Close #LGVISIUM-77: Wrong elevation value selected within bounding box

src/stratigraphy/data_extractor/data_extractor.py

@@ -130,6 +130,11 @@ def get_lines_near_key(self, lines, key_line: TextLine) -> list[TextLine]:
         )
         feature_lines = [line for line in lines if line.rect.intersects(elevation_search_rect)]
 
-        # makes sure the line with the key is included first in the extracted information and the duplicate removed
+        # Insert key_line first and remove duplicates
         feature_lines.insert(0, key_line)
-        return list(dict.fromkeys(feature_lines))
+        feature_lines = list(dict.fromkeys(feature_lines))
+
+        # Sort by vertical distance between the top of the feature line and the top of key_line
+        feature_lines_sorted = sorted(feature_lines, key=lambda line: abs(line.rect.y0 - key_line.rect.y0))
+
+        return feature_lines_sorted

src/stratigraphy/groundwater/utility.py

@@ -66,7 +66,7 @@ def extract_elevation(text: str) -> float | None:
     elevation_patterns = [
         r"(\d+(\.\d+)?)\s*m\s*u\.m\.",
         r"(\d+(\.\d+)?)\s*m\s*ur.",
-        r"(\d{3,}\.\d+)",
+        r"(\d{3,}\.\d{1,2})(?!\d)",  # Matches a float number with less than 2 digits after the decimal point
         r"(\d{3,})\s*m",
     ]
 

src/stratigraphy/metadata/elevation_extraction.py

@@ -89,35 +89,35 @@ def get_elevation_near_key(self, lines: list[TextLine], page: int) -> Elevation
         """
         # find the key that indicates the elevation information
         elevation_key_lines = self.find_feature_key(lines)
-        extracted_elevation_informations = []
+        extracted_elevation_list = []
 
         for elevation_key_line in elevation_key_lines:
-            elevation_lines = self.get_lines_near_key(lines, elevation_key_line)
+            elevation_lines = self.get_lines_near_key(lines, elevation_key_line)  # Check the sorting of the lines
 
             try:
-                extracted_elevation_information = self.get_elevation_from_lines(elevation_lines, page)
-                if extracted_elevation_information.elevation:
-                    extracted_elevation_informations.append(extracted_elevation_information)
+                extracted_elevation = self.get_elevation_from_lines(elevation_lines, page)
+                if extracted_elevation.elevation:
+                    extracted_elevation_list.append(extracted_elevation)
             except ValueError as error:
                 logger.warning("ValueError: %s", error)
                 logger.warning("Could not extract all required information from the lines provided.")
 
-        return self.select_best_elevation_information(extracted_elevation_informations)
+        return self.select_best_elevation_information(extracted_elevation_list)
 
-    def select_best_elevation_information(self, extracted_elevation_informations: list[Elevation]) -> Elevation | None:
+    def select_best_elevation_information(self, extracted_elevation_list: list[Elevation]) -> Elevation | None:
         """Select the best elevation information from a list of extracted elevation information.
 
         Args:
-            extracted_elevation_informations (list[Elevation]): A list of extracted elevation information.
+            extracted_elevation_list (list[Elevation]): A list of extracted elevation information.
 
         Returns:
             Elevation | None: The best extracted elevation information.
         """
         # Sort the extracted elevation information by elevation with the highest elevation first
-        extracted_elevation_informations.sort(key=lambda x: x.elevation, reverse=True)
+        extracted_elevation_list.sort(key=lambda x: x.elevation, reverse=True)
 
         # Return the first element of the sorted list
-        return extracted_elevation_informations[0] if extracted_elevation_informations else None
+        return extracted_elevation_list[0] if extracted_elevation_list else None
 
     def get_elevation_from_lines(self, lines: list[TextLine], page: int) -> Elevation:
         r"""Matches the elevation in a string of text.