Separate schema for ML and domain tables

[Zhiweiii]

Enable three schemas in DerivaML:

1. domain_schema
2. ml_schema
3. vocab_schema

To initiate DerivaML:

class DerivaML:
    def __init__(self, hostname: str, catalog_id: str, domain_schema: str, ml_schema: str, vocab_schema: str, data_dir: str):

In Catalog ML, hard coded the three schemas for the catalog.
E.g. EyeAI project

class EyeAI(DerivaML):
    def __init__(self, hostname: str = 'www.eye-ai.org', data_dir: str = './'):
        super().__init__(hostname, 'eye-ai, 'eye-ai', 'eye-ai', data_dir)

FaceBase ML project

class FaceBaseML(DerivaML):
    def __init__(self, hostname: str = 'ml.facebase.org', data_dir: str = './'):
        super().__init__(hostname, 'isa', 'ml', 'vocab', data_dir)

[Zhiweiii]

Changes:
Initiate the Deriva-ML with the ML catalog schema name. For all the rest of the tables, search for the schema it belongs to.