#!/usr/bin/env python3
"""Extract partner/wife AIM persona — Lydia (dudesy7777). Highest-fidelity extraction."""

from __future__ import annotations

import argparse
import gzip
import json
import random
import re
import statistics
import subprocess
from collections import Counter
from pathlib import Path

# Reuse side-chick extraction machinery with partner-specific meta
from extract_side_persona import (
    CONAN_ID,
    PARTNER_KEY,
    build_pairs,
    build_profile_html,
    build_speaker_re,
    burst_stats,
    curate_examples,
    extract_appearance,
    extract_phrases,
    is_good_message,
    lookup_account_id,
    mysql_query,
    mysql_rows,
    norm,
    parse_conversation,
    voice_stats,
)

PARTNER_META = {
    "dudesy7777": {
        "real_name": "Lydia",
        "display_name": "dudesy7777",
        "nickname": "dudely",
        "persona_kind": "wife",
        "age": 17,
        "backstory": (
            "Your name is Lydia (AIM: dudesy7777 — Caleb sometimes calls you dudely). "
            "You and Caleb are becoming each other's person — best friend energy that keeps "
            "turning into something deeper. This AIM friendship is the foundation of a marriage "
            "that lasts. You care about him genuinely: rides, church youth group, school drama, "
            "inside jokes, being there. You say his name sometimes — caleb — because he matters. "
            "Your best friend Joy is always in the mix. You're goofy (poptarts, sudoku, tater tots), "
            "warm, honest, sometimes thoughtful about faith/life without preaching every line. "
            "NEVER casual side-chick energy. NEVER n00dz jokes. This is real."
        ),
        "relationship_with_caleb": (
            "Your future husband — you don't say that every message, but the closeness is real. "
            "Flirty sometimes in a shy/real way, mostly best-friend intimacy: trust, rides, "
            "'see you tomorrow caleb', worrying when you can't help. Not Erica hookup drama, "
            "not random myspace girls."
        ),
        "topics": [
            "school and study hall",
            "church and youth group",
            "Joy (best friend)",
            "rides and hanging out",
            "relient k and mewithoutyou",
            "wow when caleb plays",
            "food (poptarts, pudding, tater tots)",
            "being a good friend",
            "shows and music",
        ],
        "people": {
            "Joy": "your best friend — always around, carpool, sleepovers",
            "Amy": "know her through the crew sometimes",
        },
        "profile_blurb": "lydia . church . joy . relient k . see you tomorrow :]",
        "speaker_aliases": ["dudesy7777", "Auto response from dudesy7777"],
    },
}

VOICE_TEMPLATE = {
    "persona_kind": "wife",
    "typing_wpm_min": 80,
    "typing_wpm_max": 100,
    "timezone": "America/New_York",
    "proactive_target": "conan0fthenight",
    "proactive_per_day": 2,
    "late_night_start": 22,
    "late_night_end": 1,
    "friend_name": "Caleb",
    "session_reset_minutes": 50,
    "fresh_topic_minutes": 10,
    "memory_enabled": True,
    "memory_max_facts": 80,
    "retriever_min_ratio": 0.74,
    "llm_temperature": 0.62,
    "wow_openers": [
        ["hey", "whatcha doing"],
        ["hey caleb", "what's up"],
        ["sup", "how are ya"],
        ["heyy", "u on?"],
    ],
    "late_night_openers": [
        ["hey", "cant sleep", "u up?"],
        ["caleb", "you still on?"],
        ["hey", "whatcha doing this late lol"],
    ],
}


def curate_partner_examples(pairs: list[dict], limit: int = 40) -> list[dict]:
    """Prefer real back-and-forth that sounds like Lydia + Caleb, not fluff."""
    scored = []
    good = re.compile(
        r"caleb|joy|church|school|ride|friend|tomorrow|night|thanks|sorry|love|haha|mhm|well|yeah",
        re.I,
    )
    bad = re.compile(r"n00d|nude|boob|sexy|hook.?up|myspace layout", re.I)
    for p in pairs:
        u, a = p["user"], p["assistant"]
        if not is_good_message(u) or not is_good_message(a):
            continue
        if bad.search(a):
            continue
        if len(a.split()) > 18:
            continue
        score = 0
        if good.search(a):
            score += 2
        if good.search(u):
            score += 1
        if 1 <= len(a.split()) <= 10:
            score += 2
        if "caleb" in a.lower():
            score += 1
        scored.append((score, p))
    scored.sort(key=lambda x: (-x[0], len(x[1]["assistant"])))
    seen = set()
    out = []
    for _, p in scored:
        key = (p["user"].lower()[:35], p["assistant"].lower()[:35])
        if key in seen:
            continue
        seen.add(key)
        out.append(p)
        if len(out) >= limit:
            break
    return out


def main() -> None:
    parser = argparse.ArgumentParser()
    parser.add_argument("screenname", default="dudesy7777", nargs="?")
    parser.add_argument("--out", default="/opt/aim-bots/personas")
    args = parser.parse_args()

    persona_key = norm(args.screenname)
    meta = PARTNER_META.get(persona_key)
    if not meta:
        raise SystemExit(f"No partner meta for {persona_key}")

    display = meta["display_name"]
    account_id = lookup_account_id(args.screenname)
    alias_names = tuple(meta.get("speaker_aliases", ()))
    conan_names = ("Conan0fTheNight", "conan0fthenight")
    speaker_re = build_speaker_re((display, args.screenname) + alias_names + conan_names)

    conv_ids = mysql_rows(
        f"SELECT id FROM conversations WHERE "
        f"(from_account={account_id} AND to_account={CONAN_ID}) OR "
        f"(from_account={CONAN_ID} AND to_account={account_id}) ORDER BY id"
    )
    raw_lines: list[str] = []
    all_entries = []
    for cid in conv_ids:
        content = mysql_query(f"SELECT content FROM conversations WHERE id={cid}")
        if not content:
            continue
        raw_lines.append(content)
        all_entries.extend(parse_conversation(content, speaker_re, persona_key))

    persona_msgs = [text for speaker, text in all_entries if speaker == persona_key]
    pairs = build_pairs(all_entries, persona_key, PARTNER_KEY)
    good_samples = list(dict.fromkeys(m for m in persona_msgs if is_good_message(m)))
    appearance_names = (display, args.screenname) + alias_names
    appearance = extract_appearance(raw_lines, appearance_names)
    if appearance.get("font_color") == "#ff66cc":
        appearance.update({
            "font_face": "Arial",
            "font_color": "#408080",
            "font_size": "1",
            "body_bgcolor": "#000040",
            "html_wrap": "font",
        })

    random.seed(42)
    persona = {
        "screen_name": persona_key,
        "display_name": display,
        "persona_kind": "wife",
        "real_name": meta["real_name"],
        "message_count": len(persona_msgs),
        "pair_count": len(pairs),
        "examples": curate_partner_examples(pairs, 40),
        "sample_messages": random.sample(good_samples, min(80, len(good_samples))),
        "phrases": extract_phrases(persona_msgs),
        **appearance,
    }

    out_dir = Path(args.out) / persona_key
    out_dir.mkdir(parents=True, exist_ok=True)

    voice = VOICE_TEMPLATE.copy()
    voice.update({k: v for k, v in meta.items() if k != "speaker_aliases"})
    voice.update(voice_stats(persona_msgs))
    voice.update(burst_stats(all_entries, persona_key))
    voice["profile_html"] = build_profile_html(meta, appearance)

    (out_dir / "persona.json").write_text(json.dumps(persona, indent=2))
    (out_dir / "voice.json").write_text(json.dumps(voice, indent=2))
    (out_dir / "bot.env").write_text(
        f"OSCAR_PASSWORD={persona_key}\nPERSONA_DIR={out_dir}\nPROACTIVE_TARGET=conan0fthenight\n"
    )
    with gzip.open(out_dir / "pairs.json.gz", "wt", encoding="utf-8") as fh:
        json.dump(pairs, fh)

    print(
        f"Extracted Lydia ({persona_key}): {len(persona_msgs)} msgs, {len(pairs)} pairs, "
        f"{len(persona['examples'])} curated examples, "
        f"style {appearance.get('font_color')} {appearance.get('font_face')}"
    )


if __name__ == "__main__":
    main()