add match_tables function

python/docrecjson-utilities/utilities.py

+from typing import List, Dict, Optional
+
+from docrecjson.elements import Table
+
+import shapely.geometry as shapely
+from shapely.validation import make_valid
+
+
+def match_tables(tables_to_find_candidates: List[Table], matching_tables: List[Table]) -> Dict[Table, Table]:
+    """
+    :param: tables_to_find_candidates: these are the tables where the candidates are found from.
+    :param: matching_tables: These are the tables which are tried to match to the tables_to_find_candidates.
+    :return: a dictionary with each table in tables_to_find_candidates, with one corresponding entry if matched table
+            was found, else none.
+    """
+
+    table: Table
+    matched_tables: Dict[Table, Table] = {}
+    for table in tables_to_find_candidates:
+        table_gt_area: shapely.Polygon = shapely.Polygon(table.get_table_coordinates())
+        matching_table_with_highest_intersection: Optional[Table] = None
+        highest_intersection: float = 0
+        table_to_match: Table
+        for table_to_match in matching_tables:
+            table_prediction_area: shapely.Polygon = shapely.Polygon(table_to_match.get_table_coordinates())
+            if not table_prediction_area.is_valid:
+                table_prediction_area = make_valid(table_prediction_area)
+            if table_gt_area.intersects(table_prediction_area):
+                intersection = table_gt_area.intersection(table_prediction_area)
+                if intersection.area > highest_intersection:
+                    highest_intersection = intersection.area
+                    matching_table_with_highest_intersection = table_to_match
+
+        matching_tables.remove(matching_table_with_highest_intersection)
+        matched_tables[table] = matching_table_with_highest_intersection
+
+    return matched_tables

python/requirements.txt

 -e .
 -e shared-file-format/python
+
+shapely
+loguru
\ No newline at end of file