Skip to content
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

chore: Supporting dataset level job queuing #386

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

chore: Supporting dataset level job queuing #386

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
96 changes: 66 additions & 30 deletions ingestion_tools/scripts/enqueue_runs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -360,6 +360,12 @@ def db_import(
multiple=False,
help="Exclude runs matching this regex. If not specified, all runs are processed",
)
@click.option(
"--queue-datasets/--queue-runs",
type=bool,
default=False,
help="Queue jobs for datasets instead of runs",
)
@ingest_common_options
@enqueue_common_options
@click.pass_context
Expand All @@ -374,6 +380,7 @@ def queue(
force_overwrite: bool,
swipe_wdl_key: str,
skip_until_run_name: str,
queue_datasets: bool,
**kwargs,
):
handle_common_options(ctx, kwargs)
Expand All @@ -396,21 +403,63 @@ def queue(
filter_datasets = [re.compile(pattern) for pattern in kwargs.get("filter_dataset_name", [])]
exclude_datasets = [re.compile(pattern) for pattern in kwargs.get("exclude_dataset_name", [])]

wdl_args = {
"config_file": config_file,
"input_bucket": input_bucket,
"output_path": output_path,
}

# Always iterate over depostions, datasets and runs.
for deposition in DepositionImporter.finder(config):
print(f"Processing deposition: {deposition.name}")
datasets = DatasetImporter.finder(config, deposition=deposition)
for dataset in datasets:
if list(filter(lambda x: x.match(dataset.name), exclude_datasets)):
print(f"Excluding {dataset.name}..")
continue
if filter_datasets and not list(filter(lambda x: x.match(dataset.name), filter_datasets)):
print(f"Skipping {dataset.name}..")
continue
print(f"Processing dataset: {dataset.name}")
runs = RunImporter.finder(config, dataset=dataset)
futures = []
with ProcessPoolExecutor(max_workers=ctx.obj["parallelism"]) as workerpool:
futures = []

with ProcessPoolExecutor(max_workers=ctx.obj["parallelism"]) as workerpool:
for dataset in datasets:
if list(filter(lambda x: x.match(dataset.name), exclude_datasets)):
print(f"Excluding {dataset.name}..")
continue
if filter_datasets and not list(filter(lambda x: x.match(dataset.name), filter_datasets)):
print(f"Skipping {dataset.name}..")
continue
print(f"Processing dataset: {dataset.name}")

per_dataset_args = {}
# Don't copy over dataset and run name filters to the queued jobs - they're intended to be
# batched into 1-dataset chunks.
excluded_args = ["filter_dataset_name"]
for k, v in kwargs.items():
if any(substring in k for substring in excluded_args):
break
per_dataset_args[k] = v

dataset_id = dataset.name
deposition_id = deposition.name
ds_execution_name = f"{int(time.time())}-dep{deposition_id}-ds{dataset_id}"
if queue_datasets:
new_args = to_args(
import_everything=import_everything,
write_mrc=write_mrc,
write_zarr=write_zarr,
force_overwrite=force_overwrite,
**per_dataset_args,
) # make a copy
new_args.append(f"--filter-dataset-name '^{dataset.name}$'")
futures.append(
workerpool.submit(
partial(
run_job,
ds_execution_name,
{**{"flags": " ".join(new_args)}, **wdl_args},
swipe_wdl_key=swipe_wdl_key,
**ctx.obj,
),
),
)
continue

runs = RunImporter.finder(config, dataset=dataset)
for run in runs:
if skip_run and not skip_run_until_regex.match(run.name):
print(f"Skipping {run.name}..")
Expand All @@ -425,14 +474,10 @@ def queue(
continue
print(f"Processing {run.name}...")

per_run_args = {}
# Don't copy over dataset and run name filters to the queued jobs - they're intended to be
per_run_args = dict(per_dataset_args)
# Don't copy over run name filters to the queued jobs - they're intended to be
# batched into 1-run chunks.
excluded_args = ["filter_dataset_name", "filter_run_name"]
for k, v in kwargs.items():
if any(substring in k for substring in excluded_args):
break
per_run_args[k] = v
per_run_args.pop("filter_run_name", None)
new_args = to_args(
import_everything=import_everything,
write_mrc=write_mrc,
Expand All @@ -443,32 +488,23 @@ def queue(
new_args.append(f"--filter-dataset-name '^{dataset.name}$'")
new_args.append(f"--filter-run-name '^{run.name}$'")

dataset_id = dataset.name
deposition_id = deposition.name
execution_name = f"{int(time.time())}-dep{deposition_id}-ds{dataset_id}-run{run.name}"
execution_name = f"{int(time.time())}-{ds_execution_name}-run{run.name}"

# execution name greater than 80 chars causes boto ValidationException
if len(execution_name) > 80:
execution_name = execution_name[-80:]

wdl_args = {
"config_file": config_file,
"input_bucket": input_bucket,
"output_path": output_path,
"flags": " ".join(new_args),
}
futures.append(
workerpool.submit(
partial(
run_job,
execution_name,
wdl_args,
{**{"flags": " ".join(new_args)}, **wdl_args},
swipe_wdl_key=swipe_wdl_key,
**ctx.obj,
),
),
)
wait(futures)
wait(futures)


class OrderedSyncFilters(click.Command):
Expand Down
Loading