core/seed.py

"""
seed.py — Dados de demonstração para o mapa de Brasília, DF.
Chamado no startup do app.py. Garante que os 6 animais fixos estão sempre no banco.
"""
import json
import logging
import shutil
from datetime import datetime, timedelta
from pathlib import Path

import numpy as np

log = logging.getLogger(__name__)

_SEED_ANIMALS = [
    {
        "species": "dog",
        "name": "Pretão",
        "photo_filename": "pretao.jpg",
        "description": {
            "species": "dog",
            "breed_estimate": "SRD",
            "size": "médio",
            "primary_color": "preto",
            "secondary_colors": ["caramelo"],
            "distinctive_marks": ["focinho caramelo", "patas caramelo"],
            "condition": "saudável",
            "description_text": "medium black dog with caramel markings on face and paws, calm and friendly",
            "_ai_success": True,
        },
        "sightings": [
            {"lat": -15.7560, "lng": -47.8920, "days_ago": 10, "notes": "Deitado na calçada da 406 Norte, aceitou carinho"},
            {"lat": -15.7555, "lng": -47.8912, "days_ago": 3,  "notes": "Voltou pro mesmo ponto, parece ter território fixo"},
        ],
    },
    {
        "species": "dog",
        "name": "Mel",
        "photo_filename": "mel.webp",
        "description": {
            "species": "dog",
            "breed_estimate": "SRD",
            "size": "médio",
            "primary_color": "branco",
            "secondary_colors": ["caramelo"],
            "distinctive_marks": ["manchas caramelo nas costas", "orelhas caramelo"],
            "condition": "saudável",
            "description_text": "medium white dog with caramel patches, resting near closed shop",
            "_ai_success": True,
        },
        "sightings": [
            {"lat": -15.8450, "lng": -47.8700, "days_ago": 20, "notes": "Próximo à feira do Lago Sul, dócil"},
            {"lat": -15.8442, "lng": -47.8692, "days_ago": 8,  "notes": "Mesma área, aparenta bem alimentada"},
            {"lat": -15.8438, "lng": -47.8705, "days_ago": 1,  "notes": "Vista descansando na entrada de um comércio"},
        ],
    },
    {
        "species": "dog",
        "name": "Caramelo",
        "photo_filename": "caramelo.jpg",
        "description": {
            "species": "dog",
            "breed_estimate": "SRD",
            "size": "grande",
            "primary_color": "caramelo",
            "secondary_colors": [],
            "distinctive_marks": ["pelagem densa", "cauda enrolada"],
            "condition": "saudável",
            "description_text": "large golden caramel mixed breed dog, standing alert near road in Taguatinga",
            "_ai_success": True,
        },
        "sightings": [
            {"lat": -15.8300, "lng": -47.9900, "days_ago": 45, "notes": "Visto na marginal de Taguatinga, sozinho"},
            {"lat": -15.8312, "lng": -47.9888, "days_ago": 15, "notes": "Ainda na área, precisa de atenção"},
        ],
    },
    {
        "species": "dog",
        "name": "Manchinha",
        "photo_filename": "manchinha.jpg",
        "description": {
            "species": "dog",
            "breed_estimate": "SRD",
            "size": "médio",
            "primary_color": "branco",
            "secondary_colors": ["preto"],
            "distinctive_marks": ["manchas pretas na cabeça", "focinho malhado", "coleira verde"],
            "condition": "saudável",
            "description_text": "medium white dog with black patches on head, spotted muzzle, wearing green collar, very friendly",
            "_ai_success": True,
        },
        "sightings": [
            {"lat": -15.8010, "lng": -47.8960, "days_ago": 5, "notes": "Super dócil, tem coleira mas sem identificação — 308 Sul"},
            {"lat": -15.8005, "lng": -47.8950, "days_ago": 1, "notes": "Brincando com passantes perto do comércio"},
        ],
    },
    {
        "species": "cat",
        "name": "Cinzinha",
        "photo_filename": "cinzinha.jpg",
        "description": {
            "species": "cat",
            "breed_estimate": "SRD",
            "size": "pequeno",
            "primary_color": "cinza",
            "secondary_colors": [],
            "distinctive_marks": ["listras tigradas", "olhos amarelos grandes"],
            "condition": "saudável",
            "description_text": "small grey tabby kitten with large yellow eyes, timid but curious",
            "_ai_success": True,
        },
        "sightings": [
            {"lat": -15.7620, "lng": -47.8875, "days_ago": 2, "notes": "Filhote na W3 Norte, aceita comida com cuidado"},
        ],
    },
    {
        "species": "cat",
        "name": "Laranja",
        "photo_filename": "laranja.jpg",
        "description": {
            "species": "cat",
            "breed_estimate": "SRD",
            "size": "médio",
            "primary_color": "laranja",
            "secondary_colors": [],
            "distinctive_marks": ["pelagem totalmente ruiva", "parte de colônia"],
            "condition": "saudável",
            "description_text": "medium orange tabby cat, part of a feral colony near dumpsters in Ceilândia",
            "_ai_success": True,
        },
        "sightings": [
            {"lat": -15.8170, "lng": -48.1080, "days_ago": 7,  "notes": "Colônia de gatos ruivos perto da feira de Ceilândia"},
            {"lat": -15.8165, "lng": -48.1072, "days_ago": 0,  "notes": "Dois gatos laranja, parecem saudáveis"},
        ],
    },
]


def _random_embedding():
    v = np.random.randn(384).astype(np.float32)
    v /= np.linalg.norm(v)
    return v.tobytes()


def _copy_seed_photos():
    """Copia fotos de static/seed-photos/ (no repo) para DATA_DIR/photos/seed/ (runtime)."""
    from core.database import DATA_DIR, PHOTOS_DIR
    src_dir = Path(__file__).parent.parent / "static" / "seed-photos"
    dst_dir = PHOTOS_DIR / "seed"
    dst_dir.mkdir(parents=True, exist_ok=True)
    for src in src_dir.iterdir():
        dst = dst_dir / src.name
        if not dst.exists():
            shutil.copy2(src, dst)
            log.info("Foto copiada: %s", src.name)


def seed_if_empty(db):
    """Garante que os animais fixos estão sempre no banco.
    Insere apenas os que ainda nao existem (por nome).
    Dados reais dos usuarios nunca sao removidos.
    """
    with db._conn() as conn:
        existing_names = {
            row[0] for row in conn.execute(
                "SELECT name FROM animals WHERE name IS NOT NULL"
            )
        }

    # Sempre copia fotos (idempotente) — garante que estao disponiveis apos restart
    _copy_seed_photos()

    to_insert = [a for a in _SEED_ANIMALS if a["name"] not in existing_names]
    if not to_insert:
        log.info("Animais fixos ja presentes — seed ignorado.")
        return

    log.info("Inserindo %d animais fixos de demonstracao...", len(to_insert))
    now = datetime.utcnow()

    with db._conn() as conn:
        for animal_data in to_insert:
            sightings = animal_data["sightings"]
            desc = animal_data["description"]
            photo_path = "photos/seed/" + animal_data["photo_filename"]

            most_recent_days = min(s["days_ago"] for s in sightings)
            last_seen = now - timedelta(days=most_recent_days)
            first_seen = now - timedelta(days=max(s["days_ago"] for s in sightings))

            cur = conn.execute(
                "INSERT INTO animals"
                " (species, name, description, embedding, first_seen, last_seen, sighting_count)"
                " VALUES (?, ?, ?, ?, ?, ?, ?)",
                (
                    animal_data["species"],
                    animal_data["name"],
                    json.dumps(desc, ensure_ascii=False),
                    _random_embedding(),
                    first_seen.strftime("%Y-%m-%d %H:%M:%S"),
                    last_seen.strftime("%Y-%m-%d %H:%M:%S"),
                    len(sightings),
                ),
            )
            animal_id = cur.lastrowid

            for i, s in enumerate(sightings):
                created_at = now - timedelta(days=s["days_ago"])
                s_photo = photo_path if i == len(sightings) - 1 else None
                conn.execute(
                    "INSERT INTO sightings"
                    " (animal_id, photo_path, latitude, longitude, notes, created_at)"
                    " VALUES (?, ?, ?, ?, ?, ?)",
                    (
                        animal_id,
                        s_photo,
                        s["lat"],
                        s["lng"],
                        s.get("notes", ""),
                        created_at.strftime("%Y-%m-%d %H:%M:%S"),
                    ),
                )

    log.info("Seed concluido: %d animais inseridos (%d avistamentos).",
             len(to_insert),
             sum(len(a["sightings"]) for a in to_insert))