fix(kernel,server,tui,inference): generic resource gauge + F-002/F-003/F-004

- generic SystemResourceProbe: owner-agnostic /proc/meminfo system RAM overlaid on the vendor GPU probe, wired on both managed and static paths so the VRAM/GPU/RAM gauge renders with an external llama-server (was dormant) - F-002: classify provider 4xx (e.g. llama.cpp 400) as terminal, not retryable (except 408/429); stops retrying deterministic request failures to exhaustion - F-003: FileSystemPromptLoader expands leading ~ / ~/ to user home - F-004: map InferenceFailedEvent + RetryAttemptedEvent to new ServerMessage.InferenceFailed / RetryAttempted, decoded+rendered in tui-go; ArtifactContentStoredEvent explicitly mapped to null (internal, no warn) - tests: tilde expansion, SystemResourceProbe (static/managed/fail-soft)
2026-06-10 11:07:08 +04:00
parent a92b5a3531
commit ee24b7786a
14 changed files with 254 additions and 24 deletions
@@ -0,0 +1,55 @@
+package com.correx.testing.contracts.model.inference
+
+import com.correx.infrastructure.inference.commons.GpuStatus
+import com.correx.infrastructure.inference.commons.ResourceProbe
+import com.correx.infrastructure.inference.commons.ResourceSnapshot
+import com.correx.infrastructure.inference.commons.SystemResourceProbe
+import com.correx.infrastructure.inference.commons.UnavailableProbe
+import org.junit.jupiter.api.Assertions.assertEquals
+import org.junit.jupiter.api.Assertions.assertNull
+import org.junit.jupiter.api.Test
+
+class SystemResourceProbeTest {
+
+    private val bytesPerKb = 1024L
+
+    @Test
+    fun `overlays system RAM onto an unavailable GPU probe (static path, F-001 gauge)`() {
+        val probe = SystemResourceProbe(
+            gpuProbe = UnavailableProbe,
+            systemRamReader = { (4L * bytesPerKb) to (16L * bytesPerKb) },
+        )
+        val snap = probe.probe()
+        assertNull(snap.gpu) // no managed model / no GPU, yet RAM still renders
+        assertNull(snap.processRssBytes)
+        assertEquals(4L * bytesPerKb, snap.systemRamUsedBytes)
+        assertEquals(16L * bytesPerKb, snap.systemRamTotalBytes)
+    }
+
+    @Test
+    fun `preserves the delegate GPU and process RSS while adding system RAM (managed path)`() {
+        val gpuDelegate = ResourceProbe {
+            ResourceSnapshot(
+                gpu = GpuStatus(memoryUsedMb = 2048, memoryTotalMb = 8192, utilizationPct = 50),
+                processRssBytes = 999L,
+            )
+        }
+        val probe = SystemResourceProbe(
+            gpuProbe = gpuDelegate,
+            systemRamReader = { 100L to 200L },
+        )
+        val snap = probe.probe()
+        assertEquals(2048L, snap.gpu?.memoryUsedMb)
+        assertEquals(999L, snap.processRssBytes)
+        assertEquals(100L, snap.systemRamUsedBytes)
+        assertEquals(200L, snap.systemRamTotalBytes)
+    }
+
+    @Test
+    fun `fails soft to null system RAM when the reader is unavailable`() {
+        val probe = SystemResourceProbe(gpuProbe = UnavailableProbe, systemRamReader = { null })
+        val snap = probe.probe()
+        assertNull(snap.systemRamUsedBytes)
+        assertNull(snap.systemRamTotalBytes)
+    }
+}