#!/usr/bin/env python3
# Local seeding utilities following 12-factor config
from __future__ import annotations
import os, sys, hashlib, time
from typing import Iterable, Tuple, Callable, List
from pathlib import Path

# Lazy import psycopg2 only when needed
try:
    import psycopg2  # type: ignore
except Exception:  # pragma: no cover
    psycopg2 = None  # allow dry-run without dependency

DEFAULT_DB = os.getenv("PGDATABASE", os.getenv("DB_NAME", "brain_atomic"))
DEFAULT_USER = os.getenv("PGUSER", os.getenv("DB_USER", "postgres"))
# Prefer Brain volume sockets if present, then External, then home dir, else localhost
DEFAULT_HOST = (
    os.getenv("PGHOST")
    or os.getenv("DB_HOST")
    or next(
        (
            c
            for c in [
                "/Volumes/Brain/AI_Brain/pg",
                "/Volumes/Brain/HiveExpansion/pg",
                "/Volumes/External/AI_Brain/pg",
                str(Path.home() / "AI_Brain" / "pg"),
            ]
            if Path(c).exists()
        ),
        "localhost",
    )
)
DEFAULT_PORT = int(os.getenv("PGPORT", os.getenv("DB_PORT", "5432")))
DRY_RUN = os.getenv("DRY_RUN", "0").lower() in {"1", "true", "yes", "on"}

SOURCE_TAG = os.getenv("SEED_SOURCE", "Hive500Expansion")


def compute_hash(sentence: str) -> str:
    # Atomic uniqueness across identical sentences (global), as specified
    return hashlib.sha256(sentence.encode("utf-8")).hexdigest()


def connect_db():
    if DRY_RUN:
        return None
    if psycopg2 is None:
        sys.stderr.write(
            "psycopg2 not installed. Install with: pip install psycopg2-binary\n"
        )
        sys.exit(1)
    return psycopg2.connect(dbname=DEFAULT_DB, user=DEFAULT_USER, host=DEFAULT_HOST, port=DEFAULT_PORT)


INSERT_SQL = (
    "INSERT INTO atomic_facts (hash, sentence, category, source, created_at) "
    "VALUES (%s, %s, %s, %s, now()) ON CONFLICT (hash) DO NOTHING;"
)


def insert_many(category: str, items: Iterable[str]) -> Tuple[int, int]:
    """Insert items; returns (attempted, inserted_estimate)."""
    attempted = 0
    inserted = 0
    conn = connect_db()
    cur = conn.cursor() if conn is not None else None
    try:
        for s in items:
            sentence = s.strip()
            if not sentence:
                continue
            if not sentence.endswith("."):
                sentence = f"{sentence}."
            h = compute_hash(sentence)
            attempted += 1
            if DRY_RUN:
                inserted += 1
                continue
            cur.execute(INSERT_SQL, (h, sentence, category, SOURCE_TAG))
        if conn is not None:
            conn.commit()
    finally:
        if cur is not None:
            cur.close()
        if conn is not None:
            conn.close()
    return attempted, (inserted if DRY_RUN else attempted)


def read_file_lines(path: str, limit: int) -> List[str]:
    seen = set()
    out: List[str] = []
    with open(path, "r", encoding="utf-8") as f:
        for raw in f:
            line = raw.strip()
            if not line:
                continue
            if line in seen:
                continue
            seen.add(line)
            out.append(line)
            if len(out) >= limit:
                break
    return out


def run_seed(category: str, generator: Callable[[int], Iterable[str]], limit: int = 500) -> None:
    start = time.time()
    cat_upper = category.upper()
    file_override = os.getenv(f"SEED_FILE_{cat_upper}") or os.getenv("SEED_FILE")
    if file_override and os.path.isfile(file_override):
        items = read_file_lines(file_override, limit)
        mode_detail = f"FILE:{file_override}"
    else:
        items = list(generator(limit))
        mode_detail = "GENERATOR"
    attempted, inserted = insert_many(category, items)
    dur = time.time() - start
    mode = "DRY_RUN" if DRY_RUN else "LIVE"
    print(
        f"✅ {mode} seeded category='{category}' source={mode_detail} attempted={attempted} inserted~={inserted} in {dur:.2f}s"
    )