Support creating FeathubClient from configuration file

In files like `nyc_example.py`, when creating a FeathubClient, apart from listing all the configuration parameters in the format of a dict variable, it should also be supported to read the configurations from a json or yaml file and create the FeathubClient from this. After that, `nyc_example.py`, `nyc_example_flink_session.py`, `nyc_example_spark_client.py` could be restructured into one python file and three configuration files with different processor configurations.

Start by comparing nyc_example.py, nyc_example_flink_session.py, and nyc_example_spark_client.py to identify their FeathubClient configuration differences. Trace how the client currently accepts dict parameters, then define the JSON/YAML configuration entry point; done means one Python example can use three configuration files with distinct processor settings.

Written by the indexing model from the issue text.