mirror of
https://github.com/safishamsi/graphify.git
synced 2026-10-25 08:11:54 +00:00
Rudy CelekliandClaude Opus 5.5
bf3aac0e28
extract_pdf_text wrapped the whole PdfReader loop in except Exception: return "", so a malformed or encrypted PDF looked exactly like a PDF with no text and dropped out of the graph with no trace. Keep returning "" (callers rely on it), but print a warning naming the file and the error first. Warn once per path, since detect() and the semantic pass both read the same file. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
118 lines
3.8 KiB
Python
118 lines
3.8 KiB
Python
"""Tests for PDF text extraction warning when pypdf is missing (#3702)."""
|
|
import sys
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
|
|
from graphify import detect
|
|
|
|
|
|
@pytest.fixture(autouse=True)
|
|
def _reset_warn_flag(monkeypatch):
|
|
# The warn-once flag is process-global; reset it per test so each test
|
|
# starts from an un-warned state.
|
|
monkeypatch.setattr(detect, "_pypdf_missing_warned", False)
|
|
|
|
|
|
def test_extract_pdf_text_missing_pypdf_warns(tmp_path: Path, monkeypatch, capsys):
|
|
pdf = tmp_path / "sample.pdf"
|
|
pdf.write_bytes(b"%PDF-1.4 dummy content")
|
|
|
|
# Simulate missing pypdf
|
|
monkeypatch.setitem(sys.modules, "pypdf", None)
|
|
|
|
result = detect.extract_pdf_text(pdf)
|
|
assert result == ""
|
|
|
|
captured = capsys.readouterr()
|
|
assert "PDF text extraction skipped: 'pypdf' is not installed." in captured.err
|
|
assert "uv tool install 'graphifyy[pdf]'" in captured.err
|
|
|
|
|
|
def test_extract_pdf_text_missing_pypdf_warns_once_across_many_pdfs(
|
|
tmp_path: Path, monkeypatch, capsys
|
|
):
|
|
"""A corpus of many PDFs with pypdf missing must produce exactly one warning,
|
|
not one per file."""
|
|
monkeypatch.setitem(sys.modules, "pypdf", None)
|
|
|
|
for i in range(5):
|
|
pdf = tmp_path / f"doc{i}.pdf"
|
|
pdf.write_bytes(b"%PDF-1.4 dummy content")
|
|
assert detect.extract_pdf_text(pdf) == ""
|
|
|
|
captured = capsys.readouterr()
|
|
assert captured.err.count("PDF text extraction skipped") == 1
|
|
|
|
|
|
@pytest.fixture
|
|
def _reset_failed_warned(monkeypatch):
|
|
monkeypatch.setattr(detect, "_pdf_failed_warned", set(), raising=False)
|
|
|
|
|
|
def test_extract_pdf_text_unreadable_pdf_warns(tmp_path: Path, capsys, _reset_failed_warned):
|
|
"""#4527: a PDF pypdf cannot read still returns "", but must say so on
|
|
stderr instead of looking exactly like a PDF with no text."""
|
|
pytest.importorskip("pypdf")
|
|
pdf = tmp_path / "broken.pdf"
|
|
pdf.write_bytes(b"this is not a pdf")
|
|
|
|
assert detect.extract_pdf_text(pdf) == ""
|
|
|
|
err = capsys.readouterr().err
|
|
assert "could not extract text from" in err
|
|
assert "broken.pdf" in err
|
|
|
|
|
|
def test_extract_pdf_text_encrypted_pdf_warns(tmp_path: Path, capsys, _reset_failed_warned):
|
|
"""An encrypted PDF opened without its password is a failed read, not an
|
|
empty document (#4527)."""
|
|
pypdf = pytest.importorskip("pypdf")
|
|
writer = pypdf.PdfWriter()
|
|
writer.add_blank_page(width=72, height=72)
|
|
writer.encrypt(user_password="secret", algorithm="RC4-128")
|
|
pdf = tmp_path / "locked.pdf"
|
|
with pdf.open("wb") as fh:
|
|
writer.write(fh)
|
|
|
|
assert detect.extract_pdf_text(pdf) == ""
|
|
|
|
err = capsys.readouterr().err
|
|
assert "could not extract text from" in err
|
|
assert "locked.pdf" in err
|
|
|
|
|
|
def test_extract_pdf_text_unreadable_pdf_warns_once_per_file(
|
|
tmp_path: Path, capsys, _reset_failed_warned
|
|
):
|
|
"""detect() and the semantic pass both read the same PDF; one bad file
|
|
must warn once, while a second bad file still gets its own warning."""
|
|
pytest.importorskip("pypdf")
|
|
first = tmp_path / "a.pdf"
|
|
second = tmp_path / "b.pdf"
|
|
first.write_bytes(b"garbage")
|
|
second.write_bytes(b"garbage")
|
|
|
|
for _ in range(3):
|
|
assert detect.extract_pdf_text(first) == ""
|
|
assert detect.extract_pdf_text(second) == ""
|
|
|
|
err = capsys.readouterr().err
|
|
assert err.count("could not extract text from") == 2
|
|
|
|
|
|
def test_extract_pdf_text_empty_but_valid_pdf_does_not_warn(
|
|
tmp_path: Path, capsys, _reset_failed_warned
|
|
):
|
|
"""A readable PDF with no text layer is not a failure and stays silent."""
|
|
pypdf = pytest.importorskip("pypdf")
|
|
writer = pypdf.PdfWriter()
|
|
writer.add_blank_page(width=72, height=72)
|
|
pdf = tmp_path / "blank.pdf"
|
|
with pdf.open("wb") as fh:
|
|
writer.write(fh)
|
|
|
|
assert detect.extract_pdf_text(pdf) == ""
|
|
|
|
assert "could not extract text from" not in capsys.readouterr().err
|