feat(news): add scraper adapters, article deduplication, blocklist service, and remove tracked publish artifacts

2026-08-24 21:35:33 +02:00
parent 44b161d509
commit 8112598602
400 changed files with 1812 additions and 113367 deletions
@@ -0,0 +1,213 @@
+// <auto-generated />
+using System;
+using FinlyticNews.Database;
+using Microsoft.EntityFrameworkCore;
+using Microsoft.EntityFrameworkCore.Infrastructure;
+using Microsoft.EntityFrameworkCore.Migrations;
+using Microsoft.EntityFrameworkCore.Storage.ValueConversion;
+using Npgsql.EntityFrameworkCore.PostgreSQL.Metadata;
+
+#nullable disable
+
+namespace FinlyticNews.Migrations
+{
+    [DbContext(typeof(NewsDbContext))]
+    [Migration("20260818194224_Init")]
+    partial class Init
+    {
+        /// <inheritdoc />
+        protected override void BuildTargetModel(ModelBuilder modelBuilder)
+        {
+#pragma warning disable 612, 618
+            modelBuilder
+                .HasAnnotation("ProductVersion", "10.0.9")
+                .HasAnnotation("Relational:MaxIdentifierLength", 63);
+
+            NpgsqlModelBuilderExtensions.UseIdentityByDefaultColumns(modelBuilder);
+
+            modelBuilder.Entity("FinlyticCore.Entities.Settings.SettingEntity", b =>
+                {
+                    b.Property<Guid>("Id")
+                        .ValueGeneratedOnAdd()
+                        .HasColumnType("uuid");
+
+                    b.Property<string>("Key")
+                        .IsRequired()
+                        .HasMaxLength(150)
+                        .HasColumnType("character varying(150)");
+
+                    b.Property<DateTime>("LastUpdatedUtc")
+                        .HasColumnType("timestamp with time zone");
+
+                    b.Property<string>("ServiceIdentifier")
+                        .IsRequired()
+                        .HasMaxLength(100)
+                        .HasColumnType("character varying(100)");
+
+                    b.Property<string>("ValueJson")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.HasKey("Id");
+
+                    b.HasIndex("Key")
+                        .IsUnique();
+
+                    b.ToTable("DynamicSettings");
+                });
+
+            modelBuilder.Entity("FinlyticNews.Entities.ArticleSourceEntity", b =>
+                {
+                    b.Property<Guid>("Id")
+                        .ValueGeneratedOnAdd()
+                        .HasColumnType("uuid");
+
+                    b.Property<string>("Name")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.Property<string>("Source")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.Property<string>("Type")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.HasKey("Id");
+
+                    b.ToTable("ArticleSources");
+                });
+
+            modelBuilder.Entity("FinlyticNews.Entities.BlockedNewsUrlEntity", b =>
+                {
+                    b.Property<Guid>("Id")
+                        .ValueGeneratedOnAdd()
+                        .HasColumnType("uuid");
+
+                    b.Property<DateTime>("BlockedAtUtc")
+                        .HasColumnType("timestamp with time zone");
+
+                    b.Property<string>("Reason")
+                        .IsRequired()
+                        .HasMaxLength(256)
+                        .HasColumnType("character varying(256)");
+
+                    b.Property<string>("Url")
+                        .IsRequired()
+                        .HasMaxLength(2048)
+                        .HasColumnType("character varying(2048)");
+
+                    b.HasKey("Id");
+
+                    b.HasIndex("Url")
+                        .IsUnique();
+
+                    b.ToTable("BlockedNewsUrls");
+                });
+
+            modelBuilder.Entity("FinlyticNews.Entities.MatchedAssetEntity", b =>
+                {
+                    b.Property<Guid>("Id")
+                        .ValueGeneratedOnAdd()
+                        .HasColumnType("uuid");
+
+                    b.Property<string>("Isin")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.Property<string>("Name")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.Property<Guid>("NewsArticleId")
+                        .HasColumnType("uuid");
+
+                    b.HasKey("Id");
+
+                    b.HasIndex("Isin");
+
+                    b.HasIndex("NewsArticleId");
+
+                    b.ToTable("MatchedAssets");
+                });
+
+            modelBuilder.Entity("FinlyticNews.Entities.NewsArticleEntity", b =>
+                {
+                    b.Property<Guid>("Id")
+                        .ValueGeneratedOnAdd()
+                        .HasColumnType("uuid");
+
+                    b.Property<string>("Author")
+                        .HasColumnType("text");
+
+                    b.Property<string>("ContentRaw")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.Property<string>("Language")
+                        .HasColumnType("text");
+
+                    b.Property<DateTime>("PublishedAt")
+                        .HasColumnType("timestamp with time zone");
+
+                    b.Property<DateTime>("ScrapedAt")
+                        .HasColumnType("timestamp with time zone");
+
+                    b.Property<long?>("SimHash")
+                        .HasColumnType("bigint");
+
+                    b.Property<string>("SourceUrl")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.Property<string>("Status")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.Property<string>("Summary")
+                        .HasColumnType("text");
+
+                    b.Property<string>("Title")
+                        .IsRequired()
+                        .HasColumnType("text");
+
+                    b.Property<string>("TitleHash")
+                        .HasMaxLength(64)
+                        .HasColumnType("character varying(64)");
+
+                    b.HasKey("Id");
+
+                    b.HasIndex("SimHash");
+
+                    b.HasIndex("SourceUrl")
+                        .IsUnique();
+
+                    b.HasIndex("Status");
+
+                    b.HasIndex("TitleHash");
+
+                    b.HasIndex("PublishedAt", "ScrapedAt");
+
+                    b.ToTable("NewsArticles");
+                });
+
+            modelBuilder.Entity("FinlyticNews.Entities.MatchedAssetEntity", b =>
+                {
+                    b.HasOne("FinlyticNews.Entities.NewsArticleEntity", "NewsArticle")
+                        .WithMany("MatchedAssets")
+                        .HasForeignKey("NewsArticleId")
+                        .OnDelete(DeleteBehavior.Cascade)
+                        .IsRequired();
+
+                    b.Navigation("NewsArticle");
+                });
+
+            modelBuilder.Entity("FinlyticNews.Entities.NewsArticleEntity", b =>
+                {
+                    b.Navigation("MatchedAssets");
+                });
+#pragma warning restore 612, 618
+        }
+    }
+}