repro-bundle-v2/repro/verify_live.py

"""Byte-compare live raw published pages against local page.md files.

trackio's URI rewrite can reintroduce CRLF on Windows, which makes a page parse
as zero cells and silently kills every figure. This catches that.
"""
import hashlib
import json
import os
import urllib.request

import huggingface_hub

ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
PAGES = os.path.join(ROOT, ".trackio", "logbook", "pages")
SPACE = ("algorise/repro-ddsvm-a-differentiable-framework-for-deep-support-"
         "vector-machines-with-iterative-geometry")
TOK = huggingface_hub.get_token()

lb = json.load(open(os.path.join(ROOT, ".trackio", "logbook", "logbook.json")))
files = ["pages/index.md"] + [c["file"] for c in lb["root"]["children"]]

ok = True
for rel in files:
    url = f"https://huggingface.co/spaces/{SPACE}/raw/main/{rel}"
    req = urllib.request.Request(url, headers={"Authorization": f"Bearer {TOK}"})
    live = urllib.request.urlopen(req).read()
    local = open(os.path.join(ROOT, ".trackio", "logbook", rel), "rb").read()
    same = live == local
    crlf = b"\r\n" in live
    ncells = live.count(b"<!-- trackio-cell")
    ok &= same and not crlf
    print(f"{'OK  ' if (same and not crlf) else 'FAIL'} {rel}")
    print(f"      live {len(live)}B sha1={hashlib.sha1(live).hexdigest()[:12]} | "
          f"local {len(local)}B sha1={hashlib.sha1(local).hexdigest()[:12]} | "
          f"identical={same} CRLF={crlf} cells={ncells}")

print("\nALL PAGES BYTE-IDENTICAL AND LF-CLEAN" if ok else "\nMISMATCH DETECTED")