fix(ranker): correct column names in ranker score matrix

jina/drivers/rank/aggregate/__init__.py

@@ -3,7 +3,7 @@
 
 import numpy as np
 
-from ....executors.rankers import Chunk2DocRanker
+from ....executors.rankers import Chunk2DocRanker, COL_STR_TYPE
 from ....types.document import Document
 from ....types.score import NamedScore
 
@@ -124,9 +124,9 @@ def _apply_all(self, docs: 'DocumentSet',
         if match_idx:
             match_idx = np.array(match_idx,
                                  dtype=[
-                                     (Chunk2DocRanker.COL_MATCH_PARENT_ID, np.object),
-                                     (Chunk2DocRanker.COL_MATCH_ID, np.object),
-                                     (Chunk2DocRanker.COL_DOC_CHUNK_ID, np.object),
+                                     (Chunk2DocRanker.COL_PARENT_ID, COL_STR_TYPE),
+                                     (Chunk2DocRanker.COL_DOC_CHUNK_ID, COL_STR_TYPE),
+                                     (Chunk2DocRanker.COL_QUERY_CHUNK_ID, COL_STR_TYPE),
                                      (Chunk2DocRanker.COL_SCORE, np.float64)
                                  ]
                                  )
@@ -203,9 +203,9 @@ def _apply_all(self, docs: 'DocumentSet', context_doc: 'Document', *args,
         if match_idx:
             match_idx = np.array(match_idx,
                                  dtype=[
-                                     (Chunk2DocRanker.COL_MATCH_PARENT_ID, np.object),
-                                     (Chunk2DocRanker.COL_MATCH_ID, np.object),
-                                     (Chunk2DocRanker.COL_DOC_CHUNK_ID, np.object),
+                                     (Chunk2DocRanker.COL_PARENT_ID, COL_STR_TYPE),
+                                     (Chunk2DocRanker.COL_DOC_CHUNK_ID, COL_STR_TYPE),
+                                     (Chunk2DocRanker.COL_QUERY_CHUNK_ID, COL_STR_TYPE),
                                      (Chunk2DocRanker.COL_SCORE, np.float64)
                                  ]
                                  )

jina/executors/rankers/__init__.py

@@ -7,6 +7,8 @@
 
 from .. import BaseExecutor
 
+COL_STR_TYPE = 'U64'  #: the ID column data type for score matrix
+
 
 class BaseRanker(BaseExecutor):
     """The base class for a `Ranker`"""
@@ -31,18 +33,10 @@ class Chunk2DocRanker(BaseRanker):
     """
 
     required_keys = {'text'}  #: a set of ``str``, key-values to extracted from the chunk-level protobuf message
-    """set: Set of required keys to be extracted from matches and query to fill the information of `query` and `chunk` meta information.
-    These are the set of keys to be extracted from `Document`.
-
-    All the keys not found in the `DocumentProto` fields, will be extracted from the `tags` structure of `Document`.
-    .. seealso::
 
-        :meth:`get_attrs` of :class:`Document`
-
-    """
-    COL_MATCH_PARENT_ID = 'match_parent_id'
-    COL_MATCH_ID = 'match_id'
-    COL_DOC_CHUNK_ID = 'doc_chunk_id'
+    COL_PARENT_ID = 'match_parent_id'
+    COL_DOC_CHUNK_ID = 'match_doc_chunk_id'
+    COL_QUERY_CHUNK_ID = 'match_query_chunk_id'
     COL_SCORE = 'score'
 
     def score(self, match_idx: 'np.ndarray', query_chunk_meta: Dict, match_chunk_meta: Dict) -> 'np.ndarray':
@@ -80,7 +74,7 @@ def group_by_doc_id(self, match_idx):
         :return: an iterator over the groups.
         :rtype: :class:`Chunk2DocRanker`.
         """
-        return self._group_by(match_idx, self.COL_MATCH_PARENT_ID)
+        return self._group_by(match_idx, self.COL_PARENT_ID)
 
     @staticmethod
     def _group_by(match_idx, col_name):
@@ -103,14 +97,14 @@ def sort_doc_by_score(r):
         :rtype: np.ndarray
         """
         r = np.array(r, dtype=[
-            (Chunk2DocRanker.COL_MATCH_PARENT_ID, np.object),
+            (Chunk2DocRanker.COL_PARENT_ID, COL_STR_TYPE),
             (Chunk2DocRanker.COL_SCORE, np.float64)]
                      )
         return np.sort(r, order=Chunk2DocRanker.COL_SCORE)[::-1]
 
     def get_doc_id(self, match_with_same_doc_id):
         """Return document id that matches with given id :param:`match_with_same_doc_id`"""
-        return match_with_same_doc_id[0][self.COL_MATCH_PARENT_ID]
+        return match_with_same_doc_id[0][self.COL_PARENT_ID]
 
 
 class Match2DocRanker(BaseRanker):
@@ -124,7 +118,7 @@ class Match2DocRanker(BaseRanker):
         - BucketShuffleRanker (first buckets matches and then sort each bucket).
     """
 
-    COL_MATCH_ID = 'match_id'
+    COL_MATCH_ID = 'match_doc_chunk_id'
     COL_SCORE = 'score'
 
     def score(self, query_meta: Dict, old_match_scores: Dict, match_meta: Dict) -> 'np.ndarray':

tests/unit/drivers/rank/aggregate/test_aggregate_matches_rank_driver.py

@@ -31,7 +31,7 @@ def __init__(self, *args, **kwargs):
         self.required_keys = {'length'}
 
     def _get_score(self, match_idx, query_chunk_meta, match_chunk_meta, *args, **kwargs):
-        return match_idx[0][self.COL_MATCH_PARENT_ID], match_chunk_meta[match_idx[0][self.COL_MATCH_ID]]['length']
+        return match_idx[0][self.COL_PARENT_ID], match_chunk_meta[match_idx[0][self.COL_DOC_CHUNK_ID]]['length']
 
 
 def create_document_to_score_same_depth_level():

tests/unit/drivers/rank/aggregate/test_chunk2doc_rank_drivers.py

@@ -25,7 +25,7 @@ def __init__(self, *args, **kwargs):
         self.required_keys = {'length'}
 
     def _get_score(self, match_idx, query_chunk_meta, match_chunk_meta, *args, **kwargs):
-        return match_idx[0][self.COL_MATCH_PARENT_ID], match_chunk_meta[match_idx[0][self.COL_MATCH_ID]]['length']
+        return match_idx[0][self.COL_PARENT_ID], match_chunk_meta[match_idx[0][self.COL_DOC_CHUNK_ID]]['length']
 
 
 class SimpleChunk2DocRankDriver(Chunk2DocRankDriver):