import pandas as pd
import warnings
warnings.filterwarnings('ignore')

csv_path = "/data1/nasiruddink/rockglacier/RockGlacierNet/patch_dataset/patches_128/metadata/patch_metadata.csv"
df = pd.read_csv(csv_path)

# Filter for only the patches that are currently kept in the dataset
kept = df[df["kept_or_filtered"] == "kept"].copy()

# Define finer bins for analysis
bins = [-1, 0.0, 0.1, 1.0, 2.0, 5.0, 10.0, 100.0]
labels = [
    "0% (Pure Background)", 
    "0.01% - 0.1% (Trace)", 
    "0.1% - 1.0%", 
    "1.0% - 2.0%", 
    "2.0% - 5.0%", 
    "5.0% - 10.0%", 
    "> 10.0% (Good Coverage)"
]

kept['coverage_bin'] = pd.cut(kept['glacier_percentage'], bins=bins, labels=labels)
counts = kept['coverage_bin'].value_counts(sort=False)
percentages = (counts / len(kept)) * 100

print("\nROCK GLACIER COVERAGE ANALYSIS OF KEPT PATCHES:")
print("=" * 65)
print(f"{'Coverage Range':<25} | {'Count':<10} | {'Percentage of Dataset':<20}")
print("-" * 65)
for label in labels:
    print(f"{label:<25} | {counts[label]:<10} | {percentages[label]:.1f}%")
print("=" * 65)
print(f"Total Kept Patches: {len(kept)}\n")

# Suggest what happens if we set a threshold
threshold = 1.0
retained = len(kept[kept['glacier_percentage'] >= threshold]) + len(kept[kept['glacier_percentage'] == 0.0])
print(f"If you set MIN_GLACIER_PERCENTAGE = {threshold}% in config.py:")
print(f"- You would drop {(len(kept) - retained)} patches with trace/negligible glacier pixels.")
print(f"- The new dataset size would be {retained} patches.")