"""Sprint 4.7 upload/parser hardening contract tests.""" from __future__ import annotations import sys from pathlib import Path from types import SimpleNamespace import pytest from app.config import Config from app.services.file_parser import ParseError, parse_document, parse_pdf, parse_text def test_text_parser_uses_charset_normalizer_fallback(tmp_path, monkeypatch): source = tmp_path / "brief.txt" source.write_bytes(b"legacy bytes") calls = [] class BestMatch: encoding = "latin-1" def from_bytes(raw): calls.append(raw) return SimpleNamespace(best=lambda: BestMatch()) monkeypatch.setitem(sys.modules, "charset_normalizer", SimpleNamespace(from_bytes=from_bytes)) monkeypatch.setattr(Config, "UPLOAD_TEXT_MAX_BYTES", 1024) assert parse_text(source) == "legacy bytes" assert calls == [] # UTF-8 succeeds without invoking detection. legacy = "café".encode("latin-1") source.write_bytes(legacy) assert parse_text(source) == "café" assert calls == [legacy] def test_text_parser_rejects_oversized_document(tmp_path, monkeypatch): source = tmp_path / "brief.txt" source.write_bytes(b"x" * 11) monkeypatch.setattr(Config, "UPLOAD_TEXT_MAX_BYTES", 10) with pytest.raises(ParseError, match="text document too large"): parse_text(source) def test_text_parser_preserves_thai_utf8(tmp_path): source = tmp_path / "thai.txt" source.write_text("ระบบฝึกทักษะการขาย", encoding="utf-8") assert parse_text(source) == "ระบบฝึกทักษะการขาย" def test_pdf_parser_rejects_page_bomb(monkeypatch, tmp_path): class FakeDocument: page_count = 3 def __iter__(self): return iter(()) def close(self): pass monkeypatch.setitem(sys.modules, "fitz", SimpleNamespace(open=lambda _path: FakeDocument())) monkeypatch.setattr(Config, "UPLOAD_MAX_PDF_PAGES", 2) with pytest.raises(ParseError, match="too many pages"): parse_pdf(tmp_path / "bomb.pdf") def test_malformed_pdf_returns_controlled_parse_error(tmp_path): pytest.importorskip("fitz") source = tmp_path / "malformed.pdf" source.write_bytes(b"not a PDF") with pytest.raises(ParseError, match="cannot open PDF"): parse_pdf(source) def test_pdf_parser_rejects_excessive_extracted_text(monkeypatch, tmp_path): class FakePage: rect = SimpleNamespace(x0=0, y0=0, x1=10, y1=100) def get_text(self, *args, **kwargs): return "0123456789" class FakeDocument: page_count = 1 def __iter__(self): return iter((FakePage(),)) def close(self): pass monkeypatch.setitem( sys.modules, "fitz", SimpleNamespace(open=lambda _path: FakeDocument(), Rect=lambda *args: args), ) monkeypatch.setattr(Config, "UPLOAD_MAX_EXTRACTED_CHARS", 5) with pytest.raises(ParseError, match="extracted text too large"): parse_pdf(tmp_path / "large.pdf") def test_pdf_parser_rejects_page_without_usable_geometry(monkeypatch, tmp_path): calls = [] class FakePage: rect = None def get_text(self, *args, **kwargs): calls.append((args, kwargs)) return "must not be materialized" class FakeDocument: page_count = 1 def __iter__(self): return iter((FakePage(),)) def close(self): pass monkeypatch.setitem(sys.modules, "fitz", SimpleNamespace(open=lambda _path: FakeDocument())) with pytest.raises(ParseError, match="page geometry unavailable"): parse_pdf(tmp_path / "no-geometry.pdf") assert calls == [] def test_parser_rejects_unsupported_extension(tmp_path): source = tmp_path / "brief.exe" source.write_bytes(b"not supported") with pytest.raises(ParseError, match="unsupported document type"): parse_document(source)