🐛 fix potential problem where the words can be re-arranged by the user when calculating lines.

Author: ianardee

mindee/fields/ocr.py

@@ -41,14 +41,17 @@ def __str__(self) -> str:
 class OcrPage:
     """OCR extraction for a single page."""
 
-    all_words: List[OcrWord]
-    """All the words on the page, in semi-random order."""
+    _all_words: List[OcrWord]
     _lines: List[OcrLine]
 
     def __init__(self, prediction: TypeApiPrediction):
-        self.all_words = [
+        self._all_words = [
             OcrWord(word_prediction) for word_prediction in prediction["all_words"]
         ]
+        # make sure words are sorted from top to bottom
+        self._all_words.sort(
+            key=lambda item: get_min_max_y(item.polygon).min, reverse=False
+        )
         self._lines = []
 
     @staticmethod
@@ -70,14 +73,9 @@ def _to_lines(self) -> List[OcrLine]:
         indexes: List[int] = []
         lines: List[OcrLine] = []
 
-        # make sure words are sorted from top to bottom
-        self.all_words.sort(
-            key=lambda item: get_min_max_y(item.polygon).min, reverse=False
-        )
-
-        for _ in self.all_words:
+        for _ in self._all_words:
             line: OcrLine = OcrLine()
-            for idx, word in enumerate(self.all_words):
+            for idx, word in enumerate(self._all_words):
                 if idx in indexes:
                     continue
                 if current is None:
@@ -102,6 +100,11 @@ def all_lines(self) -> List[OcrLine]:
             self._lines = self._to_lines()
         return self._lines
 
+    @property
+    def all_words(self) -> List[OcrWord]:
+        """All the words on the page, in semi-random order."""
+        return self._all_words
+
     def __str__(self) -> str:
         return "\n".join(str(line) for line in self.all_lines) + "\n"
 

mindee/geometry.py

@@ -208,7 +208,7 @@ def get_min_max_x(points: Points) -> MinMax:
 
 def is_point_in_polygon_x(point: Point, polygon: Polygon) -> bool:
     """
-    Determine if the Point is in the Polygon's Y-axis.
+    Determine if the Point is in the Polygon's X-axis.
 
     :param point: Point to compare
     :param polygon: Polygon to look into
@@ -219,7 +219,7 @@ def is_point_in_polygon_x(point: Point, polygon: Polygon) -> bool:
 
 def is_point_in_x(point: Point, min_x: float, max_x: float) -> bool:
     """
-    Determine if the Point is in the Polygon's Y-axis.
+    Determine if the Point is within the X-axis interval.
 
     :param point: Point to compare
     :param min_x: Minimum X-axis value