Add example: export events for external workflow-mining tools - #122
aashnology wants to merge 2 commits into
Conversation
Adds examples/export_for_workflow_mining.py, which reuses the existing
categorizing + AFK-filtering canonical query (same approach as
load_dataframe.py) and writes a flat, sorted JSON list of
{timestamp, duration, data} records -- the minimal shape a downstream
sequence-clustering or process-mining tool needs, without every raw
watcher heartbeat a full bucket dump would include.
Passes ruff, mypy, and py_compile with the repo's existing config.
README updated to list it alongside the other examples.
 |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| {"timestamp": e["timestamp"], "duration": e["duration"], "data": e["data"]} | ||
| for e in data[0]["events"] | ||
| ] | ||
| events.sort(key=lambda e: e["timestamp"]) |
There was a problem hiding this comment.
| with open(args.out, "w") as f: | ||
| json.dump(events, f, indent=2) |
There was a problem hiding this comment.
"w" immediately truncates any existing export. If serialization fails or the process is interrupted, the previous complete file is left empty or partial. Write to a temporary file and replace the destination only after the export succeeds.
| if __name__ == "__main__": | ||
| main() |
There was a problem hiding this comment.
__main__ guard means it never executes the query or writes JSON. A broken export could therefore pass the existing checks. Add a test or example-test invocation that checks the resulting JSON.
Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!
- sort by iso8601.parse_date(), not the raw timestamp string: two valid ISO8601 timestamps using different UTC offsets can sort backwards as strings even though one is unambiguously earlier once parsed - write via a temp file + os.replace() so a failure partway through serialization can't leave a truncated file in place of a prior good export - add test_* functions (matching suggest_categories_gpt.py's convention) covering the query-building, sorting, field-filtering, and atomic-write behavior -- runnable via 'make test-examples'
1 participant
Adds
examples/export_for_workflow_mining.py, which reuses the existing categorizing + AFK-filtering canonical query (same pattern asload_dataframe.py) and writes a flat, time-sorted JSON list of{timestamp, duration, data}records — the minimal shape a downstream sequence-clustering or process-mining tool needs, without every raw watcher heartbeat a full bucket dump would include.Tested against the repo's own tooling: passes
py_compile,ruff check, andmypywith the existing config. README updated to list it alongside the other examples.