#!/usr/bin/env python3
"""Validate the HF dataset staging files for BlackLansky/cwi-catalog.

Checks only what can be honestly checked: valid JSON/JSONL syntax,
non-empty required fields, uniqueness of identifiers, provenance
traceability, and format invariants (e.g. the Needle Drop ledger must
ship with verified_placements empty — entries only on verification).
Nothing is assumed or inferred beyond the source files.
"""
import json
import os
import sys

BASE = os.path.dirname(os.path.abspath(__file__))
errors = []


def load_jsonl(path):
    rows = []
    with open(path, encoding="utf-8") as f:
        for i, line in enumerate(f, 1):
            line = line.strip()
            if not line:
                continue
            try:
                rows.append(json.loads(line))
            except json.JSONDecodeError as e:
                errors.append(f"{os.path.basename(path)} line {i}: invalid JSON ({e})")
    return rows


def load_json(path):
    try:
        with open(path, encoding="utf-8") as f:
            return json.load(f)
    except json.JSONDecodeError as e:
        errors.append(f"{os.path.basename(path)}: invalid JSON ({e})")
        return None


def require(rows, fname, fields):
    for i, r in enumerate(rows, 1):
        for field in fields:
            v = r.get(field)
            if v is None or (isinstance(v, str) and not v.strip()) or (isinstance(v, (list, dict)) and len(v) == 0):
                errors.append(f"{fname} line {i}: empty required field '{field}'")


def check_provenance(rows, fname, marker):
    for i, r in enumerate(rows, 1):
        src = r.get("dataset_provenance", {}).get("source", "")
        if marker not in src:
            errors.append(f"{fname} line {i}: provenance missing (expected '{marker}')")


# ---- catalog.jsonl (verified track metadata, 36 tracks as of 2026-10-06) ----
catalog = load_jsonl(os.path.join(BASE, "catalog.jsonl"))
# spotify_url is optional: pre-release / non-Spotify tracks carry null (validated below).
require(catalog, "catalog.jsonl", ["title", "artist", "label", "dataset_provenance"])
check_provenance(catalog, "catalog.jsonl", "cwi-learn/catalog.json")
titles = [r["title"] for r in catalog]
if len(titles) != len(set(titles)):
    errors.append("catalog.jsonl: duplicate track titles")
if {r["artist"] for r in catalog} != {"That Boy Hi Hat", "Dre50", "183 Wildboi", "King Akeem"}:
    errors.append("catalog.jsonl: unexpected artist values")
for i, r in enumerate(catalog, 1):
    # Pre-release / non-Spotify tracks legitimately carry null spotify_url
    # (e.g. Cyberpunk 2027, Apple Music only until release). When present it
    # must be a canonical Spotify track URL.
    su = r.get("spotify_url")
    if su is not None and not su.startswith("https://open.spotify.com/track/"):
        errors.append(f"catalog.jsonl line {i}: bad spotify_url")

# ---- products.jsonl (Agent Deck registry, 26 SKUs) ----
products = load_jsonl(os.path.join(BASE, "products.jsonl"))
require(products, "products.jsonl", ["sku_id", "name", "department", "description", "schema_url", "dataset_provenance"])
for i, r in enumerate(products, 1):
    src = r.get("dataset_provenance", {}).get("source", "")
    if "cwi-learn/datasets/agent-deck-skus.jsonl" not in src and "playback-chip" not in src:
        errors.append(f"products.jsonl line {i}: provenance missing (expected registry or playback-chip source)")
sku_ids = [r["sku_id"] for r in products]
if len(sku_ids) != len(set(sku_ids)):
    errors.append("products.jsonl: duplicate sku_id values")
if len(products) != 26:
    errors.append(f"products.jsonl: expected 26 SKUs, found {len(products)}")

# ---- tracks.jsonl (canonical training dataset, 36 rows as of 2026-10-06) ----
tracks = load_jsonl(os.path.join(BASE, "tracks.jsonl"))
require(tracks, "tracks.jsonl", ["title", "artist"])
if len(tracks) != 36:
    errors.append(f"tracks.jsonl: expected 36 rows, found {len(tracks)}")
if {r.get("artist") for r in tracks} != {"That Boy Hi Hat", "Dre50", "183 Wildboi", "King Akeem"}:
    errors.append("tracks.jsonl: unexpected artist values")

# ---- placements.jsonl (Needle Drop ledger — ships honestly empty) ----
placements = load_jsonl(os.path.join(BASE, "placements.jsonl"))
for i, r in enumerate(placements, 1):
    if r.get("verified_placements") != []:
        errors.append(f"placements.jsonl line {i}: verified_placements must stay empty until verified")

# ---- curators.jsonl (curator verification records) ----
curators = load_jsonl(os.path.join(BASE, "curators.jsonl"))
require(curators, "curators.jsonl", ["curator", "status", "as_of"])
valid_statuses = {"verified", "unverified", "unconfirmed"}
for i, r in enumerate(curators, 1):
    if r.get("status") not in valid_statuses:
        errors.append(f"curators.jsonl line {i}: unknown status '{r.get('status')}'")

# ---- press.jsonl (real press records) ----
press = load_jsonl(os.path.join(BASE, "press.jsonl"))
require(press, "press.jsonl", ["title", "outlet", "url"])

# ---- training.json (corpus manifest) ----
training = load_json(os.path.join(BASE, "training.json"))
if training is not None and "datasets" not in training:
    errors.append("training.json: missing 'datasets' key")

print(f"catalog.jsonl: {len(catalog)} rows")
print(f"products.jsonl: {len(products)} rows")
print(f"tracks.jsonl: {len(tracks)} rows")
print(f"placements.jsonl: {len(placements)} rows (verified_placements empty: "
      f"{all(r.get('verified_placements') == [] for r in placements)})")
print(f"curators.jsonl: {len(curators)} rows")
print(f"press.jsonl: {len(press)} rows")
print(f"training.json: {'valid' if training is not None else 'INVALID'}")

if errors:
    print("\nFAILURES:")
    for e in errors:
        print(" -", e)
    sys.exit(1)
print("\nVALIDATE OK — all files valid; required fields present; provenance traceable; ledger honestly empty.")