Skip to content

feat: fault-tolerant processing #33

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
alexander-held wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

feat: fault-tolerant processing #33

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
40 changes: 34 additions & 6 deletions atlas/analysis.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -270,17 +270,45 @@
" columns_to_preload = json.load(pathlib.Path(\"columns_to_preload.json\").open())[\"JET_JER_Effective\"] # or []\n",
"\n",
" with performance_report(filename=\"process_custom.html\"):\n",
" out, report = utils.custom_process(preprocess_output, processor_class=Analysis, schema=run.schema, client=client, preload=columns_to_preload)\n",
" # out, report = utils.custom_process(preprocess_output, processor_class=Analysis, schema=run.schema, client=client, preload=columns_to_preload)\n",
"\n",
" # version from https://github.com/iris-hep/integration-challenge/pull/23\n",
" import sys\n",
" sys.path.insert(0, \"..\")\n",
"\n",
" import util._dask\n",
" cloudpickle.register_pickle_by_value(util)\n",
"\n",
" futures, futurekey2item = util._dask.dask_map(\n",
" Analysis().process,\n",
" preprocess_output,\n",
" client=client,\n",
" NanoEventsFactory_kwargs = {\n",
" \"preload\": lambda b: b in [],\n",
" \"schemaclass\": run.schema,\n",
" }\n",
" )\n",
"\n",
" final_future, failed_items = util._dask.dask_reduce(\n",
" futures,\n",
" futurekey2item=futurekey2item,\n",
" client=client,\n",
" treereduction=4\n",
" )\n",
" result = final_future.result()\n",
" out, report = result[\"out\"], result[\"report\"]\n",
" print(f\"failed chunks: {failed_items}\")\n",
"\n",
" print(f\"preloaded columns: {len(columns_to_preload)}, {columns_to_preload[:4]} {\"etc.\" if len(columns_to_preload) > 4 else \"\"}\")\n",
" print(f\"preloaded but unused columns: {len([c for c in columns_to_preload if c not in report[\"columns\"]])}\")\n",
" print(f\"used but not preloaded columns: {len([c for c in report[\"columns\"] if c not in columns_to_preload])}\")\n",
"\n",
" # shortened version of report, dropping extra columns and per-chunk information\n",
" display(\n",
" dict((k, v) for k, v in report.items() if k not in [\"columns\", \"chunk_info\"]) | \n",
" {\"columns\": report[\"columns\"][0:10] + [\"...\"]} | \n",
" {\"chunk_info\": list(report[\"chunk_info\"].items())[:2] + [\"...\"]}\n",
" )\n",
" # display(\n",
" # dict((k, v) for k, v in report.items() if k not in [\"columns\", \"chunk_info\"]) | \n",
" # {\"columns\": report[\"columns\"][0:10] + [\"...\"]} | \n",
" # {\"chunk_info\": list(report[\"chunk_info\"].items())[:2] + [\"...\"]}\n",
" # )\n",
"\n",
"else:\n",
" # coffea Runner-based processing\n",
Expand Down