From 6a4df78bc060b0fcedc713d400fda788816f8242 Mon Sep 17 00:00:00 2001
From: "James R. Barlow" <jim@purplerock.ca>
Date: Tue, 27 Mar 2018 13:32:38 -0700
Subject: [PATCH] Add _naive_find_text to search for text when fitz is not
 available

---
 src/ocrmypdf/pdfinfo.py | 37 ++++++++++++++++++++++++++-----------
 tests/test_pageinfo.py  |  8 ++++++++
 2 files changed, 34 insertions(+), 11 deletions(-)

diff --git a/src/ocrmypdf/pdfinfo.py b/src/ocrmypdf/pdfinfo.py
index c18d05dd..a5c9d50d 100644
--- a/src/ocrmypdf/pdfinfo.py
+++ b/src/ocrmypdf/pdfinfo.py
@@ -554,20 +554,35 @@ def _find_images(*, pdf, container, shorthand=None):
     yield from _find_form_xobject_images(pdf, container, contentsinfo)
 
 
-def _naive_find_text(*, pdf, container):
-    if container.get('/Type') == '/Page' and '/Contents' in container:
-        # For a /Page the content stream is attached to the page's /Contents
-        page = container
-        contentstream = pypdf.pdf.ContentStream(page.getContents(), pdf)
-    elif container.get('/Type') == '/XObject' and \
-            container['/Subtype'] == '/Form':
-        # For a Form XObject that content stream is attached to the XObject
-        contentstream = pypdf.pdf.ContentStream(container, pdf)
+def _naive_find_text(*, pdf, page):
+    if not(page.get('/Type') == '/Page' and '/Contents' in page):
+        # Not a page, or has no /Contents => no text
+        return False
 
+    # First we check the main content stream    
+    contentstream = pypdf.pdf.ContentStream(page.getContents(), pdf)
     contentsinfo = _interpret_contents(contentstream, UNIT_SQUARE)
-
     if contentsinfo.found_text:
         return True
+
+    # Then see if there is a Form XObject with with a content stream
+    # that might have text.  For full completeness we should recursively
+    # search nested Form XObjects, as we do with images.  But that is
+    # rare.
+    if '/Resources' in page:
+        resources = page['/Resources']
+        if '/XObject' in resources:    
+            for xobj in resources['/XObject']:
+                candidate = resources['/XObject'][xobj]
+                if candidate['/Subtype'] != '/Form':
+                    continue
+                form_xobject = candidate                
+                # Content stream is attached to Form XObject dictionary
+                contentstream = pypdf.pdf.ContentStream(form_xobject, pdf)
+                sub_contentsinfo = _interpret_contents(
+                    contentstream, UNIT_SQUARE)
+                if sub_contentsinfo.found_text:
+                    return True
     return False
 
 
@@ -594,7 +609,7 @@ def _pdf_get_pageinfo(pdf, pageno: int, infile):
     if fitz:
         pageinfo['has_text'] = _page_has_text(str(infile), pageno)
     else:
-        pageinfo['has_text'] = _naive_find_text(pdf=pdf, container=page)
+        pageinfo['has_text'] = _naive_find_text(pdf=pdf, page=page)
 
     width_pt = page.mediaBox.getWidth()
     height_pt = page.mediaBox.getHeight()
diff --git a/tests/test_pageinfo.py b/tests/test_pageinfo.py
index 2ef27428..e421ed52 100644
--- a/tests/test_pageinfo.py
+++ b/tests/test_pageinfo.py
@@ -28,6 +28,7 @@ import pytest
 import img2pdf
 import pytest
 import sys
+import PyPDF2 as pypdf
 
 
 def test_single_page_text(outdir):
@@ -125,6 +126,13 @@ def test_form_xobject(resources):
     assert pdfimage.width == 50
 
 
+def test_naive_find_text(resources):
+    filename = resources / 'formxobject.pdf'
+    reader = pypdf.PdfFileReader(str(filename))
+    page = reader.getPage(0)
+    assert pdfinfo._naive_find_text(pdf=reader, page=page)
+
+
 def test_no_contents(resources):
     filename = resources / 'no_contents.pdf'