art-dag/celery
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix IPFS CID detection in legacy_tasks node caching

Browse Source 
- Rename misleading ipfs_cid variable to content_cid
- Detect IPFS CIDs by prefix (Qm or bafy) instead of truthy check
- Add clearer logging to show whether IPFS or local hash is used

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
gilesb
2026-01-12 17:13:25 +00:00
parent cc29311d1c
commit a0b113923e
1 changed files with 10 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
legacy_tasks.py
View File

@@ -354,15 +354,20 @@ def execute_dag(self, dag_json: str, run_id: str = None) -> dict:
cache_node_type = "dag_intermediate" cache_node_type = "dag_intermediate"
# Store in cache_manager (indexes by cid, uploads to IPFS) # Store in cache_manager (indexes by cid, uploads to IPFS)
cached, ipfs_cid = cache_manager.put( # put() returns (CachedFile, cid) where cid is IPFS CID if available, else local hash
cached, content_cid = cache_manager.put(
Path(node_path), Path(node_path),
node_type=cache_node_type, node_type=cache_node_type,
node_id=node_id, node_id=node_id,
) )
node_hashes[node_id] = ipfs_cid or cached.cid # Prefer IPFS CID # content_cid is the primary identifier (IPFS CID or local hash)
if ipfs_cid: node_hashes[node_id] = content_cid
node_ipfs_cids[node_id] = ipfs_cid # Track IPFS CIDs separately (they start with Qm or bafy)
logger.info(f"Cached node {node_id}: {cached.cid[:16]}... -> {ipfs_cid or 'no IPFS'}") if content_cid and (content_cid.startswith("Qm") or content_cid.startswith("bafy")):
node_ipfs_cids[node_id] = content_cid
logger.info(f"Cached node {node_id}: IPFS CID {content_cid}")
else:
logger.info(f"Cached node {node_id}: local hash {content_cid[:16] if content_cid else 'none'}...")
# Get output hash from the output node # Get output hash from the output node
# Use the same identifier that's in the cache index (IPFS CID if available) # Use the same identifier that's in the cache index (IPFS CID if available)