From 2d9415cdd09fadb3c024c314db0684a14eeb3f98 Mon Sep 17 00:00:00 2001
From: Samkit Jain <15127115+samkit-jain@users.noreply.github.com>
Date: Sun, 29 Nov 2020 15:41:01 +0530
Subject: [PATCH] Recursively parse metadata values to handle nested
 `PDFObjRef` objects Fixes #316

---
 pdfplumber/pdf.py    | 13 ++-----------
 pdfplumber/utils.py  | 18 ++++++++++++++++++
 tests/test_issues.py |  2 +-
 3 files changed, 21 insertions(+), 12 deletions(-)

diff --git a/pdfplumber/pdf.py b/pdfplumber/pdf.py
index 06ffb9d2..a0964617 100644
--- a/pdfplumber/pdf.py
+++ b/pdfplumber/pdf.py
@@ -1,6 +1,6 @@
 from .container import Container
 from .page import Page
-from .utils import decode_text
+from .utils import resolve_and_decode
 
 import logging
 import pathlib
@@ -31,16 +31,7 @@ def __init__(self, stream, pages=None, laparams=None, precision=0.001, password=
             self.metadata.update(info)
         for k, v in self.metadata.items():
             try:
-                if hasattr(v, "resolve"):
-                    v = v.resolve()
-                if type(v) == list:
-                    self.metadata[k] = list(map(decode_text, v))
-                elif isinstance(v, PSLiteral):
-                    self.metadata[k] = decode_text(v.name)
-                elif isinstance(v, (str, bytes)):
-                    self.metadata[k] = decode_text(v)
-                else:
-                    self.metadata[k] = v
+                self.metadata[k] = resolve_and_decode(v)
             except Exception as e:
                 # This metadata value could not be parsed. Instead of failing the PDF read,
                 # treat it as a warning.
diff --git a/pdfplumber/utils.py b/pdfplumber/utils.py
index f71d3c6c..05a4fa56 100644
--- a/pdfplumber/utils.py
+++ b/pdfplumber/utils.py
@@ -78,6 +78,24 @@ def decode_text(s):
         return "".join(PDFDocEncoding[o] for o in ords)
 
 
+def resolve_and_decode(obj):
+    """Recursively resolve the metadata values."""
+    if hasattr(obj, "resolve"):
+        obj = obj.resolve()
+    if isinstance(obj, list):
+        return list(map(resolve_and_decode, obj))
+    elif isinstance(obj, PSLiteral):
+        return decode_text(obj.name)
+    elif isinstance(obj, (str, bytes)):
+        return decode_text(obj)
+    elif isinstance(obj, dict):
+        for k, v in obj.items():
+            obj[k] = resolve_and_decode(v)
+        return obj
+
+    return obj
+
+
 def decode_psl_list(_list):
     return [
         decode_text(value.name) if isinstance(value, PSLiteral) else value
diff --git a/tests/test_issues.py b/tests/test_issues.py
index 886cbfc8..5288d2b7 100644
--- a/tests/test_issues.py
+++ b/tests/test_issues.py
@@ -175,4 +175,4 @@ def test_issue_316(self):
         """
         path = os.path.join(HERE, "pdfs/issue-316-example.pdf")
         with pdfplumber.open(path) as pdf:
-            assert pdf.metadata
+            assert pdf.metadata["Changes"][0]["CreationDate"] == "D:20061207105020Z00'00'"