"""Verify the exact public Gemma adapter/head releases without re-downloading weights."""
import hashlib
import json
from pathlib import Path
from huggingface_hub import HfApi

ROOT=Path(__file__).resolve().parent

def main():
    api=HfApi(token=False)
    for task in ['open-emotions','open-morals','open-sensations']:
        out=ROOT/'outputs/gemma'/task;stage=out/'publish'
        info=api.model_info(f'mts-the-architect/{task}-gemma4-e2b-bf16-lora',files_metadata=True)
        assert not info.private
        files={f.rfilename:f for f in info.siblings}
        expected={p.name for p in stage.iterdir() if p.is_file()}
        assert expected<=set(files),expected-set(files)
        assert not any(name.endswith(('.sqlite3','.db')) or 'api-private' in name for name in files)
        verified={}
        for name in sorted(expected):
            raw=(stage/name).read_bytes();remote=files[name]
            if remote.lfs:
                digest=hashlib.sha256(raw).hexdigest()
                assert remote.lfs.sha256==digest,name
                verified[name]=dict(method='Hub LFS SHA256',sha256=digest)
            else:
                digest=hashlib.sha1(b'blob '+str(len(raw)).encode()+b'\0'+raw).hexdigest()
                assert remote.blob_id==digest,name
                verified[name]=dict(method='Git blob SHA1',blob_sha1=digest)
        assert {'adapter_model.safetensors','classifier_head.safetensors','synthetic-candidate.zip',
                'real-only-control.zip','human-baseline.zip'}<=verified.keys()
        report=dict(repository=info.id,public=True,commit=info.sha,verified_files=verified,
                    all_release_bytes_verified=True,adapter_head_and_continuation_archives_verified=True)
        (out/'publication_verification.json').write_text(json.dumps(report,indent=2)+'\n')
        print(task,'verified',len(verified),'files at',info.sha,flush=True)

if __name__=='__main__':main()