Add support for maxcompute as source type (batch)

[vinoth-gojek]

The changes in MR are to enable MaxCompute as a datasource type for batch jobs (offline to online ingestion)

1. The MR adds the BATCH_MAXCOMPUTE enum
2. The changes in sdk enable clients to use maxcompute as source type
3. The changes include triggering spark jobs to fetch records from maxcompute
4. The prebuilt custom jar to use custom dialect while calling maxcompute is used to overcome issues

With these changes, record with batch_source maxcompute is added to the datasources table

postgres=> select * from data_sources;
42 | BATCH_MAXCOMPUTE | {"table_ref":"datasource:playground.dsp_maxcompute_dogfood"} | {} | created_timestamp | created_timestamp |

[deadlycoconuts]

LGTM! Thanks for poking around the entire Feast codebase, for picking up scala too and of course testing these changes just to get this PR up! I just left a couple of tiny comments but feel free to merge this PR whenever you're satisfied with the changes 🚀