From abf2e7e9bb2a62052ae6531100169f3df91f84ef Mon Sep 17 00:00:00 2001
From: Jim Barlow <jim@purplerock.ca>
Date: Sat, 25 Jul 2015 15:31:02 -0700
Subject: [PATCH] Copy document metadata from source document into output
 (untested)

This works for ASCII only; will do Unicode version.
---
 src/ocrmypdf.py | 13 ++++++++++++-
 src/pdfa.py     | 16 +++++++++++-----
 2 files changed, 23 insertions(+), 6 deletions(-)

diff --git a/src/ocrmypdf.py b/src/ocrmypdf.py
index 0fe71f41..be5a35e1 100755
--- a/src/ocrmypdf.py
+++ b/src/ocrmypdf.py
@@ -596,7 +596,18 @@ def generate_postscript_stub(
         input_file,
         output_file,
         log):
-    generate_pdfa_def(output_file)
+    try:
+        pdf = pypdf.PdfFileReader(input_file)
+        pdfmark = {
+            'title': pdf.documentInfo['/Title'],
+            'author': pdf.documentInfo['/Author'],
+            'keywords': pdf.documentInfo['/Keywords'],
+            'subject': pdf.documentInfo['/Subject']
+        }
+    except KeyError:
+        pdfmark = {}
+
+    generate_pdfa_def(output_file, pdfmark)
 
 
 @transform(
diff --git a/src/pdfa.py b/src/pdfa.py
index b4069a22..cdd56356 100644
--- a/src/pdfa.py
+++ b/src/pdfa.py
@@ -25,6 +25,9 @@ pdfa_def_template = u"""%!
 def
 
 [ /Title ($pdf_title)
+  /Author ($pdf_author)
+  /Subject ($pdf_subject)
+  /Keywords ($pdf_keywords)
   /DOCINFO pdfmark
 
 % Define an ICC profile :
@@ -57,11 +60,14 @@ def
 """
 
 
-def _get_pdfa_def(icc_profile, pdf_title, icc_identifier):
+def _get_pdfa_def(icc_profile, icc_identifier, pdfmark):
     t = Template(pdfa_def_template)
     result = t.substitute(icc_profile=icc_profile,
-                          pdf_title=pdf_title,
-                          icc_identifier=icc_identifier)
+                          icc_identifier=icc_identifier,
+                          pdf_title=pdfmark.get('title', ''),
+                          pdf_author=pdfmark.get('author', ''),
+                          pdf_subject=pdfmark.get('subject', ''),
+                          pdf_keywords=pdfmark.get('keywords', ''))
     return result
 
 
@@ -91,13 +97,13 @@ def _get_postscript_icc_path():
             return path
 
 
-def generate_pdfa_def(target_filename, pdf_title='', icc='sRGB'):
+def generate_pdfa_def(target_filename, pdfmark, icc='sRGB'):
     if icc == 'sRGB':
         icc_profile = os.path.join(_get_postscript_icc_path(), 'srgb.icc')
     else:
         raise NotImplementedError("Only supporting sRGB")
 
-    ps = _get_pdfa_def(icc_profile, pdf_title, icc)
+    ps = _get_pdfa_def(icc_profile, icc, pdfmark)
 
     # Since PostScript might not handle UTF-8 (it's hard to get a clear
     # answer), insist on ascii