inline_data.py

"""Inline media.json into index.html so the page works under file:// (no fetch).

Idempotent: handles both the first run (placeholder token __MEDIA__) and
subsequent re-runs (replaces the JSON inside the existing
<script id="media-data" type="application/json">...</script> tag).

Re-run after editing media.json.
"""
import json, os, re

ROOT = os.path.dirname(os.path.abspath(__file__))
TPL_PATH = os.path.join(ROOT, "index.html")

DATA_FILES = {
    "media-data": os.path.join(ROOT, "media.json"),
}

TOKEN_FALLBACK = {
    "media-data": "__MEDIA__",
}


def normalize_json(text):
    return json.dumps(json.loads(text), ensure_ascii=False)


def main():
    src = open(TPL_PATH, encoding="utf-8").read()
    for script_id, path in DATA_FILES.items():
        has_tag = re.search(
            r'<script\s+id="' + re.escape(script_id) + r'"\s+type="application/json">',
            src,
        )
        token = TOKEN_FALLBACK.get(script_id, "")
        has_token = token and token in src
        if not has_tag and not has_token:
            continue
        if not os.path.exists(path):
            print(f"  skip #{script_id} (data file missing: {path})")
            continue
        with open(path, encoding="utf-8") as f:
            payload = normalize_json(f.read())
        payload = payload.replace("</script>", "<\\/script>")

        # Replace contents of an existing tag in preference to the placeholder.
        pat = re.compile(
            r'(<script\s+id="' + re.escape(script_id) + r'"\s+type="application/json">)([\s\S]*?)(</script>)'
        )
        new_src, n = pat.subn(lambda m: m.group(1) + payload + m.group(3), src, count=1)
        if n:
            print(f"  refreshed #{script_id}  ({len(payload):,} chars)")
            src = new_src
            continue

        if has_token:
            src = src.replace(token, payload, 1)
            print(f"  inlined {token}  ({len(payload):,} chars)")

    with open(TPL_PATH, "w", encoding="utf-8") as f:
        f.write(src)
    print(f"Wrote {TPL_PATH}  ({len(src):,} bytes)")


if __name__ == "__main__":
    main()