Update table.py

Author: JorjMcKie

src/table.py

@@ -110,6 +110,29 @@
 white_spaces = set(string.whitespace)  # for checking white space only cells
 
 
+def rect_in_rect(inner, outer):
+    """Check whether rectangle 'inner' is fully inside rectangle 'outer'."""
+    return (
+        1
+        and inner[0] >= outer[0]
+        and inner[1] >= outer[1]
+        and inner[2] <= outer[2]
+        and inner[3] <= outer[3]
+    )
+
+
+def chars_in_rect(CHARS, rect):
+    """Check whether any of the chars in CHAR are inside rectangle 'rect'."""
+    return any(
+        1
+        and rect[0] <= c["x0"]
+        and c["x1"] <= rect[2]
+        and rect[1] <= c["y0"]
+        and rect[3] >= c["y1"]
+        for c in CHARS
+    )
+
+
 def _iou(r1, r2):
     """Compute intersection over union of two rectangles."""
     ix = max(0, min(r1[2], r2[2]) - max(r1[0], r2[0]))
@@ -126,7 +149,7 @@ def intersects_words_h(bbox, y, word_rects) -> bool:
     """Check whether any of the words in bbox are cut through by
     horizontal line y.
     """
-    return any(r.y0 < y < r.y1 for r in word_rects if r in bbox)
+    return any(r.y0 < y < r.y1 for r in word_rects if rect_in_rect(r, bbox))
 
 
 def get_table_dict_from_rect(textpage, rect):
@@ -182,7 +205,9 @@ def make_table_from_bbox(textpage, word_rects, rect):
     for i in range(len(nypos) - 1):
         row_box = pymupdf.Rect(bbox.x0, nypos[i], bbox.x1, nypos[i + 1])
         # Sub-select words in this row and sort them by left coordinate
-        row_words = sorted([r for r in word_rects if r in row_box], key=lambda r: r.x0)
+        row_words = sorted(
+            [r for r in word_rects if rect_in_rect(r, row_box)], key=lambda r: r.x0
+        )
         # Sub-select x values that do not cut through words
         this_xpos = [x for x in nxpos if not any(r.x0 < x < r.x1 for r in row_words)]
         for j in range(len(this_xpos) - 1):
@@ -1721,12 +1746,12 @@ def row_has_bold(bbox):
 
             Returns True if any spans are bold else False.
             """
-            blocks = page.get_text("dict", flags=pymupdf.TEXTFLAGS_TEXT, clip=bbox)[
-                "blocks"
-            ]
-            spans = [s for b in blocks for l in b["lines"] for s in l["spans"]]
-
-            return any(s["flags"] & pymupdf.TEXT_FONT_BOLD for s in spans)
+            global CHARS
+            return any(
+                c["bold"]
+                for c in CHARS
+                if rect_in_rect((c["x0"], c["y0"], c["x1"], c["y1"]), bbox)
+            )
 
         try:
             row = self.rows[0]
@@ -2171,6 +2196,9 @@ def make_chars(page, clip=None):
             for span in sorted(line["spans"], key=lambda s: s["bbox"][0]):
                 fontname = span["font"]
                 fontsize = span["size"]
+                span_bold = bool(
+                    span["flags"] & pymupdf.TEXT_FONT_BOLD or span["char_flags"] & 8
+                )
                 color = pymupdf.sRGB_to_pdf(span["color"])
                 for char in sorted(span["chars"], key=lambda c: c["bbox"][0]):
                     bbox = pymupdf.Rect(char["bbox"])
@@ -2194,6 +2222,7 @@ def make_chars(page, clip=None):
                         "size": fontsize if upright else bbox.y1 - bbox.y0,
                         "stroking_color": color,
                         "stroking_pattern": None,
+                        "bold": span_bold,
                         "text": text,
                         "top": bbox.y0,
                         "upright": upright,
@@ -2212,6 +2241,7 @@ def make_chars(page, clip=None):
 # else to lines.
 # ------------------------------------------------------------------------
 def make_edges(page, clip=None, tset=None, paths=None, add_lines=None, add_boxes=None):
+    global CHARS
     snap_x = tset.snap_x_tolerance
     snap_y = tset.snap_y_tolerance
     min_length = tset.edge_min_length
@@ -2303,7 +2333,7 @@ def clean_graphics(npaths=None):
                         repeat = True  # keep checking the rest
 
             # move rect 0 over to result list if there is some text in it
-            if not white_spaces.issuperset(page.get_textbox(prect0, textpage=TEXTPAGE)):
+            if chars_in_rect(CHARS, prect0):
                 # contains text, so accept it as a table bbox candidate
                 new_rects.append(prect0)
             del prects[0]  # remove from rect list