import json
import pandas as pd
from pathlib import Path

# Paths
base_dir = Path("src/esdac")
output_dir = Path("src/fuse_esdac/outputs")
output_dir.mkdir(parents=True, exist_ok=True)

# Read JSON file
with open(base_dir / "status.json", "r", encoding="utf-8") as f:
    data = json.load(f)

# Keep only items with status == "PROCESSED"
processed_items = [item for item in data if item.get("status") == "PROCESSED"]

# Convert to DataFrame
df = pd.DataFrame(processed_items)

# Rename url → dataset_url
df.rename(columns={"url": "dataset_url"}, inplace=True)

# Drop unwanted columns
cols_to_drop = [
    "request_needed", "status", "notes",
    "screened_by", "requested_downloaded_by", "processed_by"
]
df.drop(columns=cols_to_drop, inplace=True, errors="ignore")

# Determine fuse status and reason
fuse_statuses = []
reasons = []

for _, row in df.iterrows():
    dataset_dir = base_dir / row["name"]
    fuse_schema = dataset_dir / "fuse_schema.json"
    fuse_skip = dataset_dir / "fuse_skip.txt"

    if fuse_schema.exists():
        fuse_statuses.append("FUSED")
        reasons.append("")
    elif fuse_skip.exists():
        fuse_statuses.append("SKIPPED")
        try:
            text = fuse_skip.read_text(encoding="utf-8").strip()
        except Exception as e:
            text = f"[Error reading fuse_skip.txt: {e}]"
        reasons.append(text)
    else:
        fuse_statuses.append("AWAIT")
        reasons.append("")

df["fuse_status"] = fuse_statuses
df["reason"] = reasons

# Save to CSV
output_path = output_dir / "processed_datasets.csv"
df.to_csv(output_path, index=False, encoding="utf-8")

print(f"✅ Saved {len(df)} processed datasets to {output_path}")

# ---- existing code above stays unchanged ----

# Save to CSV (all processed)
output_path = output_dir / "processed_datasets.csv"
df.to_csv(output_path, index=False, encoding="utf-8")
print(f"✅ Saved {len(df)} processed datasets to {output_path}")

# ---- new feature: save only fused datasets ----
df_fused = df[df["fuse_status"] == "FUSED"].drop(columns=["fuse_status", "reason"])
fused_output_path = output_dir / "fused_datasets.csv"
df_fused.to_csv(fused_output_path, index=False, encoding="utf-8")
print(f"✅ Saved {len(df_fused)} fused datasets to {fused_output_path}")