Fix pypdf bug: AttributeError when /DR missing from AcroForm

[jhs]

Bug Report for Upstream pypdf Project

Purpose: This issue tracks the need to report and fix a bug in the pypdf library (py-pdf/pypdf).

---

Summary

pypdf crashes with AttributeError: 'dict' object has no attribute 'get_object' when filling form fields in PDFs that lack a /DR (Default Resources) entry in the AcroForm dictionary and reference fonts not in CORE_FONT_METRICS.

Location

• File: pypdf/generic/_appearance_stream.py
• Line: 437
• Function: TextStreamAppearance.from_text_annotation()
• Version: pypdf 5.7.0+ (tested with 6.3.0+)

Root Cause

Lines 433-437 contain a type error:

# BUGGY CODE
document_resources = cast(
    dict[Any, Any],
    acro_form.get("/DR", {}),  # ← Returns plain dict {} if /DR missing
)
document_font_resources = document_resources.get_object().get("/Font", DictionaryObject()).get_object()

The Problem:

1. cast() is only a type hint - it doesn't convert the object
2. acro_form.get("/DR", {}) returns:
   • A pypdf DictionaryObject if /DR exists (has .get_object() method)
   • A plain Python dict {} if /DR is missing (no .get_object() method)
3. Line 437 unconditionally calls .get_object(), which crashes when the default {} is returned

Correct Pattern

The same file already shows the correct pattern on lines 419-428:

# CORRECT CODE (lines 419-428)
document_resources = cast(
    DictionaryObject,
    cast(
        DictionaryObject,
        acro_form.get("/DR", DictionaryObject()),  # ← Uses DictionaryObject() as default
    ).get_object(),  # ← Calls get_object() immediately
)

Proposed Fix

Replace lines 433-437 with:

document_resources = cast(
    DictionaryObject,
    cast(
        DictionaryObject,
        acro_form.get("/DR", DictionaryObject()),  # ← Change {} to DictionaryObject()
    ).get_object(),
)
document_font_resources = document_resources.get("/Font", DictionaryObject()).get_object()

Minimal Complete Verifiable Example (MCVE)

Reproduction Code

from pypdf import PdfReader, PdfWriter

# Create a PDF with form fields using PyMuPDF
import fitz

doc = fitz.new()
page = doc.new_page(width=612, height=792)

# Add a text field with /Helv font (not in CORE_FONT_METRICS)
widget = fitz.Widget()
widget.field_type = fitz.PDF_WIDGET_TYPE_TEXT
widget.field_name = "test_field"
widget.rect = fitz.Rect(50, 50, 300, 70)
widget.text_fontsize = 10
widget.text_font = "helv"  # This becomes /Helv in default appearance
widget.fill_color = (1, 1, 1)
widget.border_color = (0.7, 0.7, 0.7)
page.add_widget(widget)

doc.save("test_form.pdf")
doc.close()

# Now try to fill the field with pypdf
reader = PdfReader("test_form.pdf")
writer = PdfWriter(clone_from=reader)

# This will crash with AttributeError
writer.update_page_form_field_values(
    writer.pages[0],
    {"test_field": "Test Value"},
    auto_regenerate=False
)

writer.set_need_appearances_writer(True)
with open("test_form_filled.pdf", "wb") as f:
    writer.write(f)

Error Output

Traceback (most recent call last):
  File "test.py", line 29, in <module>
    writer.update_page_form_field_values(
  File "/path/to/pypdf/_writer.py", line 1045, in update_page_form_field_values
    appearance_stream_obj = TextStreamAppearance.from_text_annotation(
                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/path/to/pypdf/generic/_appearance_stream.py", line 437, in from_text_annotation
    document_font_resources = document_resources.get_object().get("/Font", DictionaryObject()).get_object()
                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'dict' object has no attribute 'get_object'

PDF Characteristics

The bug occurs when:

1. AcroForm dictionary has no /DR (Default Resources) entry
2. Form field default appearance references a font not in CORE_FONT_METRICS (e.g., /Helv)
3. Code reaches the fallback path at lines 431-437

CORE_FONT_METRICS includes: Helvetica, Courier, Times-Roman, Arial, etc. (but NOT "Helv")

Workaround

Until fixed, use fonts that are in CORE_FONT_METRICS:

• Change /Helv to /Helvetica in PDF default appearance strings
• Or add a /DR dictionary to the AcroForm with proper font resources

Testing Strategy

A proper unit test should:

1. Create a PDF with form fields lacking /DR in AcroForm
2. Use a font reference not in CORE_FONT_METRICS (e.g., /Helv)
3. Attempt to update field values using update_page_form_field_values()
4. Verify it either succeeds or fails gracefully (not with AttributeError)

Labels for pypdf Issue

When reporting to pypdf, use labels:

• is-bug
• workflow-forms
• needs-pdf (will attach sample PDF)

---

Action Items

• Create minimal reproducing PDF
• Open issue on https://github.com/py-pdf/pypdf/issues
• Submit PR with fix and unit tests
• Link back to this issue

[jhs]

Prompt for Claude Code Web to Fix This Bug

Copy and paste this prompt into Claude Code Web when ready to work on the fix:

---

Task: Fix a bug in the pypdf library where AttributeError: 'dict' object has no attribute 'get_object' occurs when filling PDF form fields.

Steps to complete:

1. Clone and Setup

   • Fork https://github.com/py-pdf/pypdf to your GitHub account
   • Clone your fork locally
   • Set up the development environment according to their CONTRIBUTING.md
   • Install dependencies and verify tests run successfully

2. Confirm the Bug

   • Navigate to pypdf/generic/_appearance_stream.py line 437
   • Identify the buggy code:

     document_resources = cast(
         dict[Any, Any],
         acro_form.get("/DR", {}),  # ← Problem: returns plain dict
     )
     document_font_resources = document_resources.get_object().get("/Font", DictionaryObject()).get_object()

   • Confirm that when acro_form.get("/DR", {}) returns the default {}, calling .get_object() on it will fail

3. Understand the Test Suite

   • Explore the existing test structure in the tests/ directory
   • Look for tests related to form filling, specifically update_page_form_field_values()
   • Identify the testing conventions:
     • File naming patterns
     • Test organization (pytest style)
     • How test PDFs are handled (fixtures, sample files, etc.)
     • Assertion patterns and error checking

4. Create a Failing Unit Test

   • Create a test that reproduces this specific bug
   • The test should:
     • Create or use a PDF with form fields
     • Ensure the PDF's AcroForm lacks a /DR entry
     • Use a font not in CORE_FONT_METRICS (like /Helv)
     • Call update_page_form_field_values() to trigger the bug
     • Currently: This should raise AttributeError: 'dict' object has no attribute 'get_object'
   • Add this test to the appropriate test file following pypdf conventions
   • Run the test suite to confirm it fails with the expected error

5. Fix the Bug

   • Modify lines 433-437 in pypdf/generic/_appearance_stream.py
   • Change from:

     document_resources = cast(
         dict[Any, Any],
         acro_form.get("/DR", {}),
     )
     document_font_resources = document_resources.get_object().get("/Font", DictionaryObject()).get_object()

   • To match the correct pattern from lines 419-428:

     document_resources = cast(
         DictionaryObject,
         cast(
             DictionaryObject,
             acro_form.get("/DR", DictionaryObject()),  # ← Use DictionaryObject() not {}
         ).get_object(),
     )
     document_font_resources = document_resources.get("/Font", DictionaryObject()).get_object()

   • The key change: Replace the default {} with DictionaryObject()

6. Verify the Fix

   • Run the new unit test - it should now pass
   • Run the entire test suite to ensure no regressions
   • Verify type checking passes (if pypdf uses mypy or similar)
   • Test manually with the MCVE from the issue description

7. Create Branch and Commit

   • Create a new branch: git checkout -b fix/acroform-dr-missing-attributeerror
   • Commit changes with a descriptive message:

     Fix AttributeError when AcroForm lacks /DR entry
     
     When filling form fields in PDFs without a /DR (Default Resources)
     entry in the AcroForm dictionary, pypdf would crash with:
     AttributeError: 'dict' object has no attribute 'get_object'
     
     The bug was in _appearance_stream.py line 437 where it used an
     empty dict {} as the default value instead of DictionaryObject().
     The cast() type hint doesn't convert the object, so when acro_form.get()
     returns the default {}, calling .get_object() fails.
     
     Fixed by using DictionaryObject() as default, matching the pattern
     already used correctly on lines 419-428 in the same function.
     
     Added unit test to prevent regression.
     

   • Push the branch to your fork: git push origin fix/acroform-dr-missing-attributeerror

8. Report Results

   • Provide the branch URL
   • Show the output of the failing test (before fix)
   • Show the output of the passing test (after fix)
   • Confirm all existing tests still pass
   • Note the test file location and test function name

Expected Outcome:

• A branch in your pypdf fork with the fix
• A new unit test that failed before the fix and passes after
• All existing tests passing
• Ready to submit as a PR to py-pdf/pypdf

Important Notes:

• Follow pypdf's coding style and conventions
• Make sure the fix handles edge cases (empty DictionaryObject, nested references, etc.)
• The fix should be minimal and focused on this specific bug
• Include comments explaining why DictionaryObject() is needed instead of {}

---

When you complete this task, provide:

1. Link to your branch
2. Output showing the failing test before the fix
3. Output showing the passing test after the fix
4. Confirmation that the full test suite passes
5. Any observations about the pypdf codebase or testing practices