feat(search): weight subject above body in FTS vector #14

Merged
cesnimda merged 1 commits from feature/search-weighted-vector into develop 2026-07-01 21:29:39 +02:00
4 changed files with 794 additions and 5 deletions
Showing only changes of commit 9545f94461 - Show all commits
@@ -0,0 +1,742 @@
// <auto-generated />
using System;
using InboxIntel.Infrastructure.Persistence;
using Microsoft.EntityFrameworkCore;
using Microsoft.EntityFrameworkCore.Infrastructure;
using Microsoft.EntityFrameworkCore.Migrations;
using Microsoft.EntityFrameworkCore.Storage.ValueConversion;
using Npgsql.EntityFrameworkCore.PostgreSQL.Metadata;
using NpgsqlTypes;
#nullable disable
namespace InboxIntel.Infrastructure.Migrations
{
[DbContext(typeof(AppDbContext))]
[Migration("20260701192100_WeightSearchVectorSubjectBody")]
partial class WeightSearchVectorSubjectBody
{
/// <inheritdoc />
protected override void BuildTargetModel(ModelBuilder modelBuilder)
{
#pragma warning disable 612, 618
modelBuilder
.HasAnnotation("ProductVersion", "8.0.4")
.HasAnnotation("Relational:MaxIdentifierLength", 63);
NpgsqlModelBuilderExtensions.UseIdentityByDefaultColumns(modelBuilder);
modelBuilder.Entity("InboxIntel.Domain.Entities.AnalyticsAggregate", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<DateOnly>("Day")
.HasColumnType("date");
b.Property<string>("HourHistogramJson")
.HasColumnType("text");
b.Property<int>("NewsletterCount")
.HasColumnType("integer");
b.Property<int>("TotalReceived")
.HasColumnType("integer");
b.Property<long>("TotalSizeBytes")
.HasColumnType("bigint");
b.Property<int>("TotalUnread")
.HasColumnType("integer");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.Property<int>("WithAttachments")
.HasColumnType("integer");
b.HasKey("Id");
b.HasIndex("UserId", "Day")
.IsUnique();
b.ToTable("analytics_aggregates", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Attachment", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("EmailId")
.HasColumnType("uuid");
b.Property<string>("FileName")
.IsRequired()
.HasMaxLength(512)
.HasColumnType("character varying(512)");
b.Property<string>("GmailAttachmentId")
.HasColumnType("text");
b.Property<string>("MimeType")
.HasMaxLength(255)
.HasColumnType("character varying(255)");
b.Property<long>("SizeBytes")
.HasColumnType("bigint");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.HasKey("Id");
b.HasIndex("EmailId");
b.HasIndex("UserId", "MimeType");
b.ToTable("attachments", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Email", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<string>("BodyText")
.HasColumnType("text");
b.Property<int>("Category")
.HasColumnType("integer");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<string>("GmailMessageId")
.IsRequired()
.HasMaxLength(64)
.HasColumnType("character varying(64)");
b.Property<bool>("HasAttachments")
.HasColumnType("boolean");
b.Property<bool>("HasListUnsubscribe")
.HasColumnType("boolean");
b.Property<bool>("IsImportant")
.HasColumnType("boolean");
b.Property<bool>("IsInInbox")
.HasColumnType("boolean");
b.Property<bool>("IsStarred")
.HasColumnType("boolean");
b.Property<bool>("IsTrashed")
.HasColumnType("boolean");
b.Property<bool>("IsUnread")
.HasColumnType("boolean");
b.Property<string>("ListUnsubscribeRaw")
.HasMaxLength(2048)
.HasColumnType("character varying(2048)");
b.Property<DateTimeOffset?>("ReceivedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<NpgsqlTsVector>("SearchVector")
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("tsvector")
.HasComputedColumnSql("setweight(to_tsvector('english', coalesce(\"Subject\",'')), 'A') || setweight(to_tsvector('english', coalesce(\"BodyText\",'')), 'B')", true);
b.Property<Guid>("SenderId")
.HasColumnType("uuid");
b.Property<DateTimeOffset>("SentAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<long>("SizeEstimateBytes")
.HasColumnType("bigint");
b.Property<string>("Snippet")
.HasMaxLength(2048)
.HasColumnType("character varying(2048)");
b.Property<string>("Subject")
.HasMaxLength(1024)
.HasColumnType("character varying(1024)");
b.Property<bool>("SupportsOneClickUnsubscribe")
.HasColumnType("boolean");
b.Property<Guid>("ThreadId")
.HasColumnType("uuid");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.HasKey("Id");
b.HasIndex("SearchVector");
NpgsqlIndexBuilderExtensions.HasMethod(b.HasIndex("SearchVector"), "GIN");
b.HasIndex("SenderId");
b.HasIndex("ThreadId");
b.HasIndex("UserId", "Category");
b.HasIndex("UserId", "GmailMessageId")
.IsUnique();
b.HasIndex("UserId", "IsInInbox");
b.HasIndex("UserId", "IsUnread");
b.HasIndex("UserId", "SenderId");
b.HasIndex("UserId", "SentAtUtc");
b.ToTable("emails", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.EmailLabel", b =>
{
b.Property<Guid>("EmailId")
.HasColumnType("uuid");
b.Property<Guid>("LabelId")
.HasColumnType("uuid");
b.HasKey("EmailId", "LabelId");
b.HasIndex("LabelId");
b.ToTable("email_labels", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Label", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<string>("ColorHex")
.HasColumnType("text");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<string>("GmailLabelId")
.IsRequired()
.HasMaxLength(64)
.HasColumnType("character varying(64)");
b.Property<string>("Name")
.IsRequired()
.HasMaxLength(255)
.HasColumnType("character varying(255)");
b.Property<string>("Type")
.IsRequired()
.HasColumnType("text");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.HasKey("Id");
b.HasIndex("UserId", "GmailLabelId")
.IsUnique();
b.ToTable("labels", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.MailDomain", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<int>("EmailCount")
.HasColumnType("integer");
b.Property<bool>("IsBulkSender")
.HasColumnType("boolean");
b.Property<string>("Name")
.IsRequired()
.HasMaxLength(255)
.HasColumnType("character varying(255)");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.HasKey("Id");
b.HasIndex("UserId", "Name")
.IsUnique();
b.ToTable("domains", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.MailThread", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<DateTimeOffset?>("FirstMessageUtc")
.HasColumnType("timestamp with time zone");
b.Property<string>("GmailThreadId")
.IsRequired()
.HasMaxLength(64)
.HasColumnType("character varying(64)");
b.Property<DateTimeOffset?>("LastMessageUtc")
.HasColumnType("timestamp with time zone");
b.Property<int>("MessageCount")
.HasColumnType("integer");
b.Property<string>("Snippet")
.HasColumnType("text");
b.Property<string>("Subject")
.HasMaxLength(1024)
.HasColumnType("character varying(1024)");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.HasKey("Id");
b.HasIndex("UserId", "GmailThreadId")
.IsUnique();
b.ToTable("threads", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Sender", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<string>("Address")
.IsRequired()
.HasMaxLength(320)
.HasColumnType("character varying(320)");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<string>("DisplayName")
.HasMaxLength(255)
.HasColumnType("character varying(255)");
b.Property<Guid>("DomainId")
.HasColumnType("uuid");
b.Property<int>("EmailCount")
.HasColumnType("integer");
b.Property<bool>("HasUnsubscribe")
.HasColumnType("boolean");
b.Property<DateTimeOffset?>("LastReceivedUtc")
.HasColumnType("timestamp with time zone");
b.Property<long>("TotalSizeBytes")
.HasColumnType("bigint");
b.Property<int>("UnreadCount")
.HasColumnType("integer");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.HasKey("Id");
b.HasIndex("DomainId");
b.HasIndex("UserId", "Address")
.IsUnique();
b.HasIndex("UserId", "EmailCount");
b.ToTable("senders", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.SyncState", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<DateTimeOffset?>("CompletedUtc")
.HasColumnType("timestamp with time zone");
b.Property<int>("ConsecutiveFailures")
.HasColumnType("integer");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<string>("LastError")
.HasMaxLength(4000)
.HasColumnType("character varying(4000)");
b.Property<string>("LastHistoryId")
.HasColumnType("text");
b.Property<DateTimeOffset?>("LastSuccessfulSyncUtc")
.HasColumnType("timestamp with time zone");
b.Property<int>("LastSyncType")
.HasColumnType("integer");
b.Property<int>("MessagesProcessed")
.HasColumnType("integer");
b.Property<string>("ResumePageToken")
.HasColumnType("text");
b.Property<DateTimeOffset?>("StartedUtc")
.HasColumnType("timestamp with time zone");
b.Property<int>("Status")
.HasColumnType("integer");
b.Property<int>("TotalMessagesEstimate")
.HasColumnType("integer");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.HasKey("Id");
b.HasIndex("UserId")
.IsUnique();
b.ToTable("sync_states", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.UnsubscribeItem", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<double>("Confidence")
.HasColumnType("double precision");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<int>("EmailCount")
.HasColumnType("integer");
b.Property<DateTimeOffset?>("LastAttemptUtc")
.HasColumnType("timestamp with time zone");
b.Property<int>("Method")
.HasColumnType("integer");
b.Property<string>("ResultMessage")
.HasColumnType("text");
b.Property<Guid>("SenderId")
.HasColumnType("uuid");
b.Property<int>("Status")
.HasColumnType("integer");
b.Property<string>("UnsubscribeTarget")
.HasMaxLength(2048)
.HasColumnType("character varying(2048)");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.HasKey("Id");
b.HasIndex("SenderId");
b.HasIndex("UserId", "SenderId")
.IsUnique();
b.ToTable("unsubscribe_items", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.User", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<DateTimeOffset?>("AccessTokenExpiresAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<bool>("DigestEnabled")
.HasColumnType("boolean");
b.Property<string>("DisplayName")
.HasColumnType("text");
b.Property<string>("Email")
.IsRequired()
.HasMaxLength(320)
.HasColumnType("character varying(320)");
b.Property<byte[]>("EncryptedRefreshToken")
.HasColumnType("bytea");
b.Property<string>("GoogleSubjectId")
.IsRequired()
.HasMaxLength(64)
.HasColumnType("character varying(64)");
b.Property<DateTimeOffset?>("LastDigestSentUtc")
.HasColumnType("timestamp with time zone");
b.Property<DateTimeOffset?>("LastLoginUtc")
.HasColumnType("timestamp with time zone");
b.Property<string>("PictureUrl")
.HasColumnType("text");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.HasKey("Id");
b.HasIndex("Email")
.IsUnique();
b.HasIndex("GoogleSubjectId")
.IsUnique();
b.ToTable("users", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.WidgetLayout", b =>
{
b.Property<Guid>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("uuid");
b.Property<DateTimeOffset>("CreatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<int>("H")
.HasColumnType("integer");
b.Property<string>("SettingsJson")
.HasColumnType("text");
b.Property<int>("SortOrder")
.HasColumnType("integer");
b.Property<DateTimeOffset?>("UpdatedAtUtc")
.HasColumnType("timestamp with time zone");
b.Property<Guid>("UserId")
.HasColumnType("uuid");
b.Property<bool>("Visible")
.HasColumnType("boolean");
b.Property<int>("W")
.HasColumnType("integer");
b.Property<string>("WidgetKey")
.IsRequired()
.HasMaxLength(64)
.HasColumnType("character varying(64)");
b.Property<int>("X")
.HasColumnType("integer");
b.Property<int>("Y")
.HasColumnType("integer");
b.HasKey("Id");
b.HasIndex("UserId", "WidgetKey")
.IsUnique();
b.ToTable("widget_layouts", (string)null);
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Attachment", b =>
{
b.HasOne("InboxIntel.Domain.Entities.Email", "Email")
.WithMany("Attachments")
.HasForeignKey("EmailId")
.OnDelete(DeleteBehavior.Cascade)
.IsRequired();
b.Navigation("Email");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Email", b =>
{
b.HasOne("InboxIntel.Domain.Entities.Sender", "Sender")
.WithMany("Emails")
.HasForeignKey("SenderId")
.OnDelete(DeleteBehavior.Restrict)
.IsRequired();
b.HasOne("InboxIntel.Domain.Entities.MailThread", "Thread")
.WithMany("Emails")
.HasForeignKey("ThreadId")
.OnDelete(DeleteBehavior.Cascade)
.IsRequired();
b.HasOne("InboxIntel.Domain.Entities.User", null)
.WithMany("Emails")
.HasForeignKey("UserId")
.OnDelete(DeleteBehavior.Cascade)
.IsRequired();
b.Navigation("Sender");
b.Navigation("Thread");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.EmailLabel", b =>
{
b.HasOne("InboxIntel.Domain.Entities.Email", "Email")
.WithMany("EmailLabels")
.HasForeignKey("EmailId")
.OnDelete(DeleteBehavior.Cascade)
.IsRequired();
b.HasOne("InboxIntel.Domain.Entities.Label", "Label")
.WithMany("EmailLabels")
.HasForeignKey("LabelId")
.OnDelete(DeleteBehavior.Cascade)
.IsRequired();
b.Navigation("Email");
b.Navigation("Label");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Sender", b =>
{
b.HasOne("InboxIntel.Domain.Entities.MailDomain", "Domain")
.WithMany("Senders")
.HasForeignKey("DomainId")
.OnDelete(DeleteBehavior.Restrict)
.IsRequired();
b.Navigation("Domain");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.UnsubscribeItem", b =>
{
b.HasOne("InboxIntel.Domain.Entities.Sender", "Sender")
.WithMany()
.HasForeignKey("SenderId")
.OnDelete(DeleteBehavior.Cascade)
.IsRequired();
b.Navigation("Sender");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.WidgetLayout", b =>
{
b.HasOne("InboxIntel.Domain.Entities.User", null)
.WithMany("WidgetLayouts")
.HasForeignKey("UserId")
.OnDelete(DeleteBehavior.Cascade)
.IsRequired();
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Email", b =>
{
b.Navigation("Attachments");
b.Navigation("EmailLabels");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Label", b =>
{
b.Navigation("EmailLabels");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.MailDomain", b =>
{
b.Navigation("Senders");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.MailThread", b =>
{
b.Navigation("Emails");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.Sender", b =>
{
b.Navigation("Emails");
});
modelBuilder.Entity("InboxIntel.Domain.Entities.User", b =>
{
b.Navigation("Emails");
b.Navigation("WidgetLayouts");
});
#pragma warning restore 612, 618
}
}
}
@@ -0,0 +1,45 @@
using Microsoft.EntityFrameworkCore.Migrations;
using NpgsqlTypes;
#nullable disable
namespace InboxIntel.Infrastructure.Migrations
{
/// <inheritdoc />
public partial class WeightSearchVectorSubjectBody : Migration
{
/// <inheritdoc />
protected override void Up(MigrationBuilder migrationBuilder)
{
migrationBuilder.AlterColumn<NpgsqlTsVector>(
name: "SearchVector",
table: "emails",
type: "tsvector",
nullable: true,
computedColumnSql: "setweight(to_tsvector('english', coalesce(\"Subject\",'')), 'A') || setweight(to_tsvector('english', coalesce(\"BodyText\",'')), 'B')",
stored: true,
oldClrType: typeof(NpgsqlTsVector),
oldType: "tsvector",
oldNullable: true,
oldComputedColumnSql: "to_tsvector('english', coalesce(\"Subject\",'') || ' ' || coalesce(\"BodyText\",''))",
oldStored: true);
}
/// <inheritdoc />
protected override void Down(MigrationBuilder migrationBuilder)
{
migrationBuilder.AlterColumn<NpgsqlTsVector>(
name: "SearchVector",
table: "emails",
type: "tsvector",
nullable: true,
computedColumnSql: "to_tsvector('english', coalesce(\"Subject\",'') || ' ' || coalesce(\"BodyText\",''))",
stored: true,
oldClrType: typeof(NpgsqlTsVector),
oldType: "tsvector",
oldNullable: true,
oldComputedColumnSql: "setweight(to_tsvector('english', coalesce(\"Subject\",'')), 'A') || setweight(to_tsvector('english', coalesce(\"BodyText\",'')), 'B')",
oldStored: true);
}
}
}
@@ -160,7 +160,7 @@ namespace InboxIntel.Infrastructure.Migrations
b.Property<NpgsqlTsVector>("SearchVector")
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("tsvector")
.HasComputedColumnSql("to_tsvector('english', coalesce(\"Subject\",'') || ' ' || coalesce(\"BodyText\",''))", true);
.HasComputedColumnSql("setweight(to_tsvector('english', coalesce(\"Subject\",'')), 'A') || setweight(to_tsvector('english', coalesce(\"BodyText\",'')), 'B')", true);
b.Property<Guid>("SenderId")
.HasColumnType("uuid");
@@ -59,15 +59,17 @@ public class AppDbContext : DbContext, IAppDbContext
modelBuilder.Entity<UnsubscribeItem>().HasQueryFilter(e => CurrentUserId == Guid.Empty || e.UserId == CurrentUserId);
// PostgreSQL full-text search: generated tsvector over subject + body with a
// GIN index, maintained by the DB and read-only in code. The tsvector type is
// Npgsql-only, so map it only for relational providers and ignore it otherwise
// (e.g. the InMemory provider used by tests). Production behaviour is unchanged.
// GIN index, maintained by the DB and read-only in code. Subject is weighted 'A'
// and body 'B' so ts_rank_cd ranks a subject match above a body-only mention. The
// tsvector type is Npgsql-only, so map it only for relational providers and ignore
// it otherwise (e.g. the InMemory provider used by tests).
if (Database.IsRelational())
{
modelBuilder.Entity<Email>().Property(e => e.SearchVector)
.HasColumnType("tsvector")
.HasComputedColumnSql(
"to_tsvector('english', coalesce(\"Subject\",'') || ' ' || coalesce(\"BodyText\",''))",
"setweight(to_tsvector('english', coalesce(\"Subject\",'')), 'A') || " +
"setweight(to_tsvector('english', coalesce(\"BodyText\",'')), 'B')",
stored: true);
modelBuilder.Entity<Email>().HasIndex(e => e.SearchVector).HasMethod("GIN");
}