Implementation of PyTorch ut parsing script - QA helper function

[BLOrange-AMD]

This PR is to implement helper scripts/functions for running PyTorch unit tests and extracting results using xml test-reports.
The scripts returns the test results in the form of a python multi-level dictionary with the following structure:

{'workflow_name_1': {
test_file_and_status(file_name='workflow_1_aggregate', status='STATISTICS'): {},
test_file_and_status(file_name='test_file_name_1', status='ERROR'): {},
test_file_and_status(file_name='test_file_name_1', status='FAILED'): {},
test_file_and_status(file_name='test_file_name_1', status='PASSED'): {},
test_file_and_status(file_name='test_file_name_1', status='SKIPPED'): {},
test_file_and_status(file_name='test_file_name_1', status='STATISTICS'): {},
test_file_and_status(file_name='test_file_name_1', status='XFAILED'): {},
test_file_and_status(file_name='test_file_name_2', status='ERROR'): {},
test_file_and_status(file_name='test_file_name_2', status='FAILED'): {},
test_file_and_status(file_name='test_file_name_2', status='PASSED'): {},
test_file_and_status(file_name='test_file_name_2', status='SKIPPED'): {},
test_file_and_status(file_name='test_file_name_2', status='STATISTICS'): {},
test_file_and_status(file_name='test_file_name_2', status='XFAILED'): {},}
'workflow_name_2': {
test_file_and_status(file_name='workflow_2_aggregate', status='STATISTICS'): {},
test_file_and_status(file_name='test_file_name_3', status='ERROR'): {},
test_file_and_status(file_name='test_file_name_3', status='FAILED'): {},
test_file_and_status(file_name='test_file_name_3', status='PASSED'): {},
test_file_and_status(file_name='test_file_name_3', status='SKIPPED'): {},
test_file_and_status(file_name='test_file_name_3', status='STATISTICS'): {},
test_file_and_status(file_name='test_file_name_3', status='XFAILED'): {},
test_file_and_status(file_name='test_file_name_4', status='ERROR'): {},
test_file_and_status(file_name='test_file_name_4', status='FAILED'): {},
test_file_and_status(file_name='test_file_name_4', status='PASSED'): {},
test_file_and_status(file_name='test_file_name_4', status='SKIPPED'): {},
test_file_and_status(file_name='test_file_name_4', status='STATISTICS'): {},
test_file_and_status(file_name='test_file_name_4', status='XFAILED'): {},}
.
.
.
.
.
.
}

Confluence page for design: https://confluence.amd.com/display/~boli1003/PyTorch+test+script+for+QA+team

Tests:
entire_tests: test_entire_tests.log
python .automation_script/run_pytorch_unit_tests.py 2>&1 | tee test_entire_tests.log

priority_test: test_priority_tests.log
python .automation_script/run_pytorch_unit_tests.py --test_config default distributed --priority_test 2>&1 | tee test_priority_tests.log

The priority log also shows separate entries for the "nccl" and "gloo" backends for distributed.test_distributed_spawn.

selected_tests: test_selected_tests.log
python .automation_script/run_pytorch_unit_tests.py --default_list test_weak test_dlpack --inductor_list inductor/test_torchinductor 2>&1 | tee test_selected_tests.log

Run from API test

[BLOrange-AMD]

Link for the implementation design: https://confluence.amd.com/display/~boli1003/PyTorch+test+script+for+QA+team

[jithunnair-amd]

@pruthvistony @BLOrange-AMD Some naming-related change requests:

• Let's name the directory to start with a dot (".") to distinguish it from the directories related to actual PyTorch functionality: from automation_script to .automation_scripts (with an "s" at the end)
• Rename ut_results_parse_automation.py to run_pytorch_unit_tests.py

[jithunnair-amd]

@BLOrange-AMD Please update the description of this PR (the "< TBD >" part) as well as add documentation to the script itself to describe:

1. the two ways of using this script: Implementation of PyTorch ut parsing script - QA helper function #1386 (comment)
2. the data structure of the results returned by the script, with examples

[jithunnair-amd]

We'll address refactor-related changes in another PR. Let's merge this if the interface is finalized.