src/catalog.py

"""Catalog module: JSON-based wardrobe storage with search.

The catalog is a simple JSON file that stores garment entries. Each
garment has a unique ID and structured attributes extracted by the
vision pipeline.

No vector DB, no embeddings. In-memory filtering over a list of dicts.
A wardrobe will never exceed a few hundred items.
"""

import json
import logging
from pathlib import Path

from .storage import save_image, get_image_path, delete_image

logger = logging.getLogger(__name__)

CATALOG_PATH = Path(__file__).resolve().parent.parent / "data" / "catalog.json"


def load_catalog() -> list[dict]:
    """Load the catalog from disk. Returns empty list if file doesn't exist."""
    if not CATALOG_PATH.exists():
        logger.debug("Catalog file not found: %s", CATALOG_PATH)
        return []
    try:
        with open(CATALOG_PATH, "r", encoding="utf-8") as f:
            catalog = json.load(f)
        logger.debug("Loaded catalog: %d garments from %s", len(catalog), CATALOG_PATH)
        return catalog
    except (json.JSONDecodeError, IOError) as e:
        logger.error("Failed to load catalog: %s", e)
        return []


def save_catalog(garments: list[dict]) -> None:
    """Write the full catalog to disk."""
    CATALOG_PATH.parent.mkdir(parents=True, exist_ok=True)
    with open(CATALOG_PATH, "w", encoding="utf-8") as f:
        json.dump(garments, f, indent=2, ensure_ascii=False)
    logger.info("Catalog saved: %d garments", len(garments))


def _next_id(catalog: list[dict]) -> int:
    """Get the next available garment ID number."""
    if not catalog:
        return 1
    existing_ids = []
    for g in catalog:
        gid = g.get("id", "")
        if isinstance(gid, str) and gid.startswith("garment_"):
            try:
                existing_ids.append(int(gid.split("_")[1]))
            except (ValueError, IndexError):
                pass
    return max(existing_ids, default=0) + 1


def add_garments(
    garments_with_images: list[tuple[dict, bytes | None]],
) -> list[dict]:
    """Add new garments to the catalog. Assigns IDs and persists images.

    Accepts a list of (garment_dict, image_bytes) tuples. Each garment
    gets its own individual image saved.

    Returns the garments with their assigned IDs.
    """
    catalog = load_catalog()
    next_num = _next_id(catalog)
    added = []

    for i, (garment, img_bytes) in enumerate(garments_with_images):
        garment_id = f"garment_{next_num + i:03d}"
        garment["id"] = garment_id

        if img_bytes:
            filename = save_image(garment_id, img_bytes)
            garment["image_ref"] = filename

        catalog.append(garment)
        added.append(garment)

    save_catalog(catalog)
    return added


def search(
    garment_type: str | None = None,
    color: str | None = None,
    season: str | None = None,
    formality: str | None = None,
) -> list[dict]:
    """Filter catalog by attributes. All filters are AND-combined."""
    catalog = load_catalog()
    results = catalog

    if garment_type:
        garment_type = garment_type.lower().strip()
        results = [g for g in results if garment_type in g.get("type", "")]

    if color:
        color = color.lower().strip()
        results = [g for g in results if color in g.get("color", "")]

    if season:
        season = season.lower().strip()
        results = [g for g in results if g.get("season", "") in (season, "all")]

    if formality:
        formality = formality.lower().strip()
        results = [g for g in results if formality in g.get("formality", "")]

    return results


def get_catalog_summary() -> str:
    """Format the catalog as a text block for LLM context injection."""
    catalog = load_catalog()
    if not catalog:
        return "The wardrobe is empty. No garments have been cataloged yet."

    lines = [f"Wardrobe catalog ({len(catalog)} garments):\n"]
    for g in catalog:
        gid = g.get("id", "?")
        gtype = g.get("type", "?")
        color = g.get("color", "?")
        material = g.get("material", "?")
        pattern = g.get("pattern", "?")
        season = g.get("season", "?")
        formality = g.get("formality", "?")
        description = g.get("description", "")

        line = (
            f"- [{gid}] {color} {gtype} ({material}, {pattern}) "
            f"| season: {season} | style: {formality}"
        )
        if description:
            line += f"\n  → {description}"
        lines.append(line)

    return "\n".join(lines)


def get_catalog_stats() -> dict:
    """Return aggregate statistics about the catalog."""
    catalog = load_catalog()
    if not catalog:
        return {"total": 0}

    types: dict[str, int] = {}
    colors: dict[str, int] = {}
    seasons: dict[str, int] = {}

    for g in catalog:
        t = g.get("type", "unknown")
        types[t] = types.get(t, 0) + 1
        c = g.get("color", "unknown")
        colors[c] = colors.get(c, 0) + 1
        s = g.get("season", "unknown")
        seasons[s] = seasons.get(s, 0) + 1

    return {
        "total": len(catalog),
        "by_type": dict(sorted(types.items(), key=lambda x: -x[1])),
        "by_color": dict(sorted(colors.items(), key=lambda x: -x[1])),
        "by_season": dict(sorted(seasons.items(), key=lambda x: -x[1])),
    }


def get_garment_image_path(garment_id: str) -> str | None:
    """Resolve the local filesystem path for a garment's image."""
    return get_image_path(garment_id)


def clear_catalog() -> None:
    """Remove all garments from the catalog and their images."""
    catalog = load_catalog()
    for g in catalog:
        gid = g.get("id", "")
        if gid:
            delete_image(gid)
    save_catalog([])
    logger.info("Catalog cleared")