Reorganize around getting bboxes for visible/invisible text

src/ocrmypdf/pdfinfo/__init__.py

@@ -28,6 +28,7 @@ import xml.etree.ElementTree as ET
 from pikepdf import PdfMatrix
 import pikepdf
 
+from .layout import get_textblocks, bboxes
 from ..exec import ghostscript
 from ..helpers import fspath
 
@@ -597,7 +598,6 @@ def _pdf_get_pageinfo(pdf, pageno: int, infile, xmltext):
     # pageinfo['textinfo'] = _page_get_textblocks(
     #     fspath(infile), pageno, xmltext=xmltext)
 
-    from .layout import get_textblocks
     with Path(infile).open('rb') as f:
         pageinfo['objects'] = get_textblocks(f, pageno)
 
@@ -605,17 +605,6 @@ def _pdf_get_pageinfo(pdf, pageno: int, infile, xmltext):
     width_pt = mediabox[2] - mediabox[0]
     height_pt = mediabox[3] - mediabox[1]
 
-    def bboxes(hierarchical_textinfo):
-        from pdfminer.layout import LTTextLine, LTTextBox
-        for obj in hierarchical_textinfo:
-            if isinstance(hierarchical_textinfo, (LTTextLine, LTTextBox)):
-                yield hierarchical_textinfo.bbox
-            else:
-                try:
-                    yield from bboxes(obj)
-                except TypeError:
-                    continue
-
     pageinfo['has_text'] = _page_has_text(
         bboxes(pageinfo['objects']), width_pt, height_pt)
 
@@ -748,17 +737,6 @@ class PageInfo:
         return self._pageinfo['images']
 
     def get_textareas(self, visible=True, invisible=True):
-        def bboxes(objs):
-            from pdfminer.layout import LTTextBox
-            for obj in objs:
-                if isinstance(obj, LTTextBox):
-                    yield obj.bbox
-                else:
-                    try:
-                        yield from bboxes(obj)
-                    except TypeError:
-                        continue
-
         if visible:
             yield from bboxes(self._pageinfo['objects'][0])
         if invisible:

src/ocrmypdf/pdfinfo/layout.py

@@ -23,7 +23,8 @@ from pdfminer.pdfpage import PDFPage
 from pdfminer.utils import matrix2str, bbox2str, fsplit
 from pdfminer.pdffont import PDFUnicodeNotDefined
 from pdfminer.layout import (
-    LTChar, LTContainer, LTLayoutContainer, LTPage, LTTextLine, LAParams
+    LTChar, LTContainer, LTLayoutContainer, LTPage, LTTextLine, LAParams,
+    LTTextBox
 )
 
 from pdfminer.converter import PDFLayoutAnalyzer
@@ -121,7 +122,7 @@ class TextPositionTracker(PDFLayoutAnalyzer):
         return item.adv
 
     def handle_undefined_char(self, font, cid):
-        log.info('undefined: %r, %r', font, cid)
+        #log.info('undefined: %r, %r', font, cid)
         return '(cid:%d)' % cid
 
     def receive_layout(self, ltpage):
@@ -141,3 +142,14 @@ def get_textblocks(infile, pageno):
     interp.process_page(next(page))
 
     return dev.get_result()
+
+
+def bboxes(hierarchical_textinfo):
+    for obj in hierarchical_textinfo:
+        if isinstance(hierarchical_textinfo, (LTTextBox)):
+            yield hierarchical_textinfo.bbox
+        else:
+            try:
+                yield from bboxes(obj)
+            except TypeError:
+                continue

tests/test_pdfinfo.py

@@ -145,7 +145,7 @@ def test_pickle(resources):
     # For multiprocessing we must be able to pickle our information - if
     # this fails then we are probably storing some unpickleabe pikepdf or
     # other external data around
-    filename = resources / 'formxobject.pdf'
+    filename = resources / 'graph_ocred.pdf'
     pdf = pdfinfo.PdfInfo(filename)
     pickle.dumps(pdf)