14e1c1a2ea
The NEEDS-OCR and EXTRACTION-COVERAGE hints hand-wrapped the path in single quotes, so a path containing an apostrophe (e.g. "training team's notebook.pdf") produced an unparseable command — the same slip this branch already fixes for the jq hint. Both hints now go through shlex.quote, and the test round-trips the rendered command through shlex.split.
50 lines
2.1 KiB
Python
50 lines
2.1 KiB
Python
"""Recovery hints must name the user-visible notebook, not its byte-transport copy."""
|
|
import json
|
|
from pathlib import Path
|
|
import shlex
|
|
|
|
import pytest
|
|
|
|
from tools.read_extract import (
|
|
_needs_ocr_warning,
|
|
_pdf_coverage_note,
|
|
extract_document_bytes,
|
|
extract_document_text,
|
|
)
|
|
|
|
|
|
@pytest.mark.parametrize('reader', ['bytes', 'text', 'read_file'])
|
|
def test_notebook_recovery_hint_opens_original_file(tmp_path, reader):
|
|
path = tmp_path / "training team's notebook.ipynb"
|
|
output = 'epoch progress\n' * 2000
|
|
notebook = {'nbformat': 4, 'nbformat_minor': 5, 'metadata': {}, 'cells': [{
|
|
'cell_type': 'code', 'id': 'train', 'execution_count': 1, 'metadata': {},
|
|
'source': ['print(log)'], 'outputs': [{
|
|
'output_type': 'stream', 'name': 'stdout', 'text': output}]}]}
|
|
path.write_text(json.dumps(notebook), encoding='utf-8')
|
|
if reader == 'bytes':
|
|
text = extract_document_bytes(path.read_bytes(), str(path))
|
|
elif reader == 'text':
|
|
text = extract_document_text(str(path))
|
|
else:
|
|
# Real public entry point, including configured local byte transport.
|
|
from tools.file_tools import read_file_tool
|
|
result = json.loads(read_file_tool(str(path), task_id='notebook-hint-recovery'))
|
|
assert result.get('extracted_document'), result
|
|
text = result['content']
|
|
hint = next(line for line in text.splitlines() if 'full output: jq' in line)
|
|
command = hint.split('full output: ', 1)[1].removesuffix(']')
|
|
argv = shlex.split(command)
|
|
assert argv == ['jq', '-r', '.cells[0].outputs', str(path)]
|
|
recovered = json.loads(Path(argv[-1]).read_text(encoding='utf-8'))
|
|
assert recovered['cells'][0]['outputs'][0]['text'] == output
|
|
|
|
|
|
def test_pdftoppm_recovery_hints_round_trip_apostrophe_path(monkeypatch):
|
|
path = "/tmp/training team's notebook.pdf"
|
|
monkeypatch.setattr('tools.read_extract._pdf_page_texts', lambda _p: ['x' * 500, '', '', ''])
|
|
for note in (_needs_ocr_warning(path, [2]), _pdf_coverage_note('/tmp/copy.pdf', display_path=path)):
|
|
command = note.split('`', 2)[1]
|
|
argv = shlex.split(command)
|
|
assert argv[-2] == path, argv
|