[FEA] Allow users to specify custom Dependency jars (spark)

[amahussein]

Is your feature request related to a problem? Please describe.

Users with custom Spark might face issues using the python wrapper if the open source versions of spark is not compatible with the custom build.
It will be helpful to allow the users to set their own dependency in order to use the custom spark/hadoop jars.

Suggested proposal

• Add a new input argument that takes a path to a local json file.
  • --tools_config_file
  • The argument has to be local because we would not be able to access it before setting the spark dependencies.
• the config file can be used in the future to append any configurations that the user is willing to override in the default behavior.
  • the suggested current structure is:

    {
       "dependencies": {
          "deployMode": {
          "LOCAL": {
            "EnvDep1": [
               {
                  "name": "Apache Spark"
                  "uri": "path to the location of the compressed tgz (https:// or file://)"
                  "type": "archive (for tgz files) / jar for jar files"
                  "relativePath": "where to find the jars within the unarchived folder (i.e., jars/*)"
                  "sha512": "XYZ (used for verification)"
                  "size": "size in bytes (used for )"
               },
               {
                  "name": "Storage connector (this can be hadoop jar if it is not included in spark jars or any other required jar",
                  "uri": ""
                  "type": "jar"
                  "relativePath": "where to find the jars within the unarchived folder (i.e., jars/*)"
                  "sha512": "XYZ (used for verification)"
                  "size": "size in bytes (used for )"
               },
               ....define remaining dependencies if needed.
            ]
          }
       }
    }

  the above format requires some changes in current code which is basically detecting local FS vs https