Update app.py

app.py

@@ -1,35 +1,226 @@
-# 19feb2023
-#https://huggingface.co/spaces/keras-io/timeseries_forecasting_for_weather/
 
+#%%
+from matplotlib.pyplot import title
+import tensorflow as tf
+from tensorflow import keras
+from huggingface_hub import from_pretrained_keras
+import pandas as pd
+import matplotlib.pyplot as plt
 import streamlit as st
+from zipfile import ZipFile
+import os
 import datetime
-import pandas as pd
-import numpy as np
 
-backlogmax = 4
-today = datetime.date.today()
+import warnings
+warnings.filterwarnings("ignore")
+
+#
+mylist = [0, 1, 5, 7, 8, 10, 11]
+mylist = [0, 2]
+df = pd.DataFrame(columns=["Date Time","p (mbar)","T (degC)","Tpot (K)","Tdew (degC)","rh (%)","VPmax (mbar)","VPact (mbar)","VPdef (mbar)","sh (g/kg)","H2OC (mmol/mol)","rho (g/m**3)","wv (m/s)","max. wv (m/s)","wd (deg)"])
+
+if ("0" == ""):
+	uri = "https://storage.googleapis.com/tensorflow/tf-keras-datasets/jena_climate_2009_2016.csv.zip"
+	zip_path = keras.utils.get_file(origin=uri, fname="jena_climate_2009_2016.csv.zip")
+	zip_file = ZipFile(zip_path)
+	zip_file.extractall()
+	csv_path = "jena_climate_2009_2016.csv"
+	df = pd.read_csv(csv_path)
+
+if ("0" != ""):
+	backlogmax = 4
+	today = datetime.date.today()
+
+	ayear = int(today.strftime("%Y"))-0
+	amonth = int(today.strftime("%m"))
+	amonthday = int(today.strftime("%d"))
+
+	adf = pd.DataFrame(columns=["Date Time","p (mbar)","T (degC)","Tpot (K)","Tdew (degC)","rh (%)","VPmax (mbar)","VPact (mbar)","VPdef (mbar)","sh (g/kg)","H2OC (mmol/mol)","rho (g/m**3)","wv (m/s)","max. wv (m/s)","wd (deg)"])
+	for i in range(ayear-backlogmax,ayear,1):
+		alink = ("https://data.weather.gov.hk/weatherAPI/opendata/opendata.php?dataType=CLMTEMP&year={}&rformat=csv&station=HKO").format(str(i))
+		df = pd.read_csv(alink, skiprows=[0,1,2], skipfooter=3, engine='python', on_bad_lines='skip')
+
+		df = df.reset_index()  # make sure indexes pair with number of rows
+		for index, row in df.iterrows():
+			if (row[2]!=amonth) or (row[3]!=amonthday):
+				continue
+
+			adate = ("{:02d}.{:02d}.{} 00:00:00").format(row[3], row[2], row[1])
+			st.write(adate)
+			adf = adf.append({"Date Time":adate,"T (degC)":row[4],}, ignore_index=True)
+			break
+	df = adf
+
+#%%
+
+title = "Timeseries forecasting for weather prediction"
+
+st.title('Timeseries forecasting for weather prediction')
+
+st.write("Demonstrates how to do timeseries forecasting using a [LSTM model.](https://keras.io/api/layers/recurrent_layers/lstm/#lstm-class)This space demonstration is forecasting for weather prediction. *n* observation is selected from validation dataset." )
+st.write("Keras example authors: [Prabhanshu Attri, Yashika Sharma, Kristi Takach, Falak Shah](https://keras.io/examples/timeseries/timeseries_weather_forecasting/)")
+
+
+# %% model 
+
+titles = [
+    "Pressure",
+    "Temperature",
+    "Temperature in Kelvin",
+    "Temperature (dew point)",
+    "Relative Humidity",
+    "Saturation vapor pressure",
+    "Vapor pressure",
+    "Vapor pressure deficit",
+    "Specific humidity",
+    "Water vapor concentration",
+    "Airtight",
+    "Wind speed",
+    "Maximum wind speed",
+    "Wind direction in degrees",
+]
+
+feature_keys = [
+    "p (mbar)",
+    "T (degC)",
+    "Tpot (K)",
+    "Tdew (degC)",
+    "rh (%)",
+    "VPmax (mbar)",
+    "VPact (mbar)",
+    "VPdef (mbar)",
+    "sh (g/kg)",
+    "H2OC (mmol/mol)",
+    "rho (g/m**3)",
+    "wv (m/s)",
+    "max. wv (m/s)",
+    "wd (deg)",
+]
+
+date_time_key = "Date Time"
+split_fraction = 0.715
+train_split = int(split_fraction * int(df.shape[0]))
+step = 6
+
+past = 720
+future = 72
+learning_rate = 0.001
+batch_size = 256
+epochs = 10
+
+
+def normalize(data, train_split):
+    data_mean = data[:train_split].mean(axis=0)
+    data_std = data[:train_split].std(axis=0)
+    return (data - data_mean) / data_std
+
+
+print(
+    "The selected parameters are:",
+    ", ".join([titles[i] for i in mylist]),
+)
+selected_features = [feature_keys[i] for i in mylist]
+features = df[selected_features]
+features.index = df[date_time_key]
+features.head()
+
+features = normalize(features.values, train_split)
+features = pd.DataFrame(features)
+features.head()
+
+train_data = features.loc[0 : train_split - 1]
+val_data = features.loc[train_split:]
+
+
+split_fraction = 0.715
+train_split = int(split_fraction * int(df.shape[0]))
+step = 6
+
+past = 720
+future = 72
+learning_rate = 0.001
+batch_size = 256
+epochs = 10
+
+
+def normalize(data, train_split):
+    data_mean = data[:train_split].mean(axis=0)
+    data_std = data[:train_split].std(axis=0)
+    return (data - data_mean) / data_std
+print(
+    "The selected parameters are:",
+    ", ".join([titles[i] for i in mylist]),
+)
+selected_features = [feature_keys[i] for i in mylist]
+features = df[selected_features]
+features.index = df[date_time_key]
+features.head()
+
+features = normalize(features.values, train_split)
+features = pd.DataFrame(features)
+features.head()
+
+train_data = features.loc[0 : train_split - 1]
+val_data = features.loc[train_split:]
+start = past + future
+end = start + train_split
+
+x_train = train_data[[i for i in range(7)]].values
+y_train = features.iloc[start:end][[1]]
+
+sequence_length = int(past / step)
+x_end = len(val_data) - past - future
+
+label_start = train_split + past + future
+
+x_val = val_data.iloc[:x_end][[i for i in range(7)]].values
+y_val = features.iloc[label_start:][[1]]
+
+dataset_val = keras.preprocessing.timeseries_dataset_from_array(
+    x_val,
+    y_val,
+    sequence_length=sequence_length,
+    sampling_rate=step,
+    batch_size=batch_size,
+)
+#%%
+model = from_pretrained_keras("keras-io/timeseries_forecasting_for_weather")
 
-ayear = int(today.strftime("%Y"))-0
-amonth = int(today.strftime("%m"))
-amonthday = int(today.strftime("%d"))
+#%% 
+st.set_option('deprecation.showPyplotGlobalUse', False)
+def plot():
+    n = st.sidebar.slider("Step", min_value = 1, max_value=5, value = 1)
+    def show_plot(plot_data, delta, title):
+        labels = ["History", "True Future", "Model Prediction"]
+        marker = [".-", "rx", "go"]
+        time_steps = list(range(-(plot_data[0].shape[0]), 0))
+        if delta:
+            future = delta
+        else:
+            future = 0
 
-st.write(type(ayear))
-st.write(("{}-{}-{}").format(ayear,amonth,amonthday))
-adf = pd.DataFrame(columns=["Date Time","p (mbar)","T (degC)","Tpot (K)","Tdew (degC)","rh (%)","VPmax (mbar)","VPact (mbar)","VPdef (mbar)","sh (g/kg)","H2OC (mmol/mol)","rho (g/m**3)","wv (m/s)","max. wv (m/s)","wd (deg)"])
+        plt.title(title)
+        for i, val in enumerate(plot_data):
+            if i:
+                plt.plot(future, plot_data[i], marker[i], markersize=10, label=labels[i])
+            else:
+                plt.plot(time_steps, plot_data[i].flatten(), marker[i], label=labels[i])
+        plt.legend(loc='lower center', bbox_to_anchor=(0.5, 1.05),
+          ncol=3, fancybox=True, shadow=True)
+        plt.xlim([time_steps[0], (future + 5) * 2])
+        plt.xlabel("Time-Step")
+        plt.show()
+        return
 
-for i in range(ayear-backlogmax,ayear,1):
-	alink = ("https://data.weather.gov.hk/weatherAPI/opendata/opendata.php?dataType=CLMTEMP&year={}&rformat=csv&station=HKO").format(str(i))
-	df = pd.read_csv(alink, skiprows=[0,1,2], skipfooter=3, engine='python', on_bad_lines='skip')
-	st.write(i)
 
-	df = df.reset_index()  # make sure indexes pair with number of rows
-	for index, row in df.iterrows():
-		if (row[2]!=amonth) or (row[3]!=amonthday):
-			continue
+    for x, y in dataset_val.take(n):
+        show_plot(
+            [x[0][:, 1].numpy(), y[0].numpy(), model.predict(x)[0]],
+            12,
+            f"{n} Step Prediction",
+        )
 
-		st.write(row[0],row[1],row[2],row[3],row[4],amonth,amonthday)
-		adate = ("{}.{}.{} 00:00:00").format(row[3], row[2], row[1])
-		adf = adf.append({"Date Time":adate,"T (degC)":row[4],}, ignore_index=True)
-		break
+fig = plot()
+st.pyplot(fig)
 
-st.dataframe(adf)
+# %%