Make meeting lifecycle stateful

2026-05-27 12:55:17 +02:00
parent d607b957bb
commit e85274829a
91 changed files with 4076 additions and 479 deletions
@@ -14,7 +14,7 @@ public sealed class SpeakerIdentityServiceTests
    public async Task MatchedUnnamedIdentityEliminatesCandidatesAndPromotesCanonicalName()
    {
        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
-        var identity = await fixture.AddIdentityAsync(["John", "Mike"], [1, 2, 3]);
+        var identity = await fixture.AddIdentityAsync(["John", "Mike"], [1, 2, 3], referenceCount: 2);
        fixture.Matcher.MatchIdentityId = identity.Id;
        var service = fixture.CreateService();

@@ -27,7 +27,9 @@ public sealed class SpeakerIdentityServiceTests
        var saved = await fixture.LoadIdentityAsync(identity.Id);
        Assert.Equal("John", saved.CanonicalName);
        Assert.Equal(["John"], saved.CandidateNames.Select(candidate => candidate.Name));
-        Assert.Equal(1, saved.TranscriptionCount);
+        Assert.Equal(3, saved.References.Count);
+        Assert.All(saved.References, reference =>
+            Assert.Contains("Guest01 was identified as John", File.ReadAllText(reference.TranscriptPath)));
        Assert.Equal("John", result.Segments.Single().Speaker);
        Assert.Equal("John", result.SpeakerMappings["Guest01"]);
    }
@@ -41,7 +43,7 @@ public sealed class SpeakerIdentityServiceTests
            [],
            [1, 2, 3],
            "Chris",
-            transcriptionCount: 5,
+            referenceCount: 5,
            updatedAt: oldTimestamp);
        fixture.Matcher.MatchIdentityId = identity.Id;
        var service = fixture.CreateService();
@@ -54,6 +56,7 @@ public sealed class SpeakerIdentityServiceTests

        var saved = await fixture.LoadIdentityAsync(identity.Id);
        Assert.True(saved.UpdatedAt > oldTimestamp);
+        Assert.Equal(6, saved.References.Count);
    }

    [Fact]
@@ -73,6 +76,7 @@ public sealed class SpeakerIdentityServiceTests
        var saved = await fixture.LoadIdentityAsync(identity.Id);
        Assert.Null(saved.CanonicalName);
        Assert.Equal(["Chris", "Jane"], saved.CandidateNames.Select(candidate => candidate.Name).Order());
+        Assert.Single(saved.References);
        Assert.Equal([7, 8, 9], saved.Snippets.Single().WavBytes);
    }

@@ -93,13 +97,28 @@ public sealed class SpeakerIdentityServiceTests
        var saved = await fixture.LoadIdentityAsync(identity.Id);
        Assert.Equal("Michael", saved.CanonicalName);
        Assert.Equal(["Michael"], saved.CandidateNames.Select(candidate => candidate.Name));
+        Assert.Contains("Guest01 was identified as Michael", File.ReadAllText(saved.References.Single().TranscriptPath));
+    }
+
+    [Fact]
+    public async Task AttendeeCanonicalizerDeduplicatesExactIdentityAliases()
+    {
+        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
+        await fixture.AddIdentityAsync([], [1, 2, 3], "Karl Berger", aliases: ["Berger, Karl"]);
+        var canonicalizer = fixture.CreateAttendeeCanonicalizer();
+
+        var attendees = await canonicalizer.CanonicalizeAsync(
+            ["Karl Berger <karl.work@example.com>", "Berger, Karl <karl.private@example.com>", "Ada"],
+            CancellationToken.None);
+
+        Assert.Equal(["Karl Berger", "Ada"], attendees);
    }

    [Fact]
    public async Task UnmatchedDiarizedSpeakersCreateCandidateIdentitiesFromUnmatchedAttendees()
    {
        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
-        var chris = await fixture.AddIdentityAsync([], [1, 2, 3], "Chris", transcriptionCount: 5);
+        var chris = await fixture.AddIdentityAsync([], [1, 2, 3], "Chris", referenceCount: 5);
        fixture.Matcher.MatchIdentityId = chris.Id;
        var service = fixture.CreateService();

@@ -115,6 +134,7 @@ public sealed class SpeakerIdentityServiceTests

        var learned = await fixture.Context.SpeakerIdentities
            .Include(identity => identity.CandidateNames)
+            .Include(identity => identity.References)
            .Where(identity => identity.CanonicalName == null)
            .OrderBy(identity => identity.Id)
            .ToListAsync();
@@ -124,6 +144,7 @@ public sealed class SpeakerIdentityServiceTests
        {
            Assert.NotEqual(default, identity.CreatedAt);
            Assert.NotEqual(default, identity.UpdatedAt);
+            Assert.Single(identity.References);
            Assert.Equal(["John", "Mike"], identity.CandidateNames.Select(candidate => candidate.Name).Order());
        });
    }
@@ -132,7 +153,7 @@ public sealed class SpeakerIdentityServiceTests
    public async Task UnmatchedSpeakersExcludeAliasesOfMatchedIdentitiesFromCandidates()
    {
        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
-        var michael = await fixture.AddIdentityAsync([], [1, 2, 3], "Michael", transcriptionCount: 5, aliases: ["Mike"]);
+        var michael = await fixture.AddIdentityAsync([], [1, 2, 3], "Michael", referenceCount: 5, aliases: ["Mike"]);
        fixture.Matcher.MatchIdentityId = michael.Id;
        var service = fixture.CreateService();

@@ -147,16 +168,42 @@ public sealed class SpeakerIdentityServiceTests

        var learned = await fixture.Context.SpeakerIdentities
            .Include(identity => identity.CandidateNames)
-            .Where(identity => identity.CanonicalName == null)
+            .Include(identity => identity.References)
+            .Where(identity => identity.CanonicalName == "Jane")
            .SingleAsync();
+        Assert.Equal("Jane", learned.CanonicalName);
        Assert.Equal(["Jane"], learned.CandidateNames.Select(candidate => candidate.Name));
+        Assert.Single(learned.References);
+        Assert.Contains("Guest01 was identified as Jane", File.ReadAllText(learned.References.Single().TranscriptPath));
+    }
+
+    [Fact]
+    public async Task UnmatchedSpeakerWithSingleRemainingCandidateIsPromotedOnInsert()
+    {
+        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
+        var service = fixture.CreateService();
+
+        await service.ProcessFinishedTranscriptAsync(
+            fixture.CreateRequest(
+                ["Manuel"],
+                [new TranscriptionSegment(TimeSpan.Zero, TimeSpan.FromSeconds(4), "Guest01", "unknown")]),
+            CancellationToken.None);
+
+        var learned = await fixture.Context.SpeakerIdentities
+            .Include(identity => identity.CandidateNames)
+            .Include(identity => identity.References)
+            .SingleAsync();
+        Assert.Equal("Manuel", learned.CanonicalName);
+        Assert.Equal(["Manuel"], learned.CandidateNames.Select(candidate => candidate.Name));
+        Assert.Single(learned.References);
+        Assert.Contains("Guest01 was identified as Manuel", File.ReadAllText(learned.References.Single().TranscriptPath));
    }

    [Fact]
    public async Task SuppliedSpeakerSamplesAreUsedBeforeExtractingFromAudioFile()
    {
        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
-        var chris = await fixture.AddIdentityAsync([], [1, 2, 3], "Chris", transcriptionCount: 5);
+        var chris = await fixture.AddIdentityAsync([], [1, 2, 3], "Chris", referenceCount: 5);
        fixture.Matcher.MatchIdentityId = chris.Id;
        var service = fixture.CreateService();
        var segment = new TranscriptionSegment(TimeSpan.Zero, TimeSpan.FromSeconds(5), "Guest03", "hello from the live buffer");
@@ -177,7 +224,7 @@ public sealed class SpeakerIdentityServiceTests
    public async Task LiveKnownSpeakerIdentificationDoesNotMutateIdentityDatabase()
    {
        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
-        var identity = await fixture.AddIdentityAsync(["John", "Mike"], [1, 2, 3], transcriptionCount: 5);
+        var identity = await fixture.AddIdentityAsync(["John", "Mike"], [1, 2, 3], referenceCount: 5);
        fixture.Matcher.MatchIdentityId = identity.Id;
        var service = fixture.CreateService();

@@ -190,7 +237,7 @@ public sealed class SpeakerIdentityServiceTests
        var saved = await fixture.LoadIdentityAsync(identity.Id);
        Assert.Null(saved.CanonicalName);
        Assert.Equal(["John", "Mike"], saved.CandidateNames.Select(candidate => candidate.Name).Order());
-        Assert.Equal(5, saved.TranscriptionCount);
+        Assert.Equal(5, saved.References.Count);
        Assert.Single(saved.Snippets);
        Assert.Equal("Guest01", result.Segments.Single().Speaker);
        Assert.Empty(result.SpeakerMappings);
@@ -200,9 +247,9 @@ public sealed class SpeakerIdentityServiceTests
    public async Task SpeakerMatchingPrioritizesAttendeeDisplayNamesAndAliases()
    {
        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
-        var unrelated = await fixture.AddIdentityAsync([], [1], "Frequent", transcriptionCount: 99);
-        var aliasMatch = await fixture.AddIdentityAsync([], [2], "Michael", transcriptionCount: 1, aliases: ["Mike"]);
-        var displayNameMatch = await fixture.AddIdentityAsync([], [3], "Jane", transcriptionCount: 2);
+        var unrelated = await fixture.AddIdentityAsync([], [1], "Frequent", referenceCount: 99);
+        var aliasMatch = await fixture.AddIdentityAsync([], [2], "Michael", referenceCount: 1, aliases: ["Mike"]);
+        var displayNameMatch = await fixture.AddIdentityAsync([], [3], "Jane", referenceCount: 2);
        var service = fixture.CreateService();

        await service.IdentifyKnownSpeakersAsync(
@@ -223,10 +270,10 @@ public sealed class SpeakerIdentityServiceTests
            options.MaxMatchCandidates = 2;
            options.MatchIdentityActiveAge = TimeSpan.FromDays(365);
        });
-        var activeHigh = await fixture.AddIdentityAsync([], [1], "Active High", transcriptionCount: 50);
-        await fixture.AddIdentityAsync([], [2], "Stale High", transcriptionCount: 100, updatedAt: DateTimeOffset.UtcNow.AddYears(-2));
-        var activeLow = await fixture.AddIdentityAsync([], [3], "Active Low", transcriptionCount: 5);
-        var staleAttendee = await fixture.AddIdentityAsync([], [4], "Former Employee", transcriptionCount: 1, updatedAt: DateTimeOffset.UtcNow.AddYears(-2), aliases: ["Former"]);
+        var activeHigh = await fixture.AddIdentityAsync([], [1], "Active High", referenceCount: 50);
+        await fixture.AddIdentityAsync([], [2], "Stale High", referenceCount: 100, updatedAt: DateTimeOffset.UtcNow.AddYears(-2));
+        var activeLow = await fixture.AddIdentityAsync([], [3], "Active Low", referenceCount: 5);
+        var staleAttendee = await fixture.AddIdentityAsync([], [4], "Former Employee", referenceCount: 1, updatedAt: DateTimeOffset.UtcNow.AddYears(-2), aliases: ["Former"]);
        var service = fixture.CreateService();

        await service.IdentifyKnownSpeakersAsync(
@@ -240,6 +287,68 @@ public sealed class SpeakerIdentityServiceTests
        Assert.DoesNotContain(activeLow.Id, requestedIds);
    }

+    [Fact]
+    public async Task KnownSpeakerMappingsExcludeMappedSpeakersAndIdentitiesFromMatching()
+    {
+        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
+        var manuel = await fixture.AddIdentityAsync([], [1], "Manuel", referenceCount: 99);
+        var daniel = await fixture.AddIdentityAsync([], [2], "Daniel", referenceCount: 5);
+        var service = fixture.CreateService();
+
+        await service.IdentifyKnownSpeakersAsync(
+            fixture.CreateRequest(
+                ["Manuel", "Daniel"],
+                [
+                    new TranscriptionSegment(TimeSpan.Zero, TimeSpan.FromSeconds(4), "Guest-2", "already identified"),
+                    new TranscriptionSegment(TimeSpan.FromSeconds(4), TimeSpan.FromSeconds(8), "Guest-1", "unresolved")
+                ],
+                [
+                    new SpeakerAudioSample(
+                        "Guest-2",
+                        new TranscriptionSegment(TimeSpan.Zero, TimeSpan.FromSeconds(4), "Guest-2", "already identified"),
+                        [4],
+                        90),
+                    new SpeakerAudioSample(
+                        "Guest-1",
+                        new TranscriptionSegment(TimeSpan.FromSeconds(4), TimeSpan.FromSeconds(8), "Guest-1", "unresolved"),
+                        [5],
+                        90)
+                ],
+                new Dictionary<string, string>(StringComparer.OrdinalIgnoreCase)
+                {
+                    ["Guest-2"] = "Manuel"
+                }),
+            CancellationToken.None);
+
+        var request = fixture.Matcher.Requests.Single();
+        Assert.Equal("Guest-1", request.DiarizedSpeaker);
+        Assert.Equal([daniel.Id], request.Candidates.Select(candidate => candidate.IdentityId));
+        Assert.DoesNotContain(manuel.Id, request.Candidates.Select(candidate => candidate.IdentityId));
+    }
+
+    [Fact]
+    public async Task NamedSpeakersExcludeTheirIdentityFromFurtherMatching()
+    {
+        await using var fixture = await SpeakerIdentityFixture.CreateAsync();
+        var manuel = await fixture.AddIdentityAsync([], [1], "Manuel", referenceCount: 99);
+        var daniel = await fixture.AddIdentityAsync([], [2], "Daniel", referenceCount: 5);
+        var service = fixture.CreateService();
+
+        await service.ProcessFinishedTranscriptAsync(
+            fixture.CreateRequest(
+                ["Manuel", "Daniel"],
+                [
+                    new TranscriptionSegment(TimeSpan.Zero, TimeSpan.FromSeconds(4), "Manuel", "already relabeled"),
+                    new TranscriptionSegment(TimeSpan.FromSeconds(4), TimeSpan.FromSeconds(8), "Guest-1", "unresolved")
+                ]),
+            CancellationToken.None);
+
+        var request = fixture.Matcher.Requests.Single();
+        Assert.Equal("Guest-1", request.DiarizedSpeaker);
+        Assert.Equal([daniel.Id], request.Candidates.Select(candidate => candidate.IdentityId));
+        Assert.DoesNotContain(manuel.Id, request.Candidates.Select(candidate => candidate.IdentityId));
+    }
+
    [Fact]
    public async Task SchemaUpgradeAddsLastModifiedColumnInitializedFromCreatedAt()
    {
@@ -275,14 +384,17 @@ public sealed class SpeakerIdentityServiceTests

    private sealed class SpeakerIdentityFixture : IAsyncDisposable
    {
+        private readonly string tempDirectory;
        private readonly string dbPath;
        private readonly SpeakerIdentificationOptions options;

        private SpeakerIdentityFixture(
+            string tempDirectory,
            string dbPath,
            SpeakerIdentityDbContext context,
            SpeakerIdentificationOptions options)
        {
+            this.tempDirectory = tempDirectory;
            this.dbPath = dbPath;
            Context = context;
            this.options = options;
@@ -297,8 +409,9 @@ public sealed class SpeakerIdentityServiceTests
        public static async Task<SpeakerIdentityFixture> CreateAsync(
            Action<SpeakerIdentificationOptions>? configureOptions = null)
        {
-            var dbPath = Path.Combine(Path.GetTempPath(), "meeting-assistant-tests", $"{Guid.NewGuid():N}.db");
-            Directory.CreateDirectory(Path.GetDirectoryName(dbPath)!);
+            var tempDirectory = Path.Combine(Path.GetTempPath(), "meeting-assistant-tests", Guid.NewGuid().ToString("N"));
+            Directory.CreateDirectory(tempDirectory);
+            var dbPath = Path.Combine(tempDirectory, "speaker-identities.db");
            var options = new SpeakerIdentificationOptions
            {
                DatabasePath = dbPath,
@@ -310,7 +423,7 @@ public sealed class SpeakerIdentityServiceTests
                .UseSqlite($"Data Source={dbPath};Pooling=False")
                .Options);
            await SpeakerIdentitySchema.EnsureCreatedOrUpdatedAsync(context, CancellationToken.None);
-            return new SpeakerIdentityFixture(dbPath, context, options);
+            return new SpeakerIdentityFixture(tempDirectory, dbPath, context, options);
        }

        public SpeakerIdentityService CreateService()
@@ -323,29 +436,44 @@ public sealed class SpeakerIdentityServiceTests
                NullLogger<SpeakerIdentityService>.Instance);
        }

+        public SpeakerIdentityAttendeeCanonicalizer CreateAttendeeCanonicalizer()
+        {
+            return new SpeakerIdentityAttendeeCanonicalizer(new TestSpeakerIdentityDbContextFactory(dbPath));
+        }
+
        public SpeakerIdentificationRequest CreateRequest(
            IReadOnlyList<string> attendees,
            IReadOnlyList<TranscriptionSegment> segments,
-            IReadOnlyList<SpeakerAudioSample>? samples = null)
+            IReadOnlyList<SpeakerAudioSample>? samples = null,
+            IReadOnlyDictionary<string, string>? knownSpeakerMappings = null)
        {
+            var meetingNotePath = Path.Combine(tempDirectory, "meeting.md");
+            var transcriptPath = Path.Combine(tempDirectory, "transcript.md");
+            File.WriteAllText(transcriptPath, "Transcript");
            return new SpeakerIdentificationRequest(
                "test.wav",
-                MeetingNoteTemplate.Create(
-                "Test Meeting",
-                DateTimeOffset.Parse("2026-05-20T12:00:00+02:00"),
-                attendees: attendees,
-                transcriptPath: "transcript.md",
-                assistantContextPath: "context.md",
-                summaryPath: "summary.md"),
+                new MeetingNote(
+                    meetingNotePath,
+                    new MeetingNoteFrontmatter
+                    {
+                        Title = "Test Meeting",
+                        StartTime = DateTimeOffset.Parse("2026-05-20T12:00:00+02:00"),
+                        Attendees = attendees.ToList(),
+                        Transcript = transcriptPath,
+                        AssistantContext = Path.Combine(tempDirectory, "context.md"),
+                        Summary = Path.Combine(tempDirectory, "summary.md")
+                    },
+                    ""),
                segments,
-                samples);
+                samples,
+                knownSpeakerMappings);
        }

        public async Task<SpeakerIdentity> AddIdentityAsync(
            IReadOnlyList<string> candidates,
            byte[] snippet,
            string? canonicalName = null,
-            int transcriptionCount = 0,
+            int referenceCount = 0,
            IReadOnlyList<string>? aliases = null,
            DateTimeOffset? updatedAt = null)
        {
@@ -353,7 +481,6 @@ public sealed class SpeakerIdentityServiceTests
            var identity = new SpeakerIdentity
            {
                CanonicalName = canonicalName,
-                TranscriptionCount = transcriptionCount,
                CreatedAt = timestamp,
                UpdatedAt = timestamp,
                Aliases = aliases?.Select(alias => new SpeakerAlias { Name = alias }).ToList() ?? [],
@@ -365,7 +492,20 @@ public sealed class SpeakerIdentityServiceTests
                        WavBytes = snippet,
                        CreatedAt = timestamp.AddMinutes(-10)
                    }
-                ]
+                ],
+                References = Enumerable.Range(0, referenceCount)
+                    .Select(index =>
+                    {
+                        var transcriptPath = Path.Combine(tempDirectory, $"reference-{Guid.NewGuid():N}-{index}.md");
+                        File.WriteAllText(transcriptPath, "Transcript");
+                        return new SpeakerIdentityReference
+                        {
+                            MeetingNotePath = Path.Combine(tempDirectory, $"reference-{Guid.NewGuid():N}-{index}.md"),
+                            TranscriptPath = transcriptPath,
+                            CreatedAt = timestamp.AddMinutes(index)
+                        };
+                    })
+                    .ToList()
            };
            Context.SpeakerIdentities.Add(identity);
            await Context.SaveChangesAsync();
@@ -379,15 +519,16 @@ public sealed class SpeakerIdentityServiceTests
                .Include(identity => identity.CandidateNames)
                .Include(identity => identity.Aliases)
                .Include(identity => identity.Snippets)
+                .Include(identity => identity.References)
                .SingleAsync(identity => identity.Id == id);
        }

        public async ValueTask DisposeAsync()
        {
            await Context.DisposeAsync();
-            if (File.Exists(dbPath))
+            if (Directory.Exists(tempDirectory))
            {
-                File.Delete(dbPath);
+                Directory.Delete(tempDirectory, recursive: true);
            }
        }
    }