modify pdftomd

2026-05-14 10:16:59 +09:00
parent 2232b51fc9
commit dc11880140
69 changed files with 7784 additions and 1150 deletions
@@ -0,0 +1,106 @@
+from __future__ import annotations
+
+from pdf2md.gpu import GpuInfo
+from pdf2md.ir import WarningCode, WarningSeverity
+from pdf2md.mineru_profile import resolve_mineru_profile
+
+
+SAFE_ENV = {
+    "MINERU_PROCESSING_WINDOW_SIZE": "1",
+    "MINERU_API_MAX_CONCURRENT_REQUESTS": "1",
+    "MINERU_PDF_RENDER_THREADS": "1",
+}
+
+
+def test_auto_profile_uses_safe_values_without_gpu_inventory() -> None:
+    profile = resolve_mineru_profile("auto", selected_gpu=None, cuda_requested=True)
+
+    assert profile.applied_profile == "safe"
+    assert profile.environment == SAFE_ENV
+    assert [warning.code for warning in profile.warnings] == [WarningCode.GPU_UNAVAILABLE]
+
+
+def test_auto_profile_uses_safe_values_for_gtx_1070_ti() -> None:
+    gpu = GpuInfo(index=0, name="NVIDIA GeForce GTX 1070 Ti", memory_total_mib=8192, driver_version="577.00")
+
+    profile = resolve_mineru_profile("auto", selected_gpu=gpu, cuda_requested=True)
+
+    assert profile.requested_profile == "auto"
+    assert profile.applied_profile == "safe"
+    assert profile.environment == SAFE_ENV
+    assert profile.selected_gpu_name == "NVIDIA GeForce GTX 1070 Ti"
+
+
+def test_auto_profile_uses_moderate_values_for_16gb_turing_or_newer_gpu() -> None:
+    gpu = GpuInfo(index=0, name="NVIDIA RTX A4000", memory_total_mib=16384, driver_version="577.00")
+
+    profile = resolve_mineru_profile("auto", selected_gpu=gpu, cuda_requested=True)
+
+    assert profile.applied_profile == "auto"
+    assert profile.environment == {
+        "MINERU_PROCESSING_WINDOW_SIZE": "8",
+        "MINERU_API_MAX_CONCURRENT_REQUESTS": "1",
+        "MINERU_PDF_RENDER_THREADS": "4",
+    }
+    assert profile.warnings == ()
+
+
+def test_auto_profile_uses_conservative_values_for_12gb_to_16gb_gpu() -> None:
+    gpu = GpuInfo(index=0, name="NVIDIA RTX 4070", memory_total_mib=12288, driver_version="577.00")
+
+    profile = resolve_mineru_profile("auto", selected_gpu=gpu, cuda_requested=True)
+
+    assert profile.applied_profile == "auto-conservative"
+    assert profile.environment == {
+        "MINERU_PROCESSING_WINDOW_SIZE": "4",
+        "MINERU_API_MAX_CONCURRENT_REQUESTS": "1",
+        "MINERU_PDF_RENDER_THREADS": "2",
+    }
+
+
+def test_performance_profile_uses_performance_values_only_on_strong_gpu() -> 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)
+
+    assert profile.applied_profile == "performance"
+    assert profile.environment == {
+        "MINERU_PROCESSING_WINDOW_SIZE": "16",
+        "MINERU_API_MAX_CONCURRENT_REQUESTS": "1",
+        "MINERU_PDF_RENDER_THREADS": "4",
+    }
+    assert profile.selected_gpu_index == 1
+    assert profile.selected_gpu_vram_mib == 24564
+
+
+def test_performance_profile_downgrades_to_safe_on_weak_gpu() -> 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)
+
+    assert profile.applied_profile == "safe"
+    assert profile.environment == SAFE_ENV
+    assert [warning.code for warning in profile.warnings] == [WarningCode.MINERU_PROFILE_ADJUSTED]
+    assert profile.warnings[0].severity == WarningSeverity.WARNING
+
+
+def test_profile_details_are_json_ready() -> None:
+    gpu = GpuInfo(index=0, name="NVIDIA RTX A5000", memory_total_mib=24564, driver_version="577.00")
+
+    profile = resolve_mineru_profile("auto", selected_gpu=gpu, cuda_requested=True)
+
+    assert profile.to_engine_options() == {
+        "requested": "auto",
+        "applied": "auto",
+        "environment": {
+            "MINERU_API_MAX_CONCURRENT_REQUESTS": "1",
+            "MINERU_PDF_RENDER_THREADS": "4",
+            "MINERU_PROCESSING_WINDOW_SIZE": "8",
+        },
+        "selected_gpu": {
+            "index": 0,
+            "name": "NVIDIA RTX A5000",
+            "memory_total_mib": 24564,
+            "pre_turing_risk": False,
+        },
+    }