Adding embedding column to dataset descriptor (#3736)

Summary:
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/3736

Nit - adding embedding column in dataset descriptor
Nit - initializing cached_ds as part of class instead of post_init

Reviewed By: satymish

Differential Revision: D60858496

fbshipit-source-id: 3358d866a0668424cd6895bc7a5c620ff97e72fa

benchs/bench_fw/benchmark_io.py

@@ -10,7 +10,7 @@ import logging
 import os
 import pickle
 from dataclasses import dataclass
-from typing import Any, List, Optional
+from typing import Any, Dict, List, Optional
 from zipfile import ZipFile
 
 import faiss  # @manual=//faiss/python:pyfaiss_gpu
@@ -46,13 +46,11 @@ def merge_rcq_itq(
 @dataclass
 class BenchmarkIO:
     path: str
+    cached_ds: Dict[Any, Any] = {}
 
     def clone(self):
         return BenchmarkIO(path=self.path)
 
-    def __post_init__(self):
-        self.cached_ds = {}
-
     # TODO(kuarora): rename it as get_local_file
     def get_local_filename(self, filename):
         if len(filename) > 184:

benchs/bench_fw/descriptors.py

@@ -78,6 +78,8 @@ class DatasetDescriptor:
     # number of vectors to load from the dataset
     num_vectors: Optional[int] = None
 
+    embedding_column: Optional[str] = None
+
     def __hash__(self):
         return hash(self.get_filename())