tao-shen Claude Opus 4.6 commited on
Commit
b0f06a1
·
1 Parent(s): a902a81

test: revert dataset derivation to observe linked spaces behavior

Browse files
Files changed (1) hide show
  1. scripts/sync_hf.py +7 -5
scripts/sync_hf.py CHANGED
@@ -80,14 +80,16 @@ SPACE_ID = os.environ.get("SPACE_ID", "") # e.g. "tao-shen/HuggingClaw"
80
  SYNC_INTERVAL = int(os.environ.get("SYNC_INTERVAL", "60"))
81
  AUTO_CREATE_DATASET = os.environ.get("AUTO_CREATE_DATASET", "false").lower() in ("true", "1", "yes")
82
 
83
- # Dataset repo: user-specified takes priority, then auto-derive from SPACE_ID.
 
 
84
  HF_REPO_ID = os.environ.get("OPENCLAW_DATASET_REPO", "")
85
- if HF_REPO_ID:
86
- print(f"[SYNC] Using OPENCLAW_DATASET_REPO: {HF_REPO_ID}")
87
- elif SPACE_ID:
88
  HF_REPO_ID = f"{SPACE_ID}-data"
89
  print(f"[SYNC] OPENCLAW_DATASET_REPO not set — auto-derived from SPACE_ID: {HF_REPO_ID}")
90
- elif HF_TOKEN:
 
91
  try:
92
  _api = HfApi(token=HF_TOKEN)
93
  _username = _api.whoami()["name"]
 
80
  SYNC_INTERVAL = int(os.environ.get("SYNC_INTERVAL", "60"))
81
  AUTO_CREATE_DATASET = os.environ.get("AUTO_CREATE_DATASET", "false").lower() in ("true", "1", "yes")
82
 
83
+ # Dataset repo: always auto-derive from SPACE_ID when not explicitly set.
84
+ # Format: {username}/{SpaceName}-data (e.g. "tao-shen/HuggingClaw-data")
85
+ # This ensures each duplicated Space gets its own dataset automatically.
86
  HF_REPO_ID = os.environ.get("OPENCLAW_DATASET_REPO", "")
87
+ if not HF_REPO_ID and SPACE_ID:
88
+ # SPACE_ID = "username/SpaceName" derive "username/SpaceName-data"
 
89
  HF_REPO_ID = f"{SPACE_ID}-data"
90
  print(f"[SYNC] OPENCLAW_DATASET_REPO not set — auto-derived from SPACE_ID: {HF_REPO_ID}")
91
+ elif not HF_REPO_ID and HF_TOKEN:
92
+ # Fallback: no SPACE_ID (local Docker), derive from HF_TOKEN username
93
  try:
94
  _api = HfApi(token=HF_TOKEN)
95
  _username = _api.whoami()["name"]