#!/usr/bin/env python3
"""Load blog-training-alpaca.jsonl into PostgreSQL learning_corpus table via psql."""

import json
import subprocess
import sys
from pathlib import Path

JSONL_FILE = Path(__file__).parent.parent / "data" / "blog-training-alpaca.jsonl"
DB_URL = "postgresql://llm:llm_secure_2026@127.0.0.1:15432/llm_gateway"
TASK_TYPE = "tip_blog"

# Parse connection details
parts = DB_URL.replace("postgresql://", "").split("@")
creds = parts[0].split(":")
db_user, db_pass = creds[0], creds[1]
host_port = parts[1].split("/")
host, port = host_port[0].split(":")
db_name = host_port[1]

print("🔄 Loading blog training data into PostgreSQL...")
print(f"   File: {JSONL_FILE}")
print(f"   Host: {host}:{port}")
print(f"   Database: {db_name}")
print()

# Generate SQL statements
sql_statements = []

# Read JSONL and convert to SQL
try:
    with open(JSONL_FILE) as f:
        samples = [json.loads(line.strip()) for line in f if line.strip()]
except FileNotFoundError:
    print(f"❌ File not found: {JSONL_FILE}")
    sys.exit(1)

# Start transaction
sql_statements.append("BEGIN;")
sql_statements.append(f"SELECT COUNT(*) as before_count FROM learning_corpus WHERE task_type = '{TASK_TYPE}';")

# Insert samples
for sample in samples:
    prompt_text = sample.get("instruction", "")
    if sample.get("input"):
        prompt_text += f"\n{sample['input']}"

    # Escape single quotes for SQL
    prompt_text = prompt_text.replace("'", "''")
    completion_text = sample.get("output", "").replace("'", "''")
    quality_score = sample.get("quality_score", 8.0)
    source = sample.get("source", "unknown").replace("'", "''")

    tags = f"ARRAY['{source}', '{TASK_TYPE}']"

    sql = f"INSERT INTO learning_corpus (task_type, prompt_text, completion_text, quality_score, tags) VALUES ('{TASK_TYPE}', '{prompt_text}', '{completion_text}', {quality_score}, {tags});"
    sql_statements.append(sql)

sql_statements.append(f"SELECT COUNT(*) as after_count FROM learning_corpus WHERE task_type = '{TASK_TYPE}';")
sql_statements.append("COMMIT;")

# Write to temporary SQL file
tmpfile = Path("/tmp/load_blog_data.sql")
with open(tmpfile, "w") as f:
    f.write("\n".join(sql_statements))

# Execute via psql
try:
    env = {"PGPASSWORD": db_pass}
    result = subprocess.run(
        [
            "/opt/homebrew/bin/psql",
            "-h", host,
            "-p", port,
            "-U", db_user,
            "-d", db_name,
            "-f", str(tmpfile),
        ],
        capture_output=True,
        text=True,
        env=env,
    )

    if result.returncode == 0:
        print(result.stdout)
        print(f"\n✅ Loaded {len(samples)} samples into learning_corpus (task_type={TASK_TYPE})")
        print("🎯 Next: python3 scripts/manual_trigger.py --general --force")
    else:
        print("❌ Error loading data:")
        print(result.stderr)
        sys.exit(1)
finally:
    tmpfile.unlink(missing_ok=True)