add meteo data + intro reformulee

APIs/meteo.py

@@ -5,9 +5,9 @@ from retry_requests import retry
 
 
 # Setup the Open-Meteo API client with cache and retry on error
-cache_session = requests_cache.CachedSession('.cache', expire_after=3600)
-retry_session = retry(cache_session, retries=5, backoff_factor=0.2)
-openmeteo = openmeteo_requests.Client(session=retry_session)
+cache_session = requests_cache.CachedSession('.cache', expire_after = 3600)
+retry_session = retry(cache_session, retries = 5, backoff_factor = 0.2)
+openmeteo = openmeteo_requests.Client(session = retry_session)
 
 
 def get_info_meteo(latitude, longitude):
@@ -15,63 +15,70 @@ def get_info_meteo(latitude, longitude):
     ARGS: 
         latitude (Float) : latitude coordinate
         longitude (Float): longitude coordinate
-
+    
     """
-
     url = "https://api.open-meteo.com/v1/forecast"
     params = {
-        "latitude": latitude,
-        "longitude": longitude,
-        "hourly": {"relative_humidity_2m", "soil_temperature_0cm", "soil_moisture_0_to_1cm"},
-        "daily": {"temperature_2m_max", "precipitation_sum", "wind_speed_10m_max", "sunshine_duration"},
-        "past_days": 30,
-        "forecast_days": 7
-
-    }
-
+        "latitude": latitude, 
+        "longitude": longitude, #"cloud_cover_low", "cloud_cover_mid","cloud_cover_high"
+        "hourly": {"relative_humidity_2m", "soil_temperature_0cm", "soil_moisture_0_to_1cm", "cloud_cover"},
+        "daily": {"temperature_2m_max","precipitation_sum", "wind_speed_10m_max", "sunshine_duration", "rain_sum"},
+        "past_days" : 5,
+        "forecast_days" : 7
+        
+    } 
     responses = openmeteo.weather_api(url, params=params)
     response = responses[0]
 
-    # Hourly dataframe
+    #Hourly dataframe
+
     hourly = response.Hourly()
     hourly_relative_humidity_2m = hourly.Variables(0).ValuesAsNumpy()
     hourly_soil_temperature_0cm = hourly.Variables(1).ValuesAsNumpy()
     hourly_soil_moisture_0_to_1cm = hourly.Variables(2).ValuesAsNumpy()
+    hourly_cloud_cover = hourly.Variables(3).ValuesAsNumpy()
 
     hourly_data = {"date": pd.date_range(
-        start=pd.to_datetime(hourly.Time(), unit="s", utc=True),
-        end=pd.to_datetime(hourly.TimeEnd(), unit="s", utc=True),
-        freq=pd.Timedelta(seconds=hourly.Interval()),
-        inclusive="left"
+        start = pd.to_datetime(hourly.Time(), unit = "s", utc = True),
+        end = pd.to_datetime(hourly.TimeEnd(), unit = "s", utc = True),
+        freq = pd.Timedelta(seconds = hourly.Interval()),
+        inclusive = "left"
     )}
 
     hourly_data["relative_humidity_2m_%"] = hourly_relative_humidity_2m
     hourly_data["soil_temperature_0cm_C"] = hourly_soil_temperature_0cm
     hourly_data["soil_moisture_0_to_1cm_m³"] = hourly_soil_moisture_0_to_1cm
+    hourly_data["cloud_cover_%"] = hourly_cloud_cover
 
-    hourly_dataframe = pd.DataFrame(data=hourly_data)
+    hourly_dataframe = pd.DataFrame(data = hourly_data)
 
-    # Average hourly data per day
-    hourly_dataframe['day_date'] = hourly_dataframe['date'].dt.strftime(
-        '%Y-%m-%d')
 
-    tmp = hourly_dataframe.groupby('day_date')[
-        ['relative_humidity_2m_%', 'soil_temperature_0cm_C', 'soil_moisture_0_to_1cm_m³']].mean()
+    hourly_dataframe = pd.DataFrame(data = hourly_data)
 
-    avg_hourly_dataframe = tmp.rename(columns={'relative_humidity_2m_%': 'avg_hourly_relative_humidity_2m_%',
-                                      'soil_temperature_0cm_C': 'avg_hourly_soil_temperature_0cm_C', 'soil_moisture_0_to_1cm_m³': 'avg_hourly_soil_moisture_0_to_1cm_m³'})
+    #Average hourly data per day
+    hourly_dataframe['day_date'] = hourly_dataframe['date'].dt.strftime('%Y-%m-%d')
 
-    # Daily dataframe
+    tmp1 = hourly_dataframe.groupby('day_date')[['relative_humidity_2m_%', 'soil_temperature_0cm_C', 'soil_moisture_0_to_1cm_m³']].mean()
+
+    avg_hourly_dataframe = tmp1.rename(columns={'relative_humidity_2m_%': 'avg_hourly_relative_humidity_2m_%', 'soil_temperature_0cm_C': 'avg_hourly_soil_temperature_0cm_C', 'soil_moisture_0_to_1cm_m³': 'avg_hourly_soil_moisture_0_to_1cm_m³'})
+
+    #med cloud couverture
+    tmp2 = hourly_dataframe.groupby('day_date')[['cloud_cover_%']].median()
+
+    med_cloud_cover = tmp2.rename(columns={'cloud_cover_%': 'med_cloud_cover_%'})
+
+    #Daily dataframe 
     daily = response.Daily()
     daily_temperature_2m_max = daily.Variables(0).ValuesAsNumpy()
     daily_precipitation_sum = daily.Variables(1).ValuesAsNumpy()
     daily_wind_speed_10m_max = daily.Variables(2).ValuesAsNumpy()
     daily_sunshine_duration = daily.Variables(3).ValuesAsNumpy()
+    daily_rain_sum = daily.Variables(4).ValuesAsNumpy()
 
     daily_data = {"date": pd.date_range(
         start=pd.to_datetime(daily.Time(), unit="s", utc=True),
         end=pd.to_datetime(daily.TimeEnd(), unit="s", utc=True),
-        freq='D',
+        freq='D',  
         inclusive="left"
     )}
 
@@ -79,16 +86,17 @@ def get_info_meteo(latitude, longitude):
     daily_data["daily_precipitation_sum_mm"] = daily_precipitation_sum
     daily_data["daily_wind_speed_10m_max_km/h"] = daily_wind_speed_10m_max
     daily_data["daily_sunshine_duration_sec"] = daily_sunshine_duration
+    daily_data["daily_rain_sum_mm"] = daily_rain_sum
 
     daily_dataframe = pd.DataFrame(data=daily_data)
 
-    daily_dataframe['day_date'] = daily_dataframe['date'].dt.strftime(
-        '%Y-%m-%d')
-    total_df = pd.merge(daily_dataframe, avg_hourly_dataframe,
-                        on="day_date", how="left")
+    daily_dataframe['day_date'] = daily_dataframe['date'].dt.strftime('%Y-%m-%d')
+
+    total_df = pd.merge(daily_dataframe, avg_hourly_dataframe, on="day_date", how="left")
+    total_df = pd.merge(total_df, med_cloud_cover, on="day_date", how="left")
 
     total_df['date'] = total_df['day_date']
     total_df = total_df.drop('day_date', axis=1)
-
+    
     return total_df
-    #total_df.to_json("app/data/weather_data.json", orient='columns')
+    #total_df.to_json("weather_data.json", orient='columns')

app.py

@@ -5,9 +5,18 @@ from utils import *
 with gr.Blocks() as demo:
     temperature = gr.Number(visible=False)
     humidite = gr.Number(visible=False)
-    meteo = gr.Textbox(visible=False)
+    chutes_de_neige = gr.Number(value=0, visible=False)
+    couverture_nuageuse = gr.Number(visible=False)
+    precipitations = gr.Number(visible=False)
     gr.Markdown(
-        "## GAIA Démo \nDémo avec Mistral AI & LlamaIndex afin d'estimer les risques sanitaires d'un élevage.")
+        """
+        ## GAIA Démo \n
+        Démo avec Mistral AI & LlamaIndex afin d'estimer les risques parasitaires d'un élevage. \n
+        ### Fonctionnement: 
+            1. Rentrez votre adresse (ou les coordonnées GPS) de votre exploitation
+            2. Cochez votre type d'élevage
+            3. Posez la question liée aux risques parasitaires de votre exploitation (idée: 'Quels sont les risques')
+            4. Amusez vous à parcourir la carte et les données météo associées.""")
     with gr.Row(equal_height=True):
         with gr.Column():
             with gr.Tab("Adresse"):
@@ -44,10 +53,12 @@ with gr.Blocks() as demo:
             placeholder="Vous pouvez me poser une question, appuyez sur Entrée pour valider")
     clear = gr.ClearButton([msg, chatbot])
 
-    demo.load(on_init, [lat, lon, address], [map, fig, temperature, humidite])
-    btn.click(on_init, [lat, lon, address], [map, fig])
+    demo.load(on_init, [lat, lon, address], [
+              map, fig, temperature, humidite, couverture_nuageuse, precipitations])
+    btn.click(on_init, [lat, lon, address], [
+              map, fig, temperature, humidite, couverture_nuageuse, precipitations])
     msg.submit(respond, [type_elevage, temperature,
-               humidite, meteo, msg, chatbot], [msg, chatbot])
+               humidite, chutes_de_neige, couverture_nuageuse, precipitations, msg, chatbot], [msg, chatbot])
 
 demo.title = "Démo GAIA - Les 5 Mousquetaires"
 demo.launch()

utils.py

@@ -15,7 +15,7 @@ from llama_index.core.query_engine import RetrieverQueryEngine
 load_dotenv()
 
 Settings.llm = MistralAI(
-    max_tokens=1024,
+    max_tokens=2048,
     api_key=os.environ.get('MISTRAL_API_KEY')
 )
 
@@ -35,9 +35,12 @@ def on_init(lat, lon, address):
     df = get_info_meteo(lat, lon)
     avg_temp = df.loc[:, 'avg_hourly_soil_temperature_0cm_C'].mean()
     avg_humidity = df.loc[:, 'avg_hourly_soil_moisture_0_to_1cm_m³'].mean()
+    #avg_snowfall = df.loc[:, 'daily_precipitation_sum_mm'].mean()
+    avg_cloud_cover = df.loc[:, 'med_cloud_cover_%'].mean()
+    avg_precipitations = df.loc[:, 'daily_rain_sum_mm'].mean()
 
     print('toto')
-    return map_html, fig, avg_temp, avg_humidity
+    return map_html, fig, avg_temp, avg_humidity, avg_cloud_cover, avg_precipitations
 
 
 def print_conversation(chat_history):
@@ -48,14 +51,16 @@ def print_conversation(chat_history):
         print(response, end='\n')
 
 
-def respond(elevage, temperature, humidite, meteo, message, chat_history):
+def respond(elevage, temperature, humidite, chutes_de_neige, couverture_nuageuse, precipitations, message, chat_history):
     PROMPT = f"""
     Tu es un chatbot qui répond en français, qui doit aider à déterminer les risques parasitaires d'élevage en fonction des conditions météorologiques.
     Voici les conditions: 
     - élevage: {elevage}
-    - température moyenne : {temperature}
-    - humidité moyenne : {humidite}
-    - météo : {meteo}
+    - température moyenne : {temperature} °C
+    - humidité moyenne : {humidite} %
+    - chutes de neige : {chutes_de_neige} cm
+    - couverture nuageuse : {couverture_nuageuse} %
+    - precipitations : {precipitations} mm
 
     Fais moi un tableau avec chaque parasite et/ou tiques susceptible de toucher mon élevage dans ces conditions météo, avec les conditions de prolifération, la vitesse de propagation, les troubles causés par ces parasites
     Essaie d'écrire des données numériques précises dans le tableau