Map Dashboard¶
HiveWatch ships with a local dashboard flow based on saved run artifacts and a small HTTP server.
Generate run artifacts¶
Use SSEEmitter during training to create:
runs/<run_id>.jsonlfor the full event streamruns/<run_id>.map.jsonfor map-friendly round and client metadata
import hivewatch as hw
from hivewatch.emitters import SSEEmitter
hw.init(emitters=[SSEEmitter(serve_map=False)])
Serve the dashboard¶
hivewatch map run --runs-dir runs --port 7070
Useful flags¶
--hostchanges the bind address.--map-pathpoints at a custom HTML viewer.--poll-intervalcontrols how often the runs directory is rescanned.
The bundled viewer in the examples/ directory loads map metadata first and
falls back to the raw event history for older runs. That means the same viewer
works for both live monitoring and deferred replay.
Common map commands¶
# Watch a runs directory for live updates
hivewatch map run --runs-dir runs --port 7070
# Serve a custom viewer HTML file
hivewatch map run --runs-dir runs --map-path /path/to/viewer.html
# Bind the dashboard to a specific interface
hivewatch map run --host 0.0.0.0 --runs-dir runs --port 7070
Deferred viewing¶
Because HiveWatch persists map metadata separately from the raw JSONL log, you can train first and inspect later. This keeps a live workflow and a replay workflow compatible with the same dashboard interface.