FIXUP: Make tm_prev a tm_matrix, not a product of tm*cm

py-pdf · Aug 6, 2023 · 884055d · 884055d
1 parent 9e6512c
commit 884055d
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 4 deletions.
diff --git a/pypdf/_page.py b/pypdf/_page.py
@@ -1926,7 +1926,7 @@ def _extract_text(
  1.0,
  0.0,
  0.0,
- ] # will store cm_matrix * tm_matrix
+ ] # will store previous tm_matrix
  char_scale = 1.0
  space_scale = 1.0
  _space_width: float = 500.0 # will be set correctly at first Tf

diff --git a/pypdf/_text_extraction/__init__.py b/pypdf/_text_extraction/__init__.py
@@ -99,10 +99,11 @@ def crlf_space_check(
  visitor_text: Optional[Callable[[Any, Any, Any, Any, Any], None]],
  spacewidth: float,
 ) -> Tuple[str, str, List[float]]:
+ m_prev = mult(tm_prev, cm_matrix)
  m = mult(tm_matrix, cm_matrix)
  orientation = orient(m)
- delta_x = m[4] - tm_prev[4]
- delta_y = m[5] - tm_prev[5]
+ delta_x = m[4] - m_prev[4]
+ delta_y = m[5] - m_prev[5]
  k = math.sqrt(abs(m[0] * m[3]) + abs(m[1] * m[2]))
  f = font_size * k
  if orientation not in orientations:
@@ -186,7 +187,7 @@ def crlf_space_check(
  text += " "
  except Exception:
  pass
- tm_prev = m
+ tm_prev = tm_matrix.copy()
  return text, output, tm_prev