает атмосферные переменные для заданных уровней давления."""
+    ds_00 = xr.open_dataset(
+        download_path / date.strftime(f"atmos_{v}_%Y-%m-%d_00.grib"), engine="cfgrib"
+    )
+    ds_06 = xr.open_dataset(
+        download_path / date.strftime(f"atmos_{v}_%Y-%m-%d_06.grib"), engine="cfgrib"
+    )
+    ds_00 = ds_00[v].sel(isobaricInhPa=list(levels))
+    ds_06 = ds_06[v].sel(isobaricInhPa=list(levels))
+    data = np.stack((ds_00.values, ds_06.values), axis=0)
+    data = data[None]
+    return torch.from_numpy(data)
+
+
+def create_batch(date, levels, downloads, download_path):
+    """Создает объект Batch с данными для модели."""
+    ds = xr.open_dataset(next(iter(downloads.keys())), engine="cfgrib")
+    batch = Batch(
+        surf_vars={
+            "2t": load_surf("2t", "t2m", download_path, date),
+            "10u": load_surf("10u", "u10", download_path, date),
+            "10v": load_surf("10v", "v10", download_path, date),
+            "msl": load_surf("msl", "msl", download_path, date),
+        },
+        static_vars={
+            "z": load_surf("z", "z", download_path, date)[0, 0],
+            "slt": load_surf("slt", "slt", download_path, date)[0, 0],
+            "lsm": load_surf("lsm", "lsm", download_path, date)[0, 0],
+        },
+        atmos_vars={
+            "t": load_atmos("t", download_path, date, levels),
+            "u": load_atmos("u", download_path, date, levels),
+            "v": load_atmos("v", download_path, date, levels),
+            "q": load_atmos("q", download_path, date, levels),
+            "z": load_atmos("z", download_path, date, levels),
+        },
+        metadata=Metadata(
+            lat=torch.from_numpy(ds.latitude.values),
+            lon=torch.from_numpy(ds.longitude.values),
+            time=(date.replace(hour=6),),
+            atmos_levels=levels,
+        ),
+    )
+    return batch.regrid(res=0.1)
+
+
+def create_batch_random(levels: tuple[int], date: tuple):
+    """Создает объект Batch с рандомными данными для модели."""
+    return Batch(
+        surf_vars={k: torch.randn(1, 2, 17, 32) for k in ("2t", "10u", "10v", "msl")},
+        static_vars={k: torch.randn(17, 32) for k in ("lsm", "z", "slt")},
+        atmos_vars={k: torch.randn(1, 2, 4, 17, 32) for k in ("z", "u", "v", "t", "q")},
+        metadata=Metadata(
+            lat=torch.linspace(90, -90, 17),
+            lon=torch.linspace(0, 360, 32 + 1)[:-1],
+            time=date,
+            atmos_levels=levels,
+        ),
+    )
+
+
+def run_model(batch):
+    """Инициализирует модель AuroraSmall и выполняет предсказание."""
+    model = AuroraSmall()
+    model.load_checkpoint("microsoft/aurora", "aurora-0.25-small-pretrained.ckpt")
+    model.eval()
+    model = model.to("cpu")
+    with torch.inference_mode():
+        prediction = model.forward(batch)
+    return prediction
+
+
+def get_wind_speed_and_direction(prediction, batch: Batch, lat: float, lon: float):
+    target_lat = lat
+    target_lon = lon
+
+    lat_idx = torch.abs(batch.metadata.lat - target_lat).argmin()
+    lon_idx = torch.abs(batch.metadata.lon - target_lon).argmin()
+
+    u_values = prediction.atmos_vars["u"][:, :, :, lat_idx, lon_idx]
+    v_values = prediction.atmos_vars["v"][:, :, :, lat_idx, lon_idx]
+    wind_speeds=[]
+    wind_directions=[]
+    for i in range(u_values.numel()):
+        u_scalar = u_values.view(-1)[i].item()  # Разворачиваем тензор в одномерный и берем элемент
+        v_scalar = v_values.view(-1)[i].item()
+
+        print("u value:", u_scalar)
+        print("v value:", v_scalar)
+
+        u_with_units = u_scalar * units("m/s")
+        v_with_units = v_scalar * units("m/s")
+
+        # Рассчитайте направление и скорость ветра
+        wind_dir = metpy.calc.wind_direction(u_with_units, v_with_units)
+        wind_speed = metpy.calc.wind_speed(u_with_units, v_with_units)
+
+        wind_speeds.append(wind_speed.magnitude.item())
+        wind_directions.append(wind_dir.magnitude.item())
+
+    return wind_speeds,wind_directions
+
+
+def wind_direction_to_text(wind_dir_deg):
+    directions = [
+        "север", "северо-восток", "восток", "юго-восток",
+        "юг", "юго-запад", "запад", "северо-запад"
+    ]
+    idx = int((wind_dir_deg + 22.5) // 45) % 8
+    return directions[idx]
+
+
+def get_weather_predict(
+        dates: tuple[datetime],
+        latitude: float,
+        longitude: float,
+):
+    levels = (100,)
+    batch_actual = create_batch_random(levels, dates)
+    prediction_actual = run_model(batch_actual)
+    return get_wind_speed_and_direction(prediction_actual, batch_actual, latitude, longitude)
+
+
+def main():
+    levels = (100,)
+
+    date1 = datetime(2024, 11, 27, 12)
+    date2 = datetime(2024, 11, 28, 12)
+    date_tuple = (date1, date2,)
+    # downloads, download_path = get_download_paths(date)
+    # download_data(downloads)  # Скачиваем данные, если их нет
+    # batch_actual = create_batch(date, levels, downloads, download_path)
+    batch_actual = create_batch_random(levels, date_tuple)
+    prediction_actual = run_model(batch_actual)
+    wind_speed_and_direction = get_wind_speed_and_direction(prediction_actual, batch_actual, 50, 20)
+    return wind_speed_and_direction
+
+
+if __name__ == "__main__":
+    main()
+    print("Prediction completed!")