analyse_frequency.py

"""
Signal Frequency Analysis for Home Assistant histopry graph CSV files
========================================================

This script analyzes the frequency components of error signals exported from Home Assistant's history graph.

- GUI file picker for CSV selection
- Handles Home Assistant history graph CSV format (columns: entity_id, state, last_changed)
- Plots error signal and its spectrum
- Prints top 5 dominant frequency components
"""

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from scipy.fft import fft, fftfreq
import tkinter as tk
from tkinter import filedialog, messagebox
import os


def main():
    # GUI to select CSV file
    root = tk.Tk()
    root.withdraw()  # Hide the main window
    file_path = filedialog.askopenfilename(
        title="Select Home Assistant history graph CSV",
        filetypes=[("CSV Files", "*.csv"), ("All Files", "*")],
    )

    # Try to use example.csv if no file is selected
    if not file_path:
        example_path = os.path.join(os.path.dirname(__file__), "example.csv")
        if os.path.exists(example_path):
            use_example = messagebox.askyesno(
                "Use Example?", "No file selected. Use example.csv instead?"
            )
            if use_example:
                file_path = example_path
            else:
                messagebox.showerror(
                    "No file selected", "No CSV file was selected. Exiting."
                )
                exit()
        else:
            messagebox.showerror(
                "No file selected",
                "No CSV file was selected and example.csv not found. Exiting.",
            )
            exit()

    # Load the error signal from CSV
    # Adjusted to handle CSV with columns: entity_id,state,last_changed

    try:
        df = pd.read_csv(file_path)
        if not all(col in df.columns for col in ["state", "last_changed"]):
            raise ValueError("CSV must contain 'state' and 'last_changed' columns.")
        signal = df["state"].astype(float).values
        timestamps = pd.to_datetime(df["last_changed"])
        time_seconds = (timestamps - timestamps.iloc[0]).dt.total_seconds().values
    except Exception as e:
        messagebox.showerror("CSV Error", f"Failed to load or parse CSV: {e}")
        exit()

    # Compute average sample interval (dt)
    dt = np.mean(np.diff(time_seconds))

    # Remove the t=0 point from time and signal arrays
    if time_seconds[0] == 0:
        time_seconds = time_seconds[1:]
        signal = signal[1:]

    # Plot the raw error signal and frequency spectrum side by side
    fig, axs = plt.subplots(1, 2, figsize=(14, 4))
    axs[0].plot(time_seconds, signal)
    axs[0].set_title("PID Error Signal")
    axs[0].set_xlabel("Time (s)")
    axs[0].set_ylabel("Error")

    # Frequency analysis with correct dt
    N = len(signal)
    freqs = fftfreq(N, dt)
    fft_values = fft(signal)
    idx = np.where(freqs >= 0)
    axs[1].plot(freqs[idx], np.abs(fft_values[idx]))
    axs[1].set_title("Frequency Spectrum of Error Signal")
    axs[1].set_xlabel("Frequency (Hz)")
    axs[1].set_ylabel("Amplitude")

    plt.tight_layout()
    plt.show()

    # Print dominant frequency components
    amplitudes = np.abs(fft_values[idx])
    dominant_indices = np.argsort(amplitudes)[-5:][::-1]
    print("Top 5 dominant frequencies:")
    for i in dominant_indices:
        print(f"Frequency: {freqs[idx][i]:.4f} Hz, Amplitude: {amplitudes[i]:.2f}")


if __name__ == "__main__":
    main()