1
0
mirror of https://github.com/ARM-software/workload-automation.git synced 2024-10-06 10:51:13 +01:00

doc: Update Output processing API docs

This commit is contained in:
Marc Bonnici 2018-06-07 17:08:41 +01:00 committed by setrofim
parent 4b5d85ce40
commit 0196d3bdbe
2 changed files with 54 additions and 6 deletions

View File

@ -15,6 +15,7 @@ Developer Reference
----------------- -----------------
.. include:: developer_reference/output_processing_api.rst .. include:: developer_reference/output_processing_api.rst
.. include:: developer_reference/APIs.rst
----------------- -----------------

View File

@ -3,14 +3,61 @@
Output Processing API Output Processing API
===================== =====================
WA3 now has an output API that can be used to post process a run's WA3 now has an output API that can be used to post process a runs
:ref:`Output Directory Structure<output_directory>` which can be performed by using WA's :ref:`Output Directory Structure <output_directory>` which can be performed by using WA's
``RunOutput`` object. :class:`RunOutput` object.
Example: This object provides a simple way to access the various aspects of the run. A
few of the most useful attributes of the ``RunOutput`` object are as follows:
If we have an existing WA output called ``wa_output`` in the current working - :confval:`jobs` - A list of all the job outputs that were generated for the
directory we can initialize a ``RunOutput`` as follows: run. These can be used to used to access more detailed information about
individual jobs for example their ``status``, any ``metrics`` or ``artifacts``
that were generated etc.
- :confval:`run_config` - A :ref:`RunConfiguration <run-configuration>` object
that can be used to access all the configuration of the run itself, for
example the ``reboot_policy``, ``execution_order``, ``device_config`` etc.
- :confval:`target_info` - A :ref:`TargetInfo <target-info-api>` object which can be used to
access various information about the target that was used during the run for
example it's ``abi``, ``hostname``, ``os`` etc.
- :confval:`info` - A :ref:`RunInfo <run-info-api>` object that contains
information about the run itself for example it's duration, name, uuid etc.
- :confval:`settings` - The :ref:`Meta Configuration <meta-configuration>` used
during the run.
- :confval:`status` - The overall status of the run.
- :confval:`artifacts` - A list containing any run level artifacts (for example
logs, output generated by output processors etc.).
- :confval:`metrics` - A list of :ref:`Metrics <metrics>` that were generated
for the overall run.
- :confval:`classifiers` - A dict mapping of any run level
:ref:`classifiers <classifiers>`.
- :confval:`augmentations` - A list of the augmentations that were enabled
during the run.
- :confval:`event_summary` - A condensed summary of any events that occurred
during the run.
- :confval:`events` - A list of any events that occurred during the run like job
failures etc.
- :confval:`basepath` - The name of the base WA output directory.
- :confval:`configfile` - The relative file path to the config file generated
for the run.
- :confval:`failed_dir` - The relative file path to the directory containing any
failed jobs during the run.
- :confval:`infofile` - The file location of the json serialized file of the
:class:`RunInfo` object.
- :confval:`list_workloads()` - A method that generated a list of the workload
names that were executed during the run.
- :confval:`logfile` - A relative file path to the run log.
- :confval:`metadir` - A relative file path to the metadata directory.
- :confval:`targetfile` - The file location of the json serialized file of the
:class:`TargetInfo` object.
Output Processing API Example
------------------------------
To demonstrate how we can use the output API if we have an existing WA output
called ``wa_output`` in the current working directory we can initialize a
``RunOutput`` as follows:
.. code-block:: python .. code-block:: python