import json
from pathlib import Path

path = Path("/mnt/storage/document-processor/exports/document_training.jsonl")

count = 0
approved = 0

with path.open() as f:
    for line in f:
        row = json.loads(line)
        count += 1
        if row["review"]["is_approved"]:
            approved += 1

        if count <= 3:
            print("\n--- SAMPLE ---")
            print("ID:", row["document"]["document_id"])
            print("Merchant:", row["extracted_fields"].get("merchant_normalized"))
            print("Total:", row["extracted_fields"].get("total"))
            print("OCR len:", len(row["ocr_text"]))

print("\nTotal docs:", count)
print("Approved:", approved)