add test for aggregator and analytics

Backend/AI_stats_measurement.Backend.csproj

@@ -25,6 +25,7 @@
     </PackageReference>
     <PackageReference Include="Microsoft.Recognizers.Text.Number" Version="1.8.13" />
     <PackageReference Include="Microsoft.VisualStudio.Web.CodeGeneration.Design" Version="9.0.12" />
+    <PackageReference Include="Moq" Version="4.20.72" />
     <PackageReference Include="Npgsql.EntityFrameworkCore.PostgreSQL" Version="9.0.4" />
     <PackageReference Include="OpenAI" Version="2.9.1" />
     <PackageReference Include="System.IdentityModel.Tokens.Jwt" Version="8.17.0" />

Backend/Services/AnalyticsService.cs

@@ -384,15 +384,15 @@ private static bool MatchesLlmGroup(string provider, string? llmGroup)
                     || provider.StartsWith("grok-4.20-reasoning", StringComparison.OrdinalIgnoreCase);
             }
 
-            // fallback (normal filtering)
+            // fallback
             return provider.StartsWith(llmGroup, StringComparison.OrdinalIgnoreCase);
         }
 
         public Dictionary<string, MetricsOverTimeDto> GetWeeklyMetricsPerNsi(
-    List<FactCheckResult> results,
-    string? nsi,
-    string? llm,
-    string? theme)
+            List<FactCheckResult> results,
+            string? nsi,
+            string? llm,
+            string? theme)
         {
             var filtered = ApplyFilters(results, nsi, llm, theme);
 
@@ -436,8 +436,6 @@ public Dictionary<string, MetricsOverTimeDto> GetWeeklyMetricsPerNsi(
                 );
         }
 
-
-
         public class MetricResultDto
         {
             public double Score { get; set; }

Tests/AI_stats_measurement.Tests.csproj

@@ -13,6 +13,7 @@
       <PrivateAssets>all</PrivateAssets>
       <IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
     </PackageReference>
+    <PackageReference Include="Microsoft.EntityFrameworkCore.InMemory" Version="9.0.11" />
     <PackageReference Include="Microsoft.NET.Test.Sdk" Version="17.12.0" />
     <PackageReference Include="xunit" Version="2.9.2" />
     <PackageReference Include="xunit.runner.visualstudio" Version="2.8.2" />

Tests/AnalyticsServiceTest.cs

@@ -1,4 +1,7 @@
-using AI_stats_measurement.Backend.Services.Parsing;
+using AI_stats_measurement.Backend.Models;
+using AI_stats_measurement.Backend.Services;
+using AI_stats_measurement.Backend.Services.Parsing;
+using AI_stats_measurement.Models;
 using System;
 using System.Collections.Generic;
 using System.Linq;
@@ -7,16 +10,145 @@
 
 namespace AI_stats_measurement.Tests
 {
+
     public class AnalyticsServiceTest
     {
         [Fact]
-        public void ComputeAccuracyScore_()
+        public void GetMetricsPerNsi_Computes_FindabilityScore()
+        {
+            var service = new AnalyticsService();
+
+            var results = new List<FactCheckResult>
+        {
+            CreateResult(1, "CBS", "gpt-4o-mini", 100, 100, true),
+            CreateResult(2, "CBS", "gpt-4o-mini", 90, 100, false),
+        };
+
+            var metrics = service.GetMetricsPerNsi(results, "CBS", null, null);
+
+            Assert.Single(metrics);
+            Assert.Equal(5.0, metrics[0].FindabilityScore);
+        }
+
+        [Fact]
+        public void GetMetricsPerNsi_Computes_PerfectConsistencyScore_WhenAnswersAreEqual()
+        {
+            var service = new AnalyticsService();
+
+            var results = new List<FactCheckResult>
+        {
+            CreateResult(1, "CBS", "gpt-4o-mini", 100, 100, true),
+            CreateResult(1, "CBS", "gpt-4o-mini", 100, 100, true),
+        };
+
+            var metrics = service.GetMetricsPerNsi(results, "CBS", null, null);
+
+            Assert.Single(metrics);
+            Assert.Equal(10.0, metrics[0].ConsistencyScore);
+        }
+
+        [Fact]
+        public void GetMetricsPerNsi_GroupsResults_PerNsi()
+        {
+            var service = new AnalyticsService();
+
+            var results = new List<FactCheckResult>
+        {
+            CreateResult(1, "CBS", "gpt-4o-mini", 100, 100, true),
+            CreateResult(2, "OECD", "gpt-4o-mini", 200, 200, true),
+        };
+
+            var metrics = service.GetMetricsPerNsi(results, null, null, null);
+
+            Assert.Equal(2, metrics.Count);
+            Assert.Contains(metrics, m => m.Nsi == "CBS");
+            Assert.Contains(metrics, m => m.Nsi == "OECD");
+        }
+
+        [Fact]
+        public void GetMetricsPerNsi_Filters_ByNsi()
+        {
+            var service = new AnalyticsService();
+
+            var results = new List<FactCheckResult>
+        {
+            CreateResult(1, "CBS", "gpt-4o-mini", 100, 100, true),
+            CreateResult(2, "OECD", "gpt-4o-mini", 200, 200, true),
+        };
+
+            var metrics = service.GetMetricsPerNsi(results, "CBS", null, null);
+
+            Assert.Single(metrics);
+            Assert.Equal("CBS", metrics[0].Nsi);
+        }
+
+
+        [Fact]
+        public void GetWeeklyMetricsPerNsi_GroupsResults_ByWeek()
+        {
+            var service = new AnalyticsService();
+
+            var results = new List<FactCheckResult>
         {
-            var text = "In 2020 was de gemiddelde verkoopprijs ongeveer € 348.000.";
+            CreateResult(1, "CBS", "gpt-4o-mini", 100, 100, true, new DateTime(2026, 1, 5, 0, 0, 0, DateTimeKind.Utc)),
+            CreateResult(2, "CBS", "gpt-4o-mini", 90, 100, false, new DateTime(2026, 1, 12, 0, 0, 0, DateTimeKind.Utc)),
+        };
 
-            var parsed = ModelResponseParser.ParseDutch(0, text);
+            var weekly = service.GetWeeklyMetricsPerNsi(results, "CBS", null, null);
 
-            Assert.Equal(348_000m, parsed.Answer);
+            Assert.True(weekly.ContainsKey("CBS"));
+            Assert.Equal(2, weekly["CBS"].Findability.Count);
+            Assert.Equal(2, weekly["CBS"].Accuracy.Count);
+            Assert.Equal(2, weekly["CBS"].Consistency.Count);
         }
+
+        private static FactCheckResult CreateResult(
+            int promptId,
+            string nsi,
+            string llm,
+            decimal actualAnswer,
+            decimal expectedAnswer,
+            bool sourceIsCorrect,
+            DateTime? createdUtc = null)
+        {
+            Source source = new Source { Id = 1, Name = "Test Source" };
+            var prompt = new Prompt(nsi, "test", "test", DateTime.Now, "none", "question", expectedAnswer, source, "");
+
+            var modelResponse = ModelResponse.Import(
+                id: promptId,
+                promptId: promptId,
+                provider: llm,
+                rawText: "test response",
+                exception: null,
+                createdUtc: createdUtc ?? DateTime.UtcNow
+            );
+
+            modelResponse.Prompt = prompt;
+
+            var parsed = new ParsedModelResponse(
+                modelResponseId: modelResponse.Id,
+                answer: actualAnswer,
+                extractedSources: new List<ExtractedSource>()
+            );
+
+            parsed.ModelResponse = modelResponse;
+
+            var absoluteError = Math.Abs(expectedAnswer - actualAnswer);
+            var relativeError = expectedAnswer == 0 ? 0 : absoluteError / expectedAnswer;
+
+            var fact = new FactCheckResult(
+                parsedModelResponseId: parsed.Id,
+                absoluteError: absoluteError,
+                relativeError: relativeError,
+                answerIsCorrect: actualAnswer == expectedAnswer,
+                sourceIsCorrect: sourceIsCorrect,
+                abstained: false
+            );
+
+            fact.ParsedModelResponse = parsed;
+
+            return fact;
+        }    
     }
 }
+

Tests/LlmAggregatorTests.cs

@@ -0,0 +1,108 @@
+using AI_stats_measurement.Backend.Models;
+using AI_stats_measurement.Data;
+using AI_stats_measurement.Interface;
+using AI_stats_measurement.Services;
+using Microsoft.EntityFrameworkCore;
+using Moq;
+using Xunit;
+
+namespace AI_stats_measurement.Tests;
+
+public class LlmAggregatorTests
+{
+    private AIMeasureDbContext CreateContext()
+    {
+        var options = new DbContextOptionsBuilder<AIMeasureDbContext>()
+            .UseInMemoryDatabase(Guid.NewGuid().ToString())
+            .Options;
+
+        return new AIMeasureDbContext(options);
+    }
+
+    [Fact]
+    public async Task AskByPromptIdsAsync_Calls_Only_Selected_Models_And_Prompts()
+    {
+        Source source = new Source { Id = 1, Name = "Test Source" };
+        var prompt1 = new Prompt("CBS", "test", "test", DateTime.Now, "none", "question", 100, source, "");
+        var prompt2 = new Prompt("CBS", "test", "test", DateTime.Now, "none", "question", 100, source, "");
+
+        var context = CreateContext();
+        context.Prompts.AddRange(prompt1, prompt2);
+        await context.SaveChangesAsync();
+
+        var mock1 = new Mock<ILlmQuerier>();
+        mock1.SetupGet(q => q.Name).Returns("ChatGPT");
+        mock1.Setup(q => q.AskAsync(It.IsAny<Prompt>(), It.IsAny<CancellationToken>()))
+            .ReturnsAsync("antwoord");
+
+        var mock2 = new Mock<ILlmQuerier>();
+        mock2.SetupGet(q => q.Name).Returns("Gemini");
+        mock2.Setup(q => q.AskAsync(It.IsAny<Prompt>(), It.IsAny<CancellationToken>()))
+            .ReturnsAsync("antwoord");
+
+        var aggregator = new LlmAggregator(
+            new[] { mock1.Object, mock2.Object },
+            context
+        );
+
+        var result = await aggregator.AskByPromptIdsAsync(
+            new List<int> { 1, 2 },
+            new List<string> { "ChatGPT" }, 
+            CancellationToken.None
+        );
+
+        Assert.Equal(2, result.Count); 
+
+        mock1.Verify(q => q.AskAsync(It.IsAny<Prompt>(), It.IsAny<CancellationToken>()), Times.Exactly(2));
+        mock2.Verify(q => q.AskAsync(It.IsAny<Prompt>(), It.IsAny<CancellationToken>()), Times.Never);
+    }
+
+    [Fact]
+    public async Task AskByPromptIdsAsync_Returns_Empty_When_No_Prompts_Found()
+    {
+        var context = CreateContext();
+
+        var mock = new Mock<ILlmQuerier>();
+        mock.SetupGet(q => q.Name).Returns("ChatGPT");
+
+        var aggregator = new LlmAggregator(
+            new[] { mock.Object },
+            context
+        );
+
+        var result = await aggregator.AskByPromptIdsAsync(
+            new List<int> { 999 }, 
+            new List<string> { "ChatGPT" },
+            CancellationToken.None
+        );
+
+        Assert.Empty(result);
+    }
+
+    [Fact]
+    public async Task AskByPromptIdsAsync_Skips_Non_Selected_Models()
+    {
+        Source source = new Source { Id = 1, Name = "Test Source" };
+        var prompt = new Prompt("CBS", "test", "test", DateTime.Now, "none", "question", 100, source, "");
+
+        var context = CreateContext();
+        context.Prompts.Add(prompt);
+        await context.SaveChangesAsync();
+
+        var mock = new Mock<ILlmQuerier>();
+        mock.SetupGet(q => q.Name).Returns("Gemini");
+
+        var aggregator = new LlmAggregator(
+            new[] { mock.Object },
+            context
+        );
+
+        var result = await aggregator.AskByPromptIdsAsync(
+            new List<int> { 1 },
+            new List<string> { "ChatGPT" },
+            CancellationToken.None
+        );
+
+        Assert.Empty(result);
+    }
+}

Tests/ModelResponseParserTests.cs

@@ -197,17 +197,17 @@ public void Parse_Returns_Zero_When_No_Answer_Is_Found()
     }
 
     [Fact]
-    public void Parse_Returns_1()
+    public void Parse_Returns_Zero_Ignore_Source()
     {
-        var text = "Bron: [CBS - Arbeidsongeschiktheidsuitkeringen] (https://www.cbs.nl/nl-nl/cijfers/detail/arbeidsongeschiktheidsuitkeringen)";
+        var text = "Bron: [CBS - Arbeidsongeschiktheidsuitkeringen] (https://www.cbs.nl/nl-nl/cijfers/86165NED/detail/arbeidsongeschiktheidsuitkeringen)";
 
         var parsed = ModelResponseParser.ParseDutch(0, text);
 
         Assert.Equal(0m, parsed.Answer);
     }
 
     [Fact]
-    public void Parse_Returns_2()
+    public void Parse_Returns_Answer_Ignore_Year()
     {
         var text = "**De levensverwachting bij geboorte voor mannen in Nederland in 2022 was 80,1 jaar.**[[1]] (https://www.lifetable.de/File/GetDocument/data/NLD/NLD000020222022CU1.pdf)\r\n\r\nDit cijfer komt uit de officiële sterftetafels(levensverwachtingstafels) van het **Centraal Bureau voor de Statistiek(CBS)**. Ter vergelijking: in 2020 was het circa 79,7 jaar(daling door COVID-19), in 2024 circa 80,5 jaar.[[2]] (https://www.cbs.nl/?sc_itemid=40d28916-85d7-494e-84d6-9d97ca41e253&sc_lang=nl-%20nl)\r\n\r\n**Bron:** CBS, tabel 37360ned(Levensverwachting; geslacht, leeftijd).  \r\nDirecte link: [https://www.cbs.nl/nl-nl/cijfers/detail/37360ned](https://www.cbs.nl/nl-nl/cijfers/detail/37360ned) of de StatLine-tabel op opendata.cbs.nl.";
 
@@ -217,7 +217,7 @@ public void Parse_Returns_2()
     }
 
     [Fact]
-    public void Parse_Returns_3()
+    public void Parse_Returns_Answer_Recognize_Ton()
     {
         var text = "30.300 ton";
 
@@ -227,7 +227,7 @@ public void Parse_Returns_3()
     } 
 
     [Fact]
-    public void Parse_Returns_4()
+    public void Parse_Returns_Zero_Ignore_Dates()
     {
         var text = "1 januari 2021";
 
@@ -237,7 +237,7 @@ public void Parse_Returns_4()
     }
 
     [Fact]
-    public void Parse_Returns_5()
+    public void Parse_Returns_Zero_Ignore_Age_Range_English()
     {
         var text = "aged 25 to 29 ";
 
@@ -247,7 +247,7 @@ public void Parse_Returns_5()
     }
 
     [Fact]
-    public void Parse_Returns_6()
+    public void Parse_Returns_Zero_Ignore_Age_Range_Dutch()
     {
         var text = "18 tot 25 jaar";
 
@@ -257,16 +257,12 @@ public void Parse_Returns_6()
     }
 
     [Fact]
-    public void Parse_Returns_7()
+    public void Parse_Returns_Zero_Ignore_Base_Year()
     {
         var text = "(base year 2025=100)";
 
         var parsed = ModelResponseParser.ParseDutch(0, text);
 
         Assert.Equal(0, parsed.Answer);
     }
-
-
-
-
 }