If the current job is the latest one in the loader pool and the job is deleted,
the job goes into an infinite cycle of load-fail-delete-reload-load-fail-delete.
So we need to prevent the infinite loop during XAI data loading failed.

1. delete the unused code, catch unknown error is unused in explain_parser.py
2. fix bug the lastest file size is unused
3. rename some variables and fix some spelling error

Support to stop the loader that is currently loading data when add a new loader to cache by get_job method
Add the UT usecase to the stop mechanism

FIFO queue mechanism is adopted for loader cache

Merge pull request !816 from TonyNG/master

fix UT

del unused arguments

fix UT

explainer_api UT change to pytest style

fix UT url

explainer encapsulator UT change to pytest style

remove unnecessary deepcopy

remove import copy

fix some coding style issues