"""Export sink artifacts for the `wai.neural.elic` checkerboard context codec
decoded by `wai-rs/src/neural/context.rs`.

context.rs needs (next to the registered model path):
  hyper_synthesis.onnx   z_hat        -> [1, 2*Cy, Sy, Sy]  (scales ‖ means, base)
  context.onnx           y_anchors    -> [1, 2*Cy, Sy, Sy]  (refined params for non-anchors)
  synthesis.onnx         y_hat        -> [1, 3, H, W]
  hyper_cdfs_q<q>.json   factorized z CDF (EntropyBottleneck)
  gauss_cdfs_q<q>.json   { scale_table, cdf[], cdf_length[], offset[] } (GaussianConditional)

It wraps an ELIC-style checkpoint exposing submodules `h_s` (hyper synthesis),
`context` (the checkerboard context net: anchor-filled latent → params), `g_s`
(synthesis), `entropy_bottleneck`, `gaussian_conditional`. CDF extraction is
identical to the hyperprior tool (the entropy coders are the same).

Run (needs torch + onnxruntime; bring your own ELIC checkpoint):
  python tools/wai_elic_export.py --ckpt elic.pt --quality 3 --size 256 \
    --out-dir wai-web/demo/models/elic
"""
from __future__ import annotations

import argparse
import json
from pathlib import Path

import torch

# Reuse the CDF extractors from the hyperprior tool (same entropy coders).
from wai_mbt2018_mean_export import factorized_cdf_blob, gaussian_cdf_blob  # noqa: E402


def export_onnx(module, dummy, path: Path, in_name, out_name, opset):
    torch.onnx.export(
        module, (dummy,), path.as_posix(),
        input_names=[in_name], output_names=[out_name],
        dynamic_axes={in_name: {2: "h", 3: "w"}, out_name: {2: "h", 3: "w"}},
        opset_version=opset, do_constant_folding=True, dynamo=False)


def main() -> int:
    ap = argparse.ArgumentParser()
    ap.add_argument("--ckpt", required=True, help="ELIC checkpoint exposing h_s/context/g_s + entropy coders")
    ap.add_argument("--quality", type=int, default=3)
    ap.add_argument("--size", type=int, default=256)
    ap.add_argument("--out-dir", required=True)
    ap.add_argument("--opset", type=int, default=17)
    args = ap.parse_args()
    out = Path(args.out_dir); out.mkdir(parents=True, exist_ok=True)
    q = args.quality

    net = torch.load(args.ckpt, map_location="cpu", weights_only=False).eval()
    net.update(force=True)

    with torch.no_grad():
        y = net.g_a(torch.zeros(1, 3, args.size, args.size))
        z = net.h_a(y)
    _, M, Sy, _ = y.shape
    _, N, Sz, _ = z.shape
    print(f"  y[1,{M},{Sy},{Sy}]  z[1,{N},{Sz},{Sz}]")

    export_onnx(net.h_s, torch.zeros(1, N, Sz, Sz), out / "hyper_synthesis.onnx", "z", "params", args.opset)
    export_onnx(net.context, torch.zeros(1, M, Sy, Sy), out / "context.onnx", "y", "params", args.opset)
    export_onnx(net.g_s, torch.zeros(1, M, Sy, Sy), out / "synthesis.onnx", "y", "image", args.opset)

    (out / f"hyper_cdfs_q{q}.json").write_text(
        json.dumps(factorized_cdf_blob(net.entropy_bottleneck, q), separators=(",", ":")))
    (out / f"gauss_cdfs_q{q}.json").write_text(
        json.dumps(gaussian_cdf_blob(net.gaussian_conditional, q), separators=(",", ":")))
    print(f"wrote 3 ONNX + 2 CDF tables → {out}")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())