adding test for color extraction

libpdf/models/horizontal_box.py

@@ -17,6 +17,8 @@ class Char:  # pylint: disable=too-few-public-methods # simplicity is good.
     :vartype x1: float
     :ivar y1: distance from the bottom of the page to the upper edge of the character (greater than y0)
     :vartype y1: float
+    :ivar ncolor: non-stroking-color
+    :vartype ncolor: tuple of rgb or None
     """
 
     def __init__(
@@ -26,13 +28,15 @@ def __init__(
         y0: float = None,
         x1: float = None,
         y1: float = None,
+        ncolor: tuple = None,
     ):
         """Init the class with plain char of a character and its rectangular coordinates."""
         self.x0 = x0
         self.y0 = y0
         self.x1 = x1
         self.y1 = y1
         self.text = text
+        self.ncolor = ncolor
 
     def __repr__(self):
         """Make the text part of the repr for better debugging."""
@@ -63,13 +67,17 @@ def __init__(
         self.x1 = x1
         self.y1 = y1
         self.chars = chars
+        self.ncolor = None
         if self.chars:
             # Obtain the rectangle coordinates from a list of libpdf text objects
             self.x0 = min(text_obj.x0 for text_obj in self.chars)
             self.y0 = min(text_obj.y0 for text_obj in self.chars)
             self.x1 = max(text_obj.x1 for text_obj in self.chars)
             self.y1 = max(text_obj.y1 for text_obj in self.chars)
 
+            if all(x.ncolor == self.chars[0].ncolor for x in self.chars):
+                self.ncolor = self.chars[0].ncolor
+
     @property
     def text(self):
         """Return plain text."""
@@ -157,6 +165,15 @@ def text(self):
         """Return plain text."""
         return "\n".join([x.text for x in self.lines])
 
+    @property
+    def words(self):
+        """Return array of words"""
+        _array = []
+        for l in self.lines:
+            _array += l.words
+
+        return _array
+
     def __repr__(self):
         """Make the text part of the repr for better debugging."""
         if self.lines:

libpdf/utils.py

@@ -458,7 +458,14 @@ def assemble_to_textlines(
     for lt_obj in flatten_lt_objs:
         if lt_obj.get_text() != " " and lt_obj.get_text() != "\n":
             # instantiate Char
-            char = Char(lt_obj.get_text(), lt_obj.x0, lt_obj.y0, lt_obj.x1, lt_obj.y1)
+            char = Char(
+                lt_obj.get_text(),
+                lt_obj.x0,
+                lt_obj.y0,
+                lt_obj.x1,
+                lt_obj.y1,
+                lt_obj.graphicstate.ncolor if hasattr(lt_obj, "graphicstate") else None,
+            )
             chars.append(char)
 
             if lt_obj is flatten_lt_objs[-1]:

tests/test_word_colors.py

@@ -0,0 +1,22 @@
+"""Test catalog extraction."""
+
+import logging
+
+import libpdf
+from tests.conftest import PDF_OUTLINE_NO_DEST
+
+
+def test_word_color_chapter() -> None:
+    """Test word colors in given chapter paragraph."""
+    objects = libpdf.load(PDF_OUTLINE_NO_DEST)
+    assert objects is not None
+    assert objects.flattened.chapters
+
+    for chapter in objects.flattened.chapters:
+        if chapter.title == "Create Basic Shapes":
+            for content in chapter.content:
+                if content.type == "paragraph" and "Diamond" in content.textbox.text:
+                    words = content.textbox.words
+                    logging.debug("found words ", words)
+                    for word in words:
+                        assert word.ncolor == (0, 0, 1)