feat(parsing): LaTeX, Nougat, GROBID parsers with chunking

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

tests/parsing/test_nougat.py

@@ -0,0 +1,70 @@
+"""Tests for codex.parsing.nougat."""
+
+from __future__ import annotations
+
+from unittest.mock import MagicMock, patch
+
+import httpx
+import pytest
+
+from codex.parsing.nougat import pdf_to_markdown
+
+
+class _FakeResponse:
+    """Minimal stand-in for httpx.Response."""
+
+    def __init__(self, text: str, status_code: int = 200) -> None:
+        self.text = text
+        self.status_code = status_code
+
+    def raise_for_status(self) -> None:
+        if self.status_code >= 400:
+            raise httpx.HTTPStatusError(
+                "error",
+                request=MagicMock(),
+                response=MagicMock(status_code=self.status_code),
+            )
+
+
+class TestPdfToMarkdown:
+    def test_success_returns_response_text(self, tmp_path: pytest.TempdirFactory) -> None:
+        pdf = tmp_path / "paper.pdf"  # type: ignore[operator]
+        pdf.write_bytes(b"%PDF fake")
+
+        expected = "# Title\nContent"
+        fake_response = _FakeResponse(expected)
+
+        with patch("httpx.Client") as mock_client_cls:
+            mock_client = MagicMock()
+            mock_client_cls.return_value.__enter__.return_value = mock_client
+            mock_client.post.return_value = fake_response
+
+            result = pdf_to_markdown(str(pdf), nougat_url="http://localhost:8080")
+
+        assert result == expected
+        mock_client.post.assert_called_once()
+        call_kwargs = mock_client.post.call_args
+        assert "/predict" in call_kwargs[0][0]
+
+    def test_connect_error_retried_exactly_twice(self, tmp_path: pytest.TempdirFactory) -> None:
+        """ConnectError must trigger 2 retries (3 total attempts)."""
+        pdf = tmp_path / "paper.pdf"  # type: ignore[operator]
+        pdf.write_bytes(b"%PDF fake")
+
+        call_count = 0
+
+        def _raising_post(*args: object, **kwargs: object) -> _FakeResponse:
+            nonlocal call_count
+            call_count += 1
+            raise httpx.ConnectError("refused")
+
+        with patch("httpx.Client") as mock_client_cls:
+            mock_client = MagicMock()
+            mock_client_cls.return_value.__enter__.return_value = mock_client
+            mock_client.post.side_effect = _raising_post
+
+            with pytest.raises(httpx.ConnectError):
+                pdf_to_markdown(str(pdf), nougat_url="http://localhost:8080")
+
+        # tenacity: stop_after_attempt(3) → 1 original + 2 retries = 3 calls
+        assert call_count == 3