pak-jabit/catalogue/photoroom_to_spree.py

#!/usr/bin/env python3
"""
Pak Jabit — Photoroom return leg.

The missing half of photoroom_pipeline.py: takes a source image, runs it through
Photoroom, and pushes the result back onto the Spree product as a media asset.

  source jpg -> Photoroom /v2/edit -> Spree POST /products/:id/media -> verify

Verification is not optional here. Spree's media#create has a url path that
enqueues a background job and answers `head :accepted` — a 202 with no body and
no way to know if the image ever landed. This uses the multipart path, which
returns 201 plus the serialized asset, and then re-reads the product's media to
confirm the asset id is really there.

Usage:
  photoroom_to_spree.py --slug 19-electro-denim-lab-chino-pants --live
  photoroom_to_spree.py --slug <slug> --position 1        # make it the thumbnail
  photoroom_to_spree.py --slug <slug> --dry-run

Keys come from /root/.photoroom/credentials. Sandbox is the default and
watermarks every image; --live uses the paid allowance (10/month) and is the
only way to get a clean result onto a live storefront.
"""
import argparse
import json
import mimetypes
import pathlib
import sys
import urllib.error
import urllib.request
import uuid

import os as _os


def _spree_admin_key() -> str:
    """The Spree admin API key. Required — never defaulted."""
    key = _os.environ.get("SPREE_ADMIN_KEY", "").strip()
    if not key:
        raise SystemExit(
            "SPREE_ADMIN_KEY is not set.\n"
            "  export SPREE_ADMIN_KEY=<the admin API key>\n"
            "It is deliberately not stored in this file — see README.md."
        )
    return key


SPREE_API = "https://api.jux-lab.com/api/v3/admin"
SPREE_KEY = _spree_admin_key()
PHOTOROOM_EDIT = "https://image-api.photoroom.com/v2/edit"
CREDS = pathlib.Path("/root/.photoroom/credentials")
SOURCE_ROOT = pathlib.Path(
    "/root/Opus/Projects/PeakyBuildr/scraped-catalogue/products"
)


def fatal(msg):
    print(f"FATAL: {msg}", file=sys.stderr)
    sys.exit(1)


def load_key(live: bool) -> str:
    if not CREDS.exists():
        fatal(f"{CREDS} missing — no API key to use")
    want = "PHOTOROOM_LIVE_KEY" if live else "PHOTOROOM_SANDBOX_KEY"
    for line in CREDS.read_text().splitlines():
        if line.startswith(want + "="):
            return line.split("=", 1)[1].strip()
    fatal(f"{want} not found in {CREDS}")


def multipart(fields: dict, file_field: str, filename: str, blob: bytes, content_type: str):
    boundary = uuid.uuid4().hex
    parts = []
    for k, v in fields.items():
        parts.append(
            f'--{boundary}\r\nContent-Disposition: form-data; name="{k}"\r\n\r\n{v}\r\n'.encode()
        )
    parts.append(
        f'--{boundary}\r\nContent-Disposition: form-data; name="{file_field}"; '
        f'filename="{filename}"\r\nContent-Type: {content_type}\r\n\r\n'.encode()
        + blob
        + b"\r\n"
    )
    parts.append(f"--{boundary}--\r\n".encode())
    return b"".join(parts), f"multipart/form-data; boundary={boundary}"


def spree_get(path):
    req = urllib.request.Request(
        f"{SPREE_API}{path}", headers={"X-Spree-Api-Key": SPREE_KEY}
    )
    with urllib.request.urlopen(req, timeout=60) as r:
        return json.loads(r.read())


def find_product(slug):
    data = spree_get(f"/products?filter[slug_eq]={slug}&per_page=50")
    for p in data.get("data", []):
        if p.get("slug") == slug or p.get("attributes", {}).get("slug") == slug:
            return p
    # The filter key varies by Spree build; fall back to a full scan rather than
    # reporting "not found" for a product that is actually there.
    page = 1
    while True:
        d = spree_get(f"/products?per_page=100&page={page}")
        for p in d.get("data", []):
            if p.get("slug") == slug:
                return p
        if page >= d.get("meta", {}).get("pages", 1):
            return None
        page += 1


def photoroom(img_path: pathlib.Path, key: str, size: str, bg: str) -> bytes:
    body, ctype = multipart(
        {"background.color": bg, "outputSize": size, "padding": "0.05"},
        "imageFile",
        img_path.name,
        img_path.read_bytes(),
        mimetypes.guess_type(img_path.name)[0] or "image/jpeg",
    )
    req = urllib.request.Request(
        PHOTOROOM_EDIT, data=body, headers={"x-api-key": key, "Content-Type": ctype}
    )
    try:
        with urllib.request.urlopen(req, timeout=180) as r:
            data = r.read()
            if not data:
                fatal("Photoroom returned 200 with an empty body")
            if not r.headers.get("content-type", "").startswith("image/"):
                fatal(f"Photoroom returned non-image: {r.headers.get('content-type')}")
            return data
    except urllib.error.HTTPError as e:
        fatal(f"Photoroom HTTP {e.code}: {e.read().decode()[:300]}")


def upload_to_spree(product_id: str, blob: bytes, filename: str, position):
    # position is always sent. Spree defaults a new asset to position 1, which
    # silently makes it the product thumbnail — a watermarked sandbox image went
    # live on the storefront that way. Callers pass the append position unless
    # they explicitly want the front.
    fields = {"type": "Spree::Image", "position": str(position)}
    body, ctype = multipart(fields, "attachment", filename, blob, "image/png")
    req = urllib.request.Request(
        f"{SPREE_API}/products/{product_id}/media",
        data=body,
        headers={"X-Spree-Api-Key": SPREE_KEY, "Content-Type": ctype},
        method="POST",
    )
    try:
        with urllib.request.urlopen(req, timeout=180) as r:
            return r.status, json.loads(r.read() or b"{}")
    except urllib.error.HTTPError as e:
        fatal(f"Spree HTTP {e.code}: {e.read().decode()[:400]}")


def main():
    ap = argparse.ArgumentParser()
    ap.add_argument("--slug", required=True)
    ap.add_argument("--image", help="source file (default: products/<slug>/image-1.jpg)")
    ap.add_argument("--live", action="store_true", help="use the paid key (10/month, no watermark)")
    ap.add_argument("--position", type=int,
                    help="1 makes it the product thumbnail. Default: appended last.")
    ap.add_argument("--size", default="1000x1000")
    ap.add_argument("--bg", default="FFFFFF")
    ap.add_argument("--dry-run", action="store_true")
    args = ap.parse_args()

    src = pathlib.Path(args.image) if args.image else SOURCE_ROOT / args.slug / "image-1.jpg"
    if not src.exists():
        fatal(f"source image not found: {src}")

    product = find_product(args.slug)
    if not product:
        fatal(f"no Spree product with slug {args.slug}")
    pid = product["id"]
    name = product.get("name", "?")

    before = spree_get(f"/products/{pid}/media").get("data", [])
    print(f"product : {name}  ({pid})")
    print(f"source  : {src}  ({src.stat().st_size:,} bytes)")
    print(f"media   : {len(before)} existing")
    print(f"key     : {'LIVE (consumes paid quota)' if args.live else 'sandbox (watermarked)'}")

    if args.dry_run:
        print("\n--dry-run: nothing sent.")
        return

    key = load_key(args.live)
    position = args.position if args.position is not None else len(before) + 1
    print(f"\nprocessing...  (will land at position {position}"
          f"{' — THUMBNAIL' if position == 1 else ''})")
    out = photoroom(src, key, args.size, args.bg)
    print(f"  photoroom ok: {len(out):,} bytes png")

    status, resp = upload_to_spree(pid, out, f"{args.slug}-photoroom.png", position)
    asset_id = resp.get("data", resp).get("id")
    print(f"  spree upload: HTTP {status}  asset={asset_id}  position={position}")

    # Re-read rather than trusting the response. A 201 is not proof the asset is
    # attached and serving.
    after = spree_get(f"/products/{pid}/media").get("data", [])
    ids = [m.get("id") for m in after]
    if asset_id and asset_id in ids:
        url = next((m.get("original_url") or m.get("url") for m in after if m.get("id") == asset_id), None)
        print(f"\ncheck ok: media {len(before)} -> {len(after)}, asset present")
        if url:
            print(f"url: {url}")
    else:
        print(f"\nCHECK FAILED: asset {asset_id} not in product media after upload", file=sys.stderr)
        print(f"  media ids: {ids}", file=sys.stderr)
        sys.exit(1)


if __name__ == "__main__":
    main()