modify pdftomd

tests/test_mineru_adapter.py

@@ -1,17 +1,21 @@
 from __future__ import annotations
 
 import os
+import sys
 from pathlib import Path
 
 import pytest
 
 from pdf2md.ir import WarningCode
+from pdf2md.gpu import GpuInfo
 from pdf2md.mineru_adapter import (
     CommandResult,
     MinerUAdapter,
     MinerUOptions,
     StrictLocalViolationError,
+    _run_command,
 )
+from pdf2md.mineru_profile import resolve_mineru_profile
 
 
 class FakeRunner:
@@ -36,10 +40,16 @@ class EnvironmentRunner:
     def __init__(self) -> None:
         self.mineru_device_mode: str | None = None
         self.cuda_visible_devices: str | None = None
+        self.processing_window_size: str | None = None
+        self.max_concurrent_requests: str | None = None
+        self.pdf_render_threads: str | None = None
 
     def __call__(self, command: tuple[str, ...]) -> CommandResult:
         self.mineru_device_mode = os.environ.get("MINERU_DEVICE_MODE")
         self.cuda_visible_devices = os.environ.get("CUDA_VISIBLE_DEVICES")
+        self.processing_window_size = os.environ.get("MINERU_PROCESSING_WINDOW_SIZE")
+        self.max_concurrent_requests = os.environ.get("MINERU_API_MAX_CONCURRENT_REQUESTS")
+        self.pdf_render_threads = os.environ.get("MINERU_PDF_RENDER_THREADS")
         work_dir = Path(command[command.index("-o") + 1])
         work_dir.mkdir(parents=True, exist_ok=True)
         (work_dir / "paper.md").write_text("# Title\n", encoding="utf-8")
@@ -133,6 +143,20 @@ def test_version_empty_output_is_explicit() -> None:
     assert [warning.code for warning in result.warnings] == [WarningCode.MINERU_CLI_FAILED]
 
 
+def test_default_runner_decodes_utf8_process_output() -> None:
+    code = (
+        "import sys; "
+        "sys.stdout.buffer.write('stdout ∙\\n'.encode('utf-8')); "
+        "sys.stderr.buffer.write('stderr ∙\\n'.encode('utf-8'))"
+    )
+
+    result = _run_command((sys.executable, "-c", code))
+
+    assert result.exit_code == 0
+    assert result.stdout == "stdout ∙\n"
+    assert result.stderr == "stderr ∙\n"
+
+
 def test_build_command_is_list_based_and_deterministic(tmp_path: Path) -> None:
     adapter = MinerUAdapter(which=available, runner=FakeRunner())
     input_pdf = tmp_path / "논문 with spaces.pdf"
@@ -200,7 +224,15 @@ def test_successful_mocked_output_parses_markdown_json_and_assets(tmp_path: Path
     ]
     assert result.engine == "MinerU"
     assert result.engine_version == "3.1.0"
-    assert result.engine_options == {"strict_local": True, "gpu_device": "cuda:0"}
+    assert result.engine_options == {
+        "strict_local": True,
+        "gpu_device": "cuda:0",
+        "mineru_profile": {
+            "requested": "auto",
+            "applied": "auto",
+            "environment": {},
+        },
+    }
     assert result.exit_code == 0
     assert result.stdout == "ok"
     assert result.stderr == "warn"
@@ -209,6 +241,7 @@ def test_successful_mocked_output_parses_markdown_json_and_assets(tmp_path: Path
 def test_gpu_option_sets_mineru_environment_and_restores_previous_values(tmp_path: Path, monkeypatch) -> None:
     monkeypatch.setenv("MINERU_DEVICE_MODE", "cpu")
     monkeypatch.setenv("CUDA_VISIBLE_DEVICES", "7")
+    monkeypatch.setenv("MINERU_PROCESSING_WINDOW_SIZE", "99")
     runner = EnvironmentRunner()
     adapter = MinerUAdapter(which=available, runner=runner)
 
@@ -219,6 +252,53 @@ def test_gpu_option_sets_mineru_environment_and_restores_previous_values(tmp_pat
     assert runner.cuda_visible_devices == "0"
     assert os.environ["MINERU_DEVICE_MODE"] == "cpu"
     assert os.environ["CUDA_VISIBLE_DEVICES"] == "7"
+    assert os.environ["MINERU_PROCESSING_WINDOW_SIZE"] == "99"
+
+
+def test_profile_option_sets_allowlisted_mineru_environment_and_engine_options(tmp_path: Path) -> None:
+    gpu = GpuInfo(index=1, name="NVIDIA RTX 4090", memory_total_mib=24564, driver_version="577.00")
+    profile = resolve_mineru_profile("performance", selected_gpu=gpu, cuda_requested=True)
+    runner = EnvironmentRunner()
+    adapter = MinerUAdapter(which=available, runner=runner)
+
+    result = adapter.convert(
+        tmp_path / "paper.pdf",
+        tmp_path / "work",
+        MinerUOptions(
+            gpu_device="cuda:1",
+            mineru_profile="performance",
+            profile_environment=profile.environment,
+            profile_engine_options=profile.to_engine_options(),
+        ),
+    )
+
+    assert result.succeeded is True
+    assert runner.mineru_device_mode == "cuda"
+    assert runner.cuda_visible_devices == "1"
+    assert runner.processing_window_size == "16"
+    assert runner.max_concurrent_requests == "1"
+    assert runner.pdf_render_threads == "4"
+    assert result.engine_options["mineru_profile"]["applied"] == "performance"
+
+
+def test_profile_warnings_are_preserved_in_adapter_result(tmp_path: Path) -> None:
+    gpu = GpuInfo(index=0, name="NVIDIA GeForce GTX 1070 Ti", memory_total_mib=8192, driver_version="577.00")
+    profile = resolve_mineru_profile("performance", selected_gpu=gpu, cuda_requested=True)
+    adapter = MinerUAdapter(which=available, runner=EnvironmentRunner())
+
+    result = adapter.convert(
+        tmp_path / "paper.pdf",
+        tmp_path / "work",
+        MinerUOptions(
+            gpu_device="cuda:0",
+            mineru_profile="performance",
+            profile_environment=profile.environment,
+            profile_engine_options=profile.to_engine_options(),
+            profile_warnings=profile.warnings,
+        ),
+    )
+
+    assert [warning.code for warning in result.warnings] == [WarningCode.MINERU_PROFILE_ADJUSTED]
 
 
 def test_nonzero_exit_does_not_parse_existing_outputs_or_fallback(tmp_path: Path) -> None: