Merge pull request #42 from gnes-ai/fix_weighting

fix(weighting): add simple normalization to chunk search

gnes/preprocessor/image/segmentation.py

@@ -68,7 +68,7 @@ def apply(self, doc: 'gnes_pb2.Document'):
                 c.doc_id = doc.doc_id
                 c.blob.CopyFrom(array2blob(self._crop_image_reshape(original_image, ele[0])))
                 c.offset_1d = ci
-                c.weight = ele[1]
+                c.weight = self._cal_area(ele[0]) / (original_image.size[0] * original_image.size[1])
 
             c = doc.chunks.add()
             c.doc_id = doc.doc_id
@@ -82,3 +82,7 @@ def apply(self, doc: 'gnes_pb2.Document'):
     def _crop_image_reshape(self, original_image, coordinates):
         return np.array(original_image.crop(coordinates).resize((self.target_img_size,
                                                                  self.target_img_size)))
+
+    @classmethod
+    def _cal_area(cls, coordinate):
+        return (coordinate[2] - coordinate[0]) * (coordinate[3] - coordinate[1])

gnes/service/indexer.py

@@ -66,7 +66,7 @@ def _handler_chunk_search(self, msg: 'gnes_pb2.Message'):
                 r.chunk.doc_id = _doc_id
                 r.chunk.offset_1d = _offset
                 r.chunk.weight = _weight
-                r.score = _weight * _relevance * qc_weight
+                r.score = _weight * qc_weight * (-1 / _relevance)
                 r.score_explained = '[chunk_score at doc: %d, offset: %d] = ' \
                                     '(doc_chunk_weight: %.6f) * ' \
                                     '(query_doc_chunk_relevance: %.6f) * ' \