mroe things

data_loader.py

@@ -0,0 +1,43 @@
+import pandas as pd
+import io
+from typing import Union
+
+
+def cut_headers(stream: Union[io.StringIO, io.FileIO, io.TextIOWrapper]):
+    """Remove custom header metadata from the CSV file."""
+    headers = []
+    exist_startheader = False
+    while line := stream.readline():
+        if "-BEGIN HEADER-" in line and not exist_startheader:
+            exist_startheader = True
+            headers.append(str(line.strip()))
+        elif "-END HEADER-" in line and exist_startheader:
+            headers.append(str(line.strip()))
+            break
+        elif exist_startheader:
+            headers.append(str(line.strip()))
+        elif not exist_startheader:
+            continue
+    if not exist_startheader:
+        stream.seek(0)
+        return "", stream
+    return "\n".join(headers), stream
+
+
+def load_weather_data(file_path: str) -> pd.DataFrame:
+    """Load weather data and preprocess the datetime index."""
+    with open(file_path, "r", encoding="utf-8") as file:
+        _, csv_file = cut_headers(file)
+        data = pd.read_csv(csv_file, na_values=-999)
+
+    # Convert YEAR, MO, DY to datetime index
+    if all(col in data.columns for col in ("YEAR", "MO", "DY")):
+        data["date"] = pd.to_datetime(
+            {k: data[v] for v, k in zip(["YEAR", "MO", "DY"], ["year", "month", "day"])}
+        )
+        data.drop(columns=["YEAR", "MO", "DY"], inplace=True)
+        data.set_index("date", inplace=True)
+
+    data.dropna(inplace=True)
+    data = data.asfreq("D")  # Assumes regular daily data
+    return data

dataset_update.py

@@ -1,11 +1,12 @@
 import urllib.parse
 import urllib.request
 from datetime import datetime
-from global_config import DATA_FILE_NAME
+
+data_file = "hanoi_weather_data.csv"
 
 # Define parameters
 latitude, longitude = 21.04815477090267, 105.80123159165096  # lmao
-start_date = "1981-01-01"
+start_date = "1985-01-01"
 end_date = datetime.now().strftime("%Y-%m-%d")  # Today's date
 
 # Define desired parameters
@@ -18,19 +19,7 @@ end_date = datetime.now().strftime("%Y-%m-%d")  # Today's date
 # WD10M           MERRA-2 Wind Direction at 10 Meters (Degrees)
 # more at https://power.larc.nasa.gov/data-access-viewer/
 weather_params = [
-    "T2M",
-    "T2M_MAX",
-    "T2M_MIN",
-    "T2MDEW",
-    "PRECTOTCORR",
     "QV2M",
-    "RH2M",
-    "PS",
-    "CLOUD_AMT",
-    "WS10M",
-    "WD10M",
-    "ALLSKY_KT",
-    "ALLSKY_SFC_SW_DWN",
 ]
 
 params = {
@@ -52,13 +41,14 @@ base_url = "https://power.larc.nasa.gov/api/temporal/daily/point"
 url = f"{base_url}?{query_string}"
 
 # Make the API request using urllib
+if __name__ == "__main__":
     try:
         with urllib.request.urlopen(url) as response:
             if response.status == 200:
-            csv_data = response.read().decode("utf-8")
+                csv_data = response.read()
-            with open(DATA_FILE_NAME, "w", encoding="utf-8") as file:
+                with open(data_file, "wb") as file:
                     file.write(csv_data)
-            print("Data successfully downloaded and saved to 'hanoi_weather_data.csv'.")
+                print(f"Data successfully downloaded and saved to {data_file}.")
             else:
                 print(f"Failed to retrieve data: HTTP Status Code {response.status}")
     except urllib.error.URLError as e:

dead/sk2_port.py

@@ -0,0 +1,185 @@
+import sys
+import numpy as np
+import pandas as pd
+import matplotlib.pyplot as plt
+import seaborn as sns
+import io
+from typing import Union
+from sklearn.metrics import mean_squared_error, r2_score
+from sklearn.preprocessing import MinMaxScaler
+from sklearn.model_selection import train_test_split
+from sklearn.linear_model import LinearRegression
+from statsmodels.tsa.ar_model import AutoReg
+
+# ==== LOAD & CLEAN DATA ====
+DATA_FILE_NAME = "hanoi_weather_data.csv"
+
+def cut_headers(stream: Union[io.StringIO, io.FileIO, io.TextIOWrapper]):
+    headers = []
+    exist_startheader = False
+    while line := stream.readline():
+        if "-BEGIN HEADER-" in line and not exist_startheader:
+            exist_startheader = True
+            headers.append(str(line.strip()))
+        elif "-END HEADER-" in line and exist_startheader:
+            headers.append(str(line.strip()))
+            break
+        elif exist_startheader:
+            headers.append(str(line.strip()))
+        elif not exist_startheader:
+            continue
+    if not exist_startheader:
+        stream.seek(0)
+        return "", stream
+    return "\n".join(headers), stream
+
+with open(DATA_FILE_NAME, "r", encoding="utf-8") as file:
+    data_header, csv_file = cut_headers(file)
+    data = pd.read_csv(csv_file, na_values=-999)
+
+if all(z in data.columns for z in ("YEAR", "MO", "DY")):
+    data["date"] = pd.to_datetime(
+        {k: data[v] for v, k in zip(["YEAR", "MO", "DY"], ["year", "month", "day"])}
+    )
+    data.drop(columns=["YEAR", "MO", "DY"], inplace=True)
+    data.set_index("date", inplace=True)
+
+data.dropna(inplace=True)
+
+# ==== VISUALIZE CORRELATION ====
+plt.figure(figsize=(12, 8))
+sns.heatmap(data.corr(), annot=True, cmap="coolwarm")
+plt.title("Feature Correlation Heatmap")
+plt.show()
+
+# ==== SCALE DATA ====
+scaler = MinMaxScaler()
+scaled_data = scaler.fit_transform(data)
+
+# ==== CREATE SEQUENCES ====
+def create_sequences(data, seq_length):
+    xs, ys = [], []
+    for i in range(len(data) - seq_length):
+        x = data[i:i + seq_length].flatten()
+        y = data[i + seq_length]
+        xs.append(x)
+        ys.append(y)
+    return np.array(xs), np.array(ys)
+
+seq_length = 50
+X, y = create_sequences(scaled_data, seq_length)
+
+# ==== SPLIT ====
+date_index = data.index[seq_length:]
+X_train, X_test, y_train, y_test, date_train, date_test = train_test_split(
+    X, y, date_index, test_size=0.02, shuffle=False
+)
+
+# ==== TRAIN LINEAR REGRESSION ====
+model = LinearRegression()
+model.fit(X_train, y_train)
+predictions = model.predict(X_test)
+
+# ==== EVALUATION ====
+mse = mean_squared_error(y_test, predictions)
+r2 = r2_score(y_test, predictions)
+
+print(f"Test MSE: {mse:.4f}")
+print(f"Test R2: {r2:.4f}")
+print(f"Model Score: {model.score(X_test, y_test):.4f}")
+
+# ==== INVERSE TRANSFORM ====
+predictions_rescaled = scaler.inverse_transform(
+    predictions.reshape(-1, data.shape[1])
+)
+y_test_rescaled = scaler.inverse_transform(
+    y_test.reshape(-1, data.shape[1])
+)
+
+# ==== PLOTS ====
+
+# Actual vs Predicted with Real Dates
+plt.figure(figsize=(12, 6))
+plt.plot(date_test, y_test_rescaled[:, 0], label="Actual")
+plt.plot(date_test, predictions_rescaled[:, 0], label="Predicted")
+plt.title("Actual vs Predicted (First Feature)")
+plt.xlabel("Date")
+plt.ylabel("Original Scale")
+plt.xticks(rotation=45)
+plt.legend()
+plt.grid(True)
+plt.tight_layout()
+plt.show()
+
+# Residual Plot
+residuals = y_test[:, 0] - predictions[:, 0]
+plt.figure(figsize=(10, 4))
+plt.plot(date_test, residuals, label="Residuals")
+plt.hlines(0, 0, len(residuals), colors="red")
+plt.title("Residuals Over Time (First Feature)")
+plt.xlabel("Date")
+plt.ylabel("Error")
+plt.xticks(rotation=45)
+plt.legend()
+plt.tight_layout()
+plt.grid(True)
+plt.show()
+
+# Error Distribution
+errors = np.abs(y_test[:, 0] - predictions[:, 0])
+plt.hist(errors, bins=50)
+plt.title("Error Distribution (First Feature)")
+plt.xlabel("Absolute Error")
+plt.ylabel("Frequency")
+plt.grid(True)
+plt.show()
+
+# Scatter Plot
+plt.scatter(y_test[:, 0], predictions[:, 0], alpha=0.5)
+plt.plot([0, 1], [0, 1], "r--")
+plt.title("Predicted vs Actual (Normalized)")
+plt.xlabel("Actual")
+plt.ylabel("Predicted")
+plt.grid(True)
+plt.show()
+
+# ==== AUTOREGRESSION FORECAST ====
+target_column = data.columns[0]
+series = data[target_column]
+
+train_size = int(len(series) * 0.98)
+train_series = series[:train_size]
+test_series = series[train_size:]
+
+ar_model = AutoReg(train_series, lags=30).fit()
+ar_preds = ar_model.predict(start=len(train_series), end=len(series)-1)
+
+# Plot AR Forecast
+plt.figure(figsize=(12, 5))
+plt.plot(test_series.index, test_series, label="Actual")
+plt.plot(test_series.index, ar_preds, label="AR Forecast", linestyle="--")
+plt.title(f"AutoRegression Forecast - {target_column}")
+plt.xlabel("Date")
+plt.ylabel(target_column)
+plt.xticks(rotation=45)
+plt.legend()
+plt.grid(True)
+plt.tight_layout()
+plt.show()
+
+# ==== FUTURE FORECAST ====
+future_steps = 30
+future_index = pd.date_range(start=series.index[-1], periods=future_steps+1, freq="D")[1:]
+future_forecast = ar_model.predict(start=len(series), end=len(series) + future_steps - 1)
+
+plt.figure(figsize=(12, 5))
+plt.plot(series[-100:], label="Recent Data")
+plt.plot(future_index, future_forecast, label="Future Forecast", color="red")
+plt.title(f"{target_column} - 30 Day Forecast (AutoRegression)")
+plt.xlabel("Date")
+plt.ylabel(target_column)
+plt.xticks(rotation=45)
+plt.legend()
+plt.tight_layout()
+plt.grid(True)
+plt.show()

dead/sklearn_port.py

@@ -7,8 +7,37 @@ from sklearn.model_selection import train_test_split
 from sklearn.linear_model import LinearRegression
 from sklearn.metrics import mean_squared_error
 import matplotlib.pyplot as plt
-from global_config import DATA_FILE_NAME
+# from global_config import DATA_FILE_NAME
-from load_data import cut_headers
+DATA_FILE_NAME = "hanoi_weather_data.csv"
+
+# from load_data import cut_headers
+
+import io
+
+# from itertools import islice
+from typing import Union
+
+def cut_headers(stream: Union[io.StringIO, io.FileIO, io.TextIOWrapper]):
+    headers = []
+    exist_startheader = False
+
+    while line := stream.readline():
+        if "-BEGIN HEADER-" in line and not exist_startheader:
+            exist_startheader = True
+            headers.append(str(line.strip()))
+        elif "-END HEADER-" in line and exist_startheader:
+            headers.append(str(line.strip()))
+            break
+        elif exist_startheader:
+            # print(f"{line = }")
+            headers.append(str(line.strip()))
+        elif not exist_startheader:
+            continue
+    if not exist_startheader:
+        # If no headers are found, reset the stream and return
+        stream.seek(0)
+        return "", stream
+    return "\n".join(headers), stream
 
 # 1. Load data
 
@@ -34,7 +63,7 @@ scaled_data = scaler.fit_transform(data)
 def create_sequences(data, seq_length):
     xs, ys = [], []
     for i in range(len(data) - seq_length):
-        x = data[i : i + seq_length]
+        x = data[i : i + seq_length].flatten()
         y = data[i + seq_length]
         xs.append(x)
         ys.append(y)
@@ -67,21 +96,15 @@ predictions_rescaled = scaler.inverse_transform(predictions.reshape(-1, data.sha
 y_test_rescaled = scaler.inverse_transform(y_test.reshape(-1, data.shape[1]))
 
 # 9. Plot results
-fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5), width_ratios=[1, 3])
 
 # Loss bar
-ax1.bar(["Linear Regression"], [mse], color="orange")
+
-ax1.set_title("Validation MSE")
-ax1.set_ylabel("Loss")
-ax1.grid(True)
 
 # Prediction vs Actual
-ax2.plot(y_test_rescaled, label="Actual")
+plt.plot(y_test_rescaled, label="Actual")
-ax2.plot(predictions_rescaled, label="Predicted")
+plt.plot(predictions_rescaled, label="Predicted")
-ax2.set_title("Actual vs. Predicted")
+plt.title("Actual vs. Predicted")
-ax2.set_xlabel("Sample")
+plt.xlabel("Sample")
-ax2.set_ylabel("Value")
+plt.ylabel("Value")
-ax2.legend()
+plt.legend()
-
-plt.tight_layout()
 plt.show()

main.py

@@ -0,0 +1,71 @@
+# import pandas as pd
+from data_loader import load_weather_data
+from preprocessing import scale_data, create_sequences, inverse_scale
+from model_train import (
+    train_linear_regression_gd,
+    predict_linear_regression,
+    evaluate_model,
+)
+from visualize import (
+    plot_original_vs_scaled,
+    plot_predictions,
+    plot_residuals,
+    plot_error_hist,
+    plot_scatter,
+    plot_loss_curve,
+    plot_value_distribution,
+    visualize_sequence_window,
+)
+from sklearn.model_selection import train_test_split
+
+# Step 1: Load data
+data = load_weather_data("hanoi_weather_data.csv")
+
+plot_value_distribution(data)
+
+# Step 2: Preprocess (scale and sequence)
+scaled_data, scaler = scale_data(data)
+plot_original_vs_scaled(data, scaled_data)
+
+seq_length = 50
+X, y = create_sequences(scaled_data, seq_length)
+
+visualize_sequence_window(data["QV2M"], seq_length=10, start_idx=1000)
+
+# Step 3: Train/test split
+date_index = data.index[seq_length:]
+X_train, X_test, y_train, y_test = train_test_split(
+    X, y.reshape(-1), test_size=0.02, shuffle=False
+)
+date_train, date_test = train_test_split(date_index, test_size=0.02, shuffle=False)
+
+# Step 4: Train model with gradient descent
+# import math
+# lr_schedule = lambda epoch: max(0.005, 0.01 * (1 + math.cos(epoch / 1000 * math.pi)) / 2)
+
+theta, losses = train_linear_regression_gd(
+    X_train,
+    y_train,
+    # lr_schedule=lr_schedule,
+    n_epochs=10000,
+)
+
+predictions = predict_linear_regression(X_test, theta)
+
+# Step 5: Evaluate
+mse, r2 = evaluate_model(y_test, predictions)
+print(f"MSE: {mse:.4f}, R2: {r2:.4f}")
+
+# Step 6: Rescale predictions
+y_test_rescaled = inverse_scale(y_test, scaler, data.shape[1])
+predictions_rescaled = inverse_scale(predictions, scaler, data.shape[1])
+
+# Step 7: Visualize
+residuals = y_test - predictions
+errors = abs(residuals)
+
+plot_predictions(date_test, y_test_rescaled, predictions_rescaled)
+plot_residuals(date_test.to_numpy(), residuals)
+plot_error_hist(errors)
+plot_scatter(y_test.reshape(-1, 1), predictions.reshape(-1, 1))
+plot_loss_curve(losses)

model_train.py

@@ -0,0 +1,45 @@
+import numpy as np
+from sklearn.metrics import mean_squared_error, r2_score
+
+
+def train_linear_regression_gd(X_train, y_train, lr_schedule= lambda _e: 0.01, n_epochs=1000):
+    """
+    Train linear regression model using batch gradient descent.
+
+    Returns:
+        theta: learned weights
+        losses: list of MSE loss per epoch
+    """
+    m, n = X_train.shape
+    X_b = np.hstack([np.ones((m, 1)), X_train])
+    theta = np.zeros(n + 1)
+    losses = []
+
+    for epoch in range(n_epochs):
+        lr = lr_schedule(epoch)
+        predictions = X_b @ theta
+        errors = predictions - y_train
+        gradients = (2 / m) * X_b.T @ errors
+        gradients = np.clip(gradients, -1, 1)  # gradient clipping
+        theta -= lr * gradients
+        loss = np.mean(errors ** 2)
+        losses.append(loss)
+
+        if epoch % 1000 == 0:
+            print(f"Epoch {epoch}: Loss = {loss:.6f}, LR: {lr:.6f}")
+
+    print(f"Final Loss: {losses[-1]:.6f}")
+    return theta, losses
+
+
+def predict_linear_regression(X, theta):
+    """Predict using learned linear model."""
+    X_b = np.hstack([np.ones((X.shape[0], 1)), X])
+    return X_b @ theta
+
+
+def evaluate_model(y_true, y_pred):
+    """Evaluate using sklearn's built-in MSE and R² functions."""
+    mse = mean_squared_error(y_true, y_pred)
+    r2 = r2_score(y_true, y_pred)
+    return mse, r2

preprocessing.py

@@ -0,0 +1,26 @@
+import numpy as np
+from sklearn.preprocessing import MinMaxScaler
+from sklearn.model_selection import train_test_split
+
+
+def scale_data(data):
+    """Normalize the data using MinMaxScaler."""
+    scaler = MinMaxScaler()
+    scaled_data = scaler.fit_transform(data)
+    return scaled_data, scaler
+
+
+def inverse_scale(scaled, scaler, n_features):
+    """Inverse transform using stored scaler."""
+    return scaler.inverse_transform(scaled.reshape(-1, n_features))
+
+
+def create_sequences(data, seq_length):
+    """Transform the time series into overlapping sequences."""
+    xs, ys = [], []
+    for i in range(len(data) - seq_length):
+        x = data[i : i + seq_length].flatten()
+        y = data[i + seq_length]
+        xs.append(x)
+        ys.append(y)
+    return np.array(xs), np.array(ys)

visualize.py

@@ -0,0 +1,137 @@
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def plot_predictions(date_test, y_test_rescaled, predictions_rescaled):
+    plt.figure(figsize=(12, 6))
+    plt.plot(date_test, y_test_rescaled[:, 0], label="Actual")
+    plt.plot(date_test, predictions_rescaled[:, 0], label="Predicted")
+    plt.title("Actual vs Predicted (First Feature)")
+    plt.xlabel("Date")
+    plt.ylabel("Value")
+    plt.xticks(rotation=45)
+    plt.legend()
+    plt.tight_layout()
+    plt.grid(True)
+    plt.show()
+
+
+def plot_residuals(date_test, residuals):
+    plt.figure(figsize=(10, 4))
+    plt.plot(date_test, residuals, label="Residuals")
+    plt.hlines(0, 0, len(residuals), colors="red")
+    plt.xlim(date_test.min(), date_test.max())
+    plt.title("Residuals Over Time")
+    plt.xlabel("Date")
+    plt.ylabel("Error")
+    plt.xticks(rotation=45)
+    plt.grid(True)
+    plt.tight_layout()
+    plt.legend()
+    plt.show()
+
+
+def plot_error_hist(errors):
+    plt.hist(errors, bins=50)
+    plt.title("Error Distribution")
+    plt.xlabel("Absolute Error")
+    plt.ylabel("Frequency")
+    plt.grid(True)
+    plt.show()
+
+
+def plot_scatter(y_test, predictions):
+    plt.scatter(y_test[:, 0], predictions[:, 0], alpha=0.5)
+    plt.plot([0, 1], [0, 1], "r--")
+    plt.title("Predicted vs Actual (Normalized)")
+    plt.xlabel("Actual")
+    plt.ylabel("Predicted")
+    plt.grid(True)
+    plt.show()
+
+
+def plot_loss_curve(losses):
+    plt.plot(losses)
+    plt.title("Training Loss Curve")
+    plt.xlabel("Epoch")
+    plt.ylabel("MSE Loss")
+    plt.grid(True)
+    plt.tight_layout()
+    plt.show()
+
+
+def plot_original_vs_scaled(original_data, scaled_data, feature_names=None):
+    """
+    Plot original and scaled data, aligned by date.
+    """
+    if feature_names is None:
+        feature_names = original_data.columns[:2]
+
+    dates = original_data.index  # Grab the datetime index
+
+    fig, axs = plt.subplots(2, 1, figsize=(12, 6), sharex=True)
+
+    for col in feature_names:
+        axs[0].plot(dates, original_data[col], label=f"Original {col}")
+    axs[0].set_title("Original Data")
+    axs[0].legend()
+    axs[0].grid(True)
+
+    for i, col in enumerate(feature_names):
+        col_index = original_data.columns.get_loc(col)
+        axs[1].plot(dates, scaled_data[:, col_index], label=f"Scaled {col}")
+    axs[1].set_title("Scaled Data (MinMaxScaler)")
+    axs[1].legend()
+    axs[1].grid(True)
+
+    plt.xlabel("Date")
+    plt.tight_layout()
+    plt.show()
+
+
+def plot_value_distribution(data):
+    # import matplotlib.pyplot as plt
+
+    plt.figure(figsize=(8, 4))
+    plt.hist(data.iloc[:, 0], bins=40, color="skyblue", edgecolor="black")
+    plt.title("Distribution of Values")
+    plt.xlabel(data.columns[0])
+    plt.ylabel("Frequency")
+    plt.grid(True)
+    plt.tight_layout()
+    plt.show()
+
+
+def visualize_sequence_window(series, seq_length=5, start_idx=0):
+    """
+    Visualize how a sequence window is used to create X (input) and y (target).
+
+    Parameters:
+        series (pd.Series or np.ndarray): 1D time series data
+        seq_length (int): Length of input window
+        start_idx (int): Where to start slicing from for demo
+    """
+    if isinstance(series, np.ndarray):
+        data = series
+    else:
+        data = series.to_numpy()
+
+    x_seq = data[start_idx:start_idx + seq_length]
+    y_val = data[start_idx + seq_length]
+
+    full_seq = np.concatenate([x_seq, [y_val]])
+
+    plt.figure(figsize=(8, 3))
+    plt.plot(range(start_idx, start_idx + seq_length + 1), full_seq, marker='o')
+    plt.axvline(x=start_idx + seq_length - 0.5, linestyle="--", color="gray")
+
+    for i in range(seq_length):
+        plt.text(start_idx + i, x_seq[i], f"x[{i}]", ha="center", va="bottom", fontsize=9)
+    plt.text(start_idx + seq_length, y_val, "y", ha="center", va="bottom", color="red", fontsize=9)
+
+    plt.title(f"Sequence Window (seq_length={seq_length}) → Predict Next")
+    plt.xlabel("Time Step")
+    plt.ylabel("Value")
+    plt.grid(True)
+    plt.tight_layout()
+    plt.show()