Improve classifier fallback for flat CV parsing

2026-04-01 11:00:53 +02:00
parent 517c42250d
commit b283f8b9d2
2 changed files with 274 additions and 29 deletions
@@ -617,9 +617,47 @@ public sealed class ProfileCvControllerTests
        Assert.Contains("Atlas Systems", matchedJob!.Company ?? string.Empty, StringComparison.Ordinal);
        Assert.Contains("Python", structured.Skills);
        Assert.Contains("SQL", structured.Skills);
+        Assert.Equal("classifier", structured.Metadata.Fields["jobs[0].title"].Method);
+        Assert.Equal("block-1", structured.Metadata.Fields["jobs[0].title"].SourceBlockId);
        classifier.Verify(x => x.ClassifyBlockAsync(It.IsAny<string>(), It.IsAny<CancellationToken>()), Times.AtLeastOnce());
    }

+    [Fact]
+    public async Task Parse_uses_classifier_fallback_for_education_blocks_without_real_sections()
+    {
+        var source = "BSc Computer Science\nUniversity of Oslo\nOslo\n2016 - 2019\nGraduated with focus on distributed systems.";
+        var user = new ApplicationUser { Id = "user-1", ProfileCvText = source };
+        var userManager = CreateUserManager();
+        userManager.Setup(x => x.GetUserAsync(It.IsAny<ClaimsPrincipal>())).ReturnsAsync(user);
+        userManager.Setup(x => x.UpdateAsync(user)).ReturnsAsync(IdentityResult.Success);
+        var aiService = new Mock<ISummarizerService>();
+        aiService
+            .Setup(x => x.SummarizeSectionAsync(It.Is<string>(instruction => instruction.Contains("Extract this CV into structured JSON", StringComparison.Ordinal)), source, 3200, 900))
+            .ReturnsAsync("not-json");
+
+        var classifier = new Mock<ICvAiClassifier>();
+        classifier
+            .Setup(x => x.ClassifyBlockAsync(It.IsAny<string>(), It.IsAny<CancellationToken>()))
+            .ReturnsAsync(new CvBlockClassificationResult("Education", 0.87, "education block", "BSc Computer Science", "University of Oslo", "Oslo", "2016", "2019", new List<string> { "Graduated with focus on distributed systems." }));
+
+        await using var db = CreateDb();
+        var paths = CreatePaths();
+        var controller = CreateController(userManager.Object, aiService.Object, db, paths, classifier.Object);
+
+        var result = await controller.Parse(new ProfileCvController.ParseCvRequest(source));
+
+        var ok = Assert.IsType<OkObjectResult>(result.Result);
+        var json = JsonSerializer.Serialize(ok.Value);
+        Assert.Contains("BSc Computer Science", json);
+        Assert.Contains("University of Oslo", json);
+
+        var structured = StructuredCvProfileJson.Deserialize(user.ProfileCvStructureJson);
+        Assert.Single(structured.Education);
+        Assert.Equal("BSc Computer Science", structured.Education[0].Qualification);
+        Assert.Equal("University of Oslo", structured.Education[0].Institution);
+        Assert.Equal("classifier", structured.Metadata.Fields["education[0].qualification"].Method);
+    }
+
    [Fact]
    public async Task Parse_keeps_general_fallback_when_classifier_returns_nothing()
    {