Use meituan hdfs to read or write train data and model

[henshy]

1. Add Meituan Hadoop File and Requirements:
   Add the fedlearner-sgx-dev.dockerfile and requirements.txt for Meituan Hadoop integration.
2. Add Meituan Hadoop Environment in Gramine Template:
   Incorporate Meituan Hadoop environment settings into the python.manifest.template for Gramine.
3. Perform Kerberos Authentication in hdfs_common.sh:
   Execute Kerberos authentication within the hdfs_common.sh script.
4. Implement Meituan HDFS File Read and Write in the Training Step:
   Carry out file read and write operations for Meituan HDFS within the training step, specifically in trainer_master.py and trainer_worker.py.

[zeuson0]

LGTM