darshan.experimental.aggregators package

Aggregator functions are monkey patched into the DarshanReport object during initialization of DarshanReport class.

Note

These functions are only enabled if the darshan.enable_experimental(True) function is called.

Example usage:

import darshan
import darshan.report
dasrhan.enable_experimental(True)
report = darshan.report.DarshanReport()
...
result = report.agg_ioops()

Submodules

darshan.experimental.aggregators.agg_ioops module

darshan.experimental.aggregators.agg_ioops.agg_ioops(self, mode='append')

Compile the I/O operations summary for the current report.

Parameters:

mode (str) – Whether to ‘append’ (default) or to ‘return’ aggregation.

Returns:

Depending on mode

Return type:

None or dict

darshan.experimental.aggregators.create_dxttimeline module

darshan.experimental.aggregators.create_dxttimeline.create_dxttimeline(self, group_by='rank', mode='append')

Generate/update a timeline from dxt tracing records of current report.

Parameters:

group_by (str) – By which factor to group entries (default: rank) Allowed Parameters: rank, filename

darshan.experimental.aggregators.create_sankey module

darshan.experimental.aggregators.create_sankey.create_sankey(self, mode='append')

Generate a summary that shows the dataflow between ranks, files and their mountpoints.

darshan.experimental.aggregators.create_time_summary module

darshan.experimental.aggregators.create_time_summary.create_time_summary(self, mode='append')

TODO: port to new object report

darshan.experimental.aggregators.create_timeline module

darshan.experimental.aggregators.create_timeline.compress_pathname(pathname)

darshan.experimental.aggregators.create_timeline.configure_groups()

Prepare a dictionary to lookup high level group ordering.

darshan.experimental.aggregators.create_timeline.create_timeline(self, group_by='mod,file,rank', action='attach,overwrite', summary_name='timeline')

Generate/update a timeline from records of the current report state.

Parameters:

group_by (str) – By which factor to group entries (default: rank) Allowed Parameters: rank, filename

darshan.experimental.aggregators.create_timeline.normalize_mod(mod, inverse=False)

darshan.experimental.aggregators.create_timeline.purge_empty_nestedGroups(groups)

darshan.experimental.aggregators.create_timeline.summarized_items(self, groups, items, mod, nmod, rec, rec_id, group_id, parent_group)

darshan.experimental.aggregators.create_timeline.update_parent_item(self, groups, items, parent_id, start=None, end=None)

darshan.experimental.aggregators.mod_agg_iohist module

darshan.experimental.aggregators.mod_agg_iohist.mod_agg_iohist(self, mod, mode='append')

Generate aggregated histogram for mod_name.

Parameters:

mod_name (str) –

Returns:

None

darshan.experimental.aggregators.name_records_summary module

darshan.experimental.aggregators.name_records_summary.name_records_summary(self)

Count records for every name record.

Parameters:

mod_name (str) –

Returns:

dict with counts nested <nrec-id>/<mod>

darshan.experimental.aggregators.print_module_records module

darshan.experimental.aggregators.print_module_records.print_module_records(self, mode='append')

Compile the I/O operations summary for the current report.

Parameters:

mode (str) – Whether to ‘append’ (default) or to ‘return’ aggregation.

Returns:

Depending on mode

Return type:

None or dict

darshan.experimental.aggregators.records_as_dict module

darshan.experimental.aggregators.records_as_dict.records_as_dict(self, mode='append', recs=None)

Convert all counters to dictionaries with human-readable names.

Parameters:

mode (str) – Whether to ‘append’ (default) or to ‘return’ aggregation.

Returns:

Depending on mode

Return type:

None or dict

darshan.experimental.aggregators.summarize module

darshan.experimental.aggregators.summarize.summarize(self)

Compiles a report summary of the records present in the report object.

Parameters:

None –

Returns:

None