meeting-assistant/MeetingAssistant/Speakers/SpeakerIdentityAttendeeCanonicalizer.cs

using Microsoft.EntityFrameworkCore;
using MeetingAssistant.MeetingNotes;

namespace MeetingAssistant.Speakers;

public interface ISpeakerIdentityAttendeeCanonicalizer
{
    Task<IReadOnlyList<string>> CanonicalizeAsync(
        IReadOnlyList<string> attendees,
        CancellationToken cancellationToken);
}

public sealed class SpeakerIdentityAttendeeCanonicalizer : ISpeakerIdentityAttendeeCanonicalizer
{
    private readonly IDbContextFactory<SpeakerIdentityDbContext> dbContextFactory;

    public SpeakerIdentityAttendeeCanonicalizer(IDbContextFactory<SpeakerIdentityDbContext> dbContextFactory)
    {
        this.dbContextFactory = dbContextFactory;
    }

    public async Task<IReadOnlyList<string>> CanonicalizeAsync(
        IReadOnlyList<string> attendees,
        CancellationToken cancellationToken)
    {
        var attendeeEntries = attendees
            .Select(attendee => new
            {
                Raw = attendee.Trim(),
                DisplayName = NormalizeName(attendee)
            })
            .Where(entry => !string.IsNullOrWhiteSpace(entry.Raw) && !string.IsNullOrWhiteSpace(entry.DisplayName))
            .ToList();
        if (attendeeEntries.Count == 0)
        {
            return [];
        }

        await using var context = await dbContextFactory.CreateDbContextAsync(cancellationToken);
        await SpeakerIdentitySchema.EnsureCreatedOrUpdatedAsync(context, cancellationToken);

        var identities = await context.SpeakerIdentities
            .Include(identity => identity.Aliases)
            .Include(identity => identity.References)
            .ToListAsync(cancellationToken);
        var identitiesByAcceptedName = identities
            .OrderBy(identity => string.IsNullOrWhiteSpace(identity.CanonicalName))
            .ThenByDescending(identity => identity.ReferenceCount)
            .ThenByDescending(identity => identity.UpdatedAt)
            .ThenBy(identity => identity.Id)
            .SelectMany(identity => GetExactAcceptedNames(identity).Select(name => new { Name = name, Identity = identity }))
            .GroupBy(entry => entry.Name, StringComparer.OrdinalIgnoreCase)
            .ToDictionary(group => group.Key, group => group.First().Identity, StringComparer.OrdinalIgnoreCase);

        var result = new List<string>();
        var seenNames = new HashSet<string>(StringComparer.OrdinalIgnoreCase);
        var seenIdentityIds = new HashSet<int>();
        foreach (var attendee in attendeeEntries)
        {
            if (identitiesByAcceptedName.TryGetValue(attendee.DisplayName!, out var identity))
            {
                var displayName = identity.GetDisplayName();
                if (string.IsNullOrWhiteSpace(displayName) || !seenIdentityIds.Add(identity.Id))
                {
                    continue;
                }

                if (seenNames.Add(displayName))
                {
                    result.Add(displayName);
                }

                continue;
            }

            if (seenNames.Add(attendee.Raw))
            {
                result.Add(attendee.Raw);
            }
        }

        return result;
    }

    private static IEnumerable<string> GetExactAcceptedNames(SpeakerIdentity identity)
    {
        return new[] { identity.CanonicalName }
            .Concat(identity.Aliases.Select(alias => alias.Name))
            .Select(NormalizeName)
            .Where(name => !string.IsNullOrWhiteSpace(name))
            .Select(name => name!);
    }

    private static string? NormalizeName(string? name)
    {
        if (string.IsNullOrWhiteSpace(name))
        {
            return null;
        }

        return MeetingAttendeeNames.NormalizeDisplayName(name);
    }
}

public sealed class PassthroughSpeakerIdentityAttendeeCanonicalizer : ISpeakerIdentityAttendeeCanonicalizer
{
    public static PassthroughSpeakerIdentityAttendeeCanonicalizer Instance { get; } = new();

    private PassthroughSpeakerIdentityAttendeeCanonicalizer()
    {
    }

    public Task<IReadOnlyList<string>> CanonicalizeAsync(
        IReadOnlyList<string> attendees,
        CancellationToken cancellationToken)
    {
        var distinctAttendees = attendees
            .Select(attendee => attendee.Trim())
            .Where(name => !string.IsNullOrWhiteSpace(name))
            .Distinct(StringComparer.OrdinalIgnoreCase)
            .ToList();
        return Task.FromResult<IReadOnlyList<string>>(distinctAttendees);
    }
}