updated code for release 0.0.7

src/pyjedai/block_building.py

@@ -13,7 +13,7 @@
 
 from .datamodel import Block, Data, PYJEDAIFeature
 from .utils import (are_matching, drop_big_blocks_by_size,
-                    drop_single_entity_blocks)
+                    drop_single_entity_blocks, get_blocks_cardinality)
 from .evaluation import Evaluation
 
 class AbstractBlockProcessing(PYJEDAIFeature):
@@ -27,7 +27,13 @@ def __init__(self):
         self.attributes_2: list
         self.num_of_blocks_dropped: int
         self.original_num_of_blocks: int
-
+        self.sum_of_sizes: int = 0
+        self.total_num_of_comparisons: int = 0
+        self.min_block_size: int = None
+        self.max_block_size: int = None
+        self.min_block_comparisons: int = None
+        self.max_block_comparisons: int = None
+
     def report(self) -> None:
         """Prints Block Building method configuration
         """
@@ -75,7 +81,8 @@ def evaluate(self,
                 id2 in entity_index and are_matching(entity_index, id1, id2):
                 true_positives += 1
 
-        eval_obj.calculate_scores(true_positives=true_positives)
+        total_matching_pairs = get_blocks_cardinality(eval_blocks, self.data.is_dirty_er)
+        eval_obj.calculate_scores(true_positives=true_positives, total_matching_pairs=total_matching_pairs)
         eval_result = eval_obj.report(self.method_configuration(),
                                 export_to_df,
                                 export_to_dict,
@@ -84,8 +91,7 @@ def evaluate(self,
         if with_stats:
             self.stats(eval_blocks)
         return eval_result
-
-
+
     def stats(self, blocks: dict) -> None:
         self.list_of_sizes = []
         self.entities_in_blocks = set()
@@ -120,7 +126,6 @@ def stats(self, blocks: dict) -> None:
         )
         print(u'\u2500' * 123)
 
-
 class AbstractBlockBuilding(AbstractBlockProcessing):
     """Abstract class for the block building method
     """
@@ -137,13 +142,7 @@ def __init__(self):
         self.attributes_2: list
         self.execution_time: float
         self.data: Data
-        self.sum_of_sizes: int = 0
         self.list_of_sizes: list = []
-        self.total_num_of_comparisons: int = 0
-        self.min_block_size: int = None
-        self.max_block_size: int = None
-        self.min_block_comparisons: int = None
-        self.max_block_comparisons: int = None
 
     def build_blocks(
             self,

src/pyjedai/block_cleaning.py

@@ -13,7 +13,7 @@
 
 from .block_building import AbstractBlockProcessing
 from .datamodel import Block, Data
-from .utils import create_entity_index, drop_single_entity_blocks
+from .utils import create_entity_index, drop_single_entity_blocks, java_math_round
 
 class AbstractBlockCleaning(AbstractBlockProcessing):
 
@@ -81,14 +81,13 @@ def process(
         )
         sorted_blocks = _sort_blocks_cardinality(blocks, self.data.is_dirty_er)
         self._progress_bar.update(1)
-        entity_index = create_entity_index(sorted_blocks, self.data.is_dirty_er)
+        self.entity_index = create_entity_index(sorted_blocks, self.data.is_dirty_er)
         self._progress_bar.update(1)
         filtered_blocks = {}
-        for entity_id, block_keys in entity_index.items():
+        for entity_id, block_keys in self.entity_index.items():
             # Create new blocks from the entity index
-            # print(list(block_keys[:int(round(self.ratio*len(block_keys)))]))
             block_keys = list(block_keys)
-            for key in list(block_keys[:int(round(self.ratio*len(block_keys)))]):
+            for key in list(block_keys[:java_math_round(self.ratio*float(len(block_keys)))]):
                 filtered_blocks.setdefault(key, Block())
                 # Entities ids start to 0 ... n-1 for 1st dataset
                 # and n ... m for 2nd dataset
@@ -97,17 +96,17 @@ def process(
         self._progress_bar.update(1)
         new_blocks = drop_single_entity_blocks(filtered_blocks, self.data.is_dirty_er)
         self._progress_bar.close()
+        self.num_of_blocks_dropped = len(blocks) - len(new_blocks)
         self.execution_time = time() - start_time
         self.blocks = new_blocks
-        
-        return new_blocks
+
+        return self.blocks
 
     def _configuration(self) -> dict:
         return {
             "Ratio" : self.ratio
         }
 
-
 class BlockPurging(AbstractBlockCleaning):
     """Discards the blocks exceeding a certain number of comparisons.
     """