scripts/make_lamp_quality_assessment.py

import json
from pathlib import Path

import numpy as np
import pandas as pd
import plotly.graph_objects as go

input_path = Path('triangulated_lamps.csv')
output_dir = Path('output')
output_dir.mkdir(exist_ok=True)
chart_path = output_dir / 'lamp_quality_rank.png'
meta_path = output_dir / 'lamp_quality_rank.png.meta.json'
csv_out = output_dir / 'lamp_quality_assessment.csv'
summary_out = output_dir / 'lamp_quality_summary.csv'

if not input_path.exists():
    raise FileNotFoundError(f'Không tìm thấy file input: {input_path.resolve()}')

df = pd.read_csv(input_path)
required_cols = ['lamp_id', 'mean_ray_distance_m', 'x_m_local', 'y_m_local']
for c in required_cols:
    if c not in df.columns:
        raise ValueError(f'Thiếu cột bắt buộc: {c}')

df['lamp_id'] = df['lamp_id'].astype(str)
df['lamp_num'] = pd.to_numeric(df['lamp_id'], errors='coerce')
df['mean_ray_distance_m'] = pd.to_numeric(df['mean_ray_distance_m'], errors='coerce')
df['x_m_local'] = pd.to_numeric(df['x_m_local'], errors='coerce')
df['y_m_local'] = pd.to_numeric(df['y_m_local'], errors='coerce')
df = df.dropna(subset=['lamp_num', 'mean_ray_distance_m']).copy()
df = df.sort_values('lamp_num').reset_index(drop=True)

q1 = df['mean_ray_distance_m'].quantile(0.25)
q3 = df['mean_ray_distance_m'].quantile(0.75)
iqr = q3 - q1
warn_thr = q3
bad_thr = q3 + 1.5 * iqr

def label_quality(v):
    if v > bad_thr:
        return 'bad'
    if v > warn_thr:
        return 'warning'
    return 'good'

df['quality_label'] = df['mean_ray_distance_m'].apply(label_quality)
df['rank_worst_first'] = df['mean_ray_distance_m'].rank(method='dense', ascending=False).astype(int)
df['z_score'] = (df['mean_ray_distance_m'] - df['mean_ray_distance_m'].mean()) / df['mean_ray_distance_m'].std(ddof=0)
df['is_outlier_iqr'] = df['mean_ray_distance_m'] > bad_thr

df = df[['lamp_id', 'lamp_num', 'x_m_local', 'y_m_local', 'mean_ray_distance_m', 'z_score', 'rank_worst_first', 'quality_label', 'is_outlier_iqr']]
df.to_csv(csv_out, index=False, encoding='utf-8-sig')

summary = pd.DataFrame({
    'metric': ['count', 'mean', 'median', 'q1', 'q3', 'iqr', 'warning_threshold_q3', 'bad_threshold_q3_plus_1p5iqr'],
    'value': [
        len(df),
        df['mean_ray_distance_m'].mean(),
        df['mean_ray_distance_m'].median(),
        q1, q3, iqr, warn_thr, bad_thr
    ]
})
summary.to_csv(summary_out, index=False, encoding='utf-8-sig')

color_map = {'good': '#2E8B57', 'warning': '#E6A700', 'bad': '#C0392B'}
bar_colors = [color_map[q] for q in df.sort_values('mean_ray_distance_m', ascending=False)['quality_label']]
plot_df = df.sort_values('mean_ray_distance_m', ascending=False).copy()
plot_df['lamp_label'] = 'L' + plot_df['lamp_id']

fig = go.Figure()
fig.add_trace(go.Bar(
    x=plot_df['lamp_label'],
    y=plot_df['mean_ray_distance_m'],
    marker_color=bar_colors,
    customdata=plot_df[['quality_label', 'rank_worst_first', 'z_score']].values,
    hovertemplate='Lamp %{x}<br>Mean ray dist.: %{y:.3f} m<br>Quality: %{customdata[0]}<br>Worst-rank: %{customdata[1]}<br>Z-score: %{customdata[2]:.2f}<extra></extra>'
))

fig.add_hline(y=warn_thr, line_width=1.5, line_dash='dash', line_color='#B8860B')
fig.add_hline(y=bad_thr, line_width=1.5, line_dash='dot', line_color='#8B0000')

fig.update_layout(
    title='Lamp quality ranking by ray error (current set)<br><span style="font-size: 18px; font-weight: normal;">Thresholds from IQR rule on mean ray distance</span>'
)
fig.update_xaxes(title_text='Lamp ID')
fig.update_yaxes(title_text='Ray err (m)')
fig.write_image(str(chart_path))

with open(meta_path, 'w', encoding='utf-8') as f:
    json.dump({
        'caption': 'Lamp quality ranking from mean ray distance',
        'description': 'Bar chart ranking lamps by mean ray distance, with IQR-based warning and bad thresholds for quality assessment.'
    }, f, ensure_ascii=False)

print(str(chart_path))
print(str(csv_out))
print(str(summary_out))