The sessions command lets you inspect the sessions created by detect and pipeline. Each session is a directory that contains all results and metadata for a single analysis run.
Syntax
./run-cli sessions [options]
Options
Print a table of all sessions with their name, creation time, status, and path. Short alias: -l.
Print the path of the most recently created session. Useful for piping into other commands.
Session storage location
Sessions are stored in a centralized directory managed by the StorageManager. Running sessions without flags prints the current storage path:
./run-cli sessions
# Storage location: /home/user/.local/share/archeo-cluster/sessions
# Use --list to see all sessions or --latest to get the latest session path.
Example output
—list
Analysis Sessions
┌──────────────────┬──────────────────┬──────────────┬──────────────────────────────────────┐
│ Session │ Created │ Status │ Path │
├──────────────────┼──────────────────┼──────────────┼──────────────────────────────────────┤
│ excavation-2025 │ 2025-06-01 14:32 │ completed │ /home/user/.local/share/... │
│ site-b-test │ 2025-05-28 09:15 │ completed │ /home/user/.local/share/... │
│ draft-run │ 2025-05-20 17:44 │ error │ /home/user/.local/share/... │
└──────────────────┴──────────────────┴──────────────┴──────────────────────────────────────┘
Storage location: /home/user/.local/share/archeo-cluster/sessions
—latest
./run-cli sessions --latest
# /home/user/.local/share/archeo-cluster/sessions/excavation-2025
Using session paths for downstream analysis
The --latest flag is convenient when chaining commands manually after a detect run:
# Run detection and capture the session path
./run-cli detect -i ./data/raw/images --session my-site
SESSION=$(./run-cli sessions --latest)
# Pass the session's features CSV directly to cluster
./run-cli cluster -i "$SESSION/detection/features.csv"
# Then analyze each clustered CSV
./run-cli analyze -i "$SESSION/clusters/image01/image01_clustered.csv"
The pipeline command handles this chaining automatically. Use sessions for manual workflows or when rerunning individual stages on existing session data.
