add WaybackMachine backup snapshot

ref #325
This commit is contained in:
Collin M. Barrett 2018-08-24 11:30:17 -05:00
parent 70e621a528
commit 209bd9a0b8
13 changed files with 842 additions and 18 deletions

View file

@ -16,4 +16,6 @@
<s:String x:Key="/Default/CodeStyle/Naming/CSharpNaming/PredefinedNamingRules/=PrivateInstanceFields/@EntryIndexedValue">&lt;Policy Inspect="True" Prefix="" Suffix="" Style="aaBb" /&gt;</s:String>
<s:String x:Key="/Default/CodeStyle/Naming/CSharpNaming/PredefinedNamingRules/=PrivateStaticFields/@EntryIndexedValue">&lt;Policy Inspect="True" Prefix="" Suffix="" Style="aaBb" /&gt;</s:String>
<s:Boolean x:Key="/Default/UserDictionary/Words/=appsettings/@EntryIndexedValue">True</s:Boolean>
<s:Boolean x:Key="/Default/UserDictionary/Words/=Wayback/@EntryIndexedValue">True</s:Boolean>
<s:Boolean x:Key="/Default/UserDictionary/Words/=Waybacks/@EntryIndexedValue">True</s:Boolean>
</wpf:ResourceDictionary>

View file

@ -0,0 +1,645 @@
// <auto-generated />
using System;
using FilterLists.Data;
using Microsoft.EntityFrameworkCore;
using Microsoft.EntityFrameworkCore.Infrastructure;
using Microsoft.EntityFrameworkCore.Metadata;
using Microsoft.EntityFrameworkCore.Migrations;
namespace FilterLists.Api.Migrations
{
[DbContext(typeof(FilterListsDbContext))]
[Migration("20180824155224_AddWaybackTimestamp")]
partial class AddWaybackTimestamp
{
protected override void BuildTargetModel(ModelBuilder modelBuilder)
{
#pragma warning disable 612, 618
modelBuilder
.HasAnnotation("ProductVersion", "2.1.2-rtm-30932")
.HasAnnotation("Relational:MaxIdentifierLength", 64);
modelBuilder.Entity("FilterLists.Data.Entities.FilterList", b =>
{
b.Property<ushort>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("SMALLINT UNSIGNED")
.HasAnnotation("MySql:ValueGenerationStrategy", MySqlValueGenerationStrategy.IdentityColumn);
b.Property<bool>("CantSnapshot");
b.Property<string>("ChatUrl")
.HasColumnType("TEXT");
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.Property<string>("Description")
.HasColumnType("TEXT");
b.Property<string>("DescriptionSourceUrl")
.HasColumnType("TEXT");
b.Property<DateTime?>("DiscontinuedDate")
.ValueGeneratedOnAdd()
.HasColumnType("DATE")
.HasDefaultValueSql("NULL");
b.Property<string>("DonateUrl")
.HasColumnType("TEXT");
b.Property<string>("EmailAddress")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(126)")
.HasDefaultValueSql("NULL");
b.Property<string>("ForumUrl")
.HasColumnType("TEXT");
b.Property<string>("HomeUrl")
.HasColumnType("TEXT");
b.Property<string>("IssuesUrl")
.HasColumnType("TEXT");
b.Property<byte>("LicenseId")
.ValueGeneratedOnAdd()
.HasDefaultValue((byte)5);
b.Property<DateTime?>("ModifiedDateUtc")
.IsRequired()
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp() ON UPDATE current_timestamp()");
b.Property<string>("Name")
.IsRequired()
.HasColumnType("VARCHAR(126)");
b.Property<string>("PolicyUrl")
.HasColumnType("TEXT");
b.Property<DateTime?>("PublishedDate")
.ValueGeneratedOnAdd()
.HasColumnType("DATE")
.HasDefaultValueSql("NULL");
b.Property<string>("SubmissionUrl")
.HasColumnType("TEXT");
b.Property<byte?>("SyntaxId");
b.Property<string>("ViewUrl")
.HasColumnType("TEXT");
b.HasKey("Id");
b.HasIndex("LicenseId");
b.HasIndex("SyntaxId");
b.ToTable("filterlists");
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.FilterListLanguage", b =>
{
b.Property<ushort>("FilterListId");
b.Property<ushort>("LanguageId");
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.HasKey("FilterListId", "LanguageId");
b.HasIndex("LanguageId");
b.ToTable("filterlists_languages");
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.FilterListMaintainer", b =>
{
b.Property<ushort>("FilterListId");
b.Property<ushort>("MaintainerId");
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.HasKey("FilterListId", "MaintainerId");
b.HasIndex("MaintainerId");
b.ToTable("filterlists_maintainers");
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.FilterListTag", b =>
{
b.Property<ushort>("FilterListId");
b.Property<byte>("TagId");
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.HasKey("FilterListId", "TagId");
b.HasIndex("TagId");
b.ToTable("filterlists_tags");
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.Fork", b =>
{
b.Property<ushort>("ForkFilterListId");
b.Property<ushort>("UpstreamFilterListId");
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.HasKey("ForkFilterListId", "UpstreamFilterListId");
b.HasIndex("UpstreamFilterListId");
b.ToTable("forks");
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.Merge", b =>
{
b.Property<ushort>("MergeFilterListId");
b.Property<ushort>("UpstreamFilterListId");
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.HasKey("MergeFilterListId", "UpstreamFilterListId");
b.HasIndex("UpstreamFilterListId");
b.ToTable("merges");
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.SnapshotRule", b =>
{
b.Property<uint>("SnapshotId");
b.Property<uint>("RuleId");
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.HasKey("SnapshotId", "RuleId");
b.HasIndex("RuleId");
b.ToTable("snapshots_rules");
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.SoftwareSyntax", b =>
{
b.Property<byte>("SyntaxId");
b.Property<byte>("SoftwareId");
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.HasKey("SyntaxId", "SoftwareId");
b.HasIndex("SoftwareId");
b.ToTable("software_syntaxes");
});
modelBuilder.Entity("FilterLists.Data.Entities.Language", b =>
{
b.Property<ushort>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("SMALLINT UNSIGNED")
.HasAnnotation("MySql:ValueGenerationStrategy", MySqlValueGenerationStrategy.IdentityColumn);
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.Property<string>("Iso6391")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(2)")
.HasDefaultValueSql("NULL");
b.Property<string>("Iso6392")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(3)")
.HasDefaultValueSql("NULL");
b.Property<string>("Iso6392B")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(3)")
.HasDefaultValueSql("NULL");
b.Property<string>("Iso6392T")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(3)")
.HasDefaultValueSql("NULL");
b.Property<string>("Iso6393")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(3)")
.HasDefaultValueSql("NULL");
b.Property<string>("LocalName")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(126)")
.HasDefaultValueSql("NULL");
b.Property<DateTime?>("ModifiedDateUtc")
.IsRequired()
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp() ON UPDATE current_timestamp()");
b.Property<string>("Name")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(126)")
.HasDefaultValueSql("NULL");
b.HasKey("Id");
b.ToTable("languages");
});
modelBuilder.Entity("FilterLists.Data.Entities.License", b =>
{
b.Property<byte>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("TINYINT UNSIGNED")
.HasAnnotation("MySql:ValueGenerationStrategy", MySqlValueGenerationStrategy.IdentityColumn);
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.Property<string>("DescriptionUrl")
.HasColumnType("TEXT");
b.Property<DateTime?>("ModifiedDateUtc")
.IsRequired()
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp() ON UPDATE current_timestamp()");
b.Property<string>("Name")
.IsRequired()
.HasColumnType("VARCHAR(126)");
b.Property<bool>("PermissiveAdaptation");
b.Property<bool>("PermissiveCommercial");
b.HasKey("Id");
b.ToTable("licenses");
});
modelBuilder.Entity("FilterLists.Data.Entities.Maintainer", b =>
{
b.Property<ushort>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("SMALLINT UNSIGNED")
.HasAnnotation("MySql:ValueGenerationStrategy", MySqlValueGenerationStrategy.IdentityColumn);
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.Property<string>("EmailAddress")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(126)")
.HasDefaultValueSql("NULL");
b.Property<string>("HomeUrl")
.HasColumnType("TEXT");
b.Property<DateTime?>("ModifiedDateUtc")
.IsRequired()
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp() ON UPDATE current_timestamp()");
b.Property<string>("Name")
.IsRequired()
.HasColumnType("VARCHAR(126)");
b.Property<string>("TwitterHandle")
.ValueGeneratedOnAdd()
.HasColumnType("VARCHAR(126)")
.HasDefaultValueSql("NULL");
b.HasKey("Id");
b.ToTable("maintainers");
});
modelBuilder.Entity("FilterLists.Data.Entities.Rule", b =>
{
b.Property<uint>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("INT UNSIGNED")
.HasAnnotation("MySql:ValueGenerationStrategy", MySqlValueGenerationStrategy.IdentityColumn);
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.Property<string>("Raw")
.IsRequired()
.HasColumnType("LONGTEXT");
b.HasKey("Id");
b.ToTable("rules");
});
modelBuilder.Entity("FilterLists.Data.Entities.Snapshot", b =>
{
b.Property<uint>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("MEDIUMINT UNSIGNED")
.HasAnnotation("MySql:ValueGenerationStrategy", MySqlValueGenerationStrategy.IdentityColumn);
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.Property<ushort>("FilterListId");
b.Property<ushort?>("HttpStatusCode")
.ValueGeneratedOnAdd()
.HasColumnType("SMALLINT UNSIGNED")
.HasDefaultValueSql("NULL");
b.Property<DateTime?>("ModifiedDateUtc")
.IsRequired()
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp() ON UPDATE current_timestamp()");
b.Property<bool>("WasSuccessful");
b.Property<DateTime?>("WaybackTimestamp")
.HasColumnType("TIMESTAMP");
b.HasKey("Id");
b.HasIndex("FilterListId");
b.ToTable("snapshots");
});
modelBuilder.Entity("FilterLists.Data.Entities.Software", b =>
{
b.Property<byte>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("TINYINT UNSIGNED")
.HasAnnotation("MySql:ValueGenerationStrategy", MySqlValueGenerationStrategy.IdentityColumn);
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.Property<string>("DownloadUrl")
.HasColumnType("TEXT");
b.Property<string>("HomeUrl")
.HasColumnType("TEXT");
b.Property<DateTime?>("ModifiedDateUtc")
.IsRequired()
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp() ON UPDATE current_timestamp()");
b.Property<string>("Name")
.IsRequired()
.HasColumnType("VARCHAR(126)");
b.HasKey("Id");
b.ToTable("software");
});
modelBuilder.Entity("FilterLists.Data.Entities.Syntax", b =>
{
b.Property<byte>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("TINYINT UNSIGNED")
.HasAnnotation("MySql:ValueGenerationStrategy", MySqlValueGenerationStrategy.IdentityColumn);
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.Property<string>("DefinitionUrl")
.HasColumnType("TEXT");
b.Property<DateTime?>("ModifiedDateUtc")
.IsRequired()
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp() ON UPDATE current_timestamp()");
b.Property<string>("Name")
.IsRequired()
.HasColumnType("VARCHAR(126)");
b.HasKey("Id");
b.ToTable("syntaxes");
});
modelBuilder.Entity("FilterLists.Data.Entities.Tag", b =>
{
b.Property<byte>("Id")
.ValueGeneratedOnAdd()
.HasColumnType("TINYINT UNSIGNED")
.HasAnnotation("MySql:ValueGenerationStrategy", MySqlValueGenerationStrategy.IdentityColumn);
b.Property<DateTime?>("CreatedDateUtc")
.IsRequired()
.ValueGeneratedOnAdd()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp()");
b.Property<string>("Description")
.HasColumnType("TEXT");
b.Property<DateTime?>("ModifiedDateUtc")
.IsRequired()
.ValueGeneratedOnAddOrUpdate()
.HasColumnType("TIMESTAMP")
.HasDefaultValueSql("current_timestamp() ON UPDATE current_timestamp()");
b.Property<string>("Name")
.IsRequired()
.HasColumnType("VARCHAR(126)");
b.HasKey("Id");
b.ToTable("tags");
});
modelBuilder.Entity("FilterLists.Data.Entities.FilterList", b =>
{
b.HasOne("FilterLists.Data.Entities.License", "License")
.WithMany("FilterLists")
.HasForeignKey("LicenseId")
.OnDelete(DeleteBehavior.Cascade);
b.HasOne("FilterLists.Data.Entities.Syntax", "Syntax")
.WithMany("FilterLists")
.HasForeignKey("SyntaxId");
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.FilterListLanguage", b =>
{
b.HasOne("FilterLists.Data.Entities.FilterList", "FilterList")
.WithMany("FilterListLanguages")
.HasForeignKey("FilterListId")
.OnDelete(DeleteBehavior.Cascade);
b.HasOne("FilterLists.Data.Entities.Language", "Language")
.WithMany("FilterListLanguages")
.HasForeignKey("LanguageId")
.OnDelete(DeleteBehavior.Cascade);
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.FilterListMaintainer", b =>
{
b.HasOne("FilterLists.Data.Entities.FilterList", "FilterList")
.WithMany("FilterListMaintainers")
.HasForeignKey("FilterListId")
.OnDelete(DeleteBehavior.Cascade);
b.HasOne("FilterLists.Data.Entities.Maintainer", "Maintainer")
.WithMany("FilterListMaintainers")
.HasForeignKey("MaintainerId")
.OnDelete(DeleteBehavior.Cascade);
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.FilterListTag", b =>
{
b.HasOne("FilterLists.Data.Entities.FilterList", "FilterList")
.WithMany("FilterListTags")
.HasForeignKey("FilterListId")
.OnDelete(DeleteBehavior.Cascade);
b.HasOne("FilterLists.Data.Entities.Tag", "Tag")
.WithMany("FilterListTags")
.HasForeignKey("TagId")
.OnDelete(DeleteBehavior.Cascade);
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.Fork", b =>
{
b.HasOne("FilterLists.Data.Entities.FilterList", "ForkFilterList")
.WithMany("ForkFilterLists")
.HasForeignKey("ForkFilterListId")
.OnDelete(DeleteBehavior.Cascade);
b.HasOne("FilterLists.Data.Entities.FilterList", "UpstreamFilterList")
.WithMany("UpstreamForkFilterLists")
.HasForeignKey("UpstreamFilterListId")
.OnDelete(DeleteBehavior.Cascade);
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.Merge", b =>
{
b.HasOne("FilterLists.Data.Entities.FilterList", "MergeFilterList")
.WithMany("MergeFilterLists")
.HasForeignKey("MergeFilterListId")
.OnDelete(DeleteBehavior.Cascade);
b.HasOne("FilterLists.Data.Entities.FilterList", "UpstreamFilterList")
.WithMany("UpstreamMergeFilterLists")
.HasForeignKey("UpstreamFilterListId")
.OnDelete(DeleteBehavior.Cascade);
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.SnapshotRule", b =>
{
b.HasOne("FilterLists.Data.Entities.Rule", "Rule")
.WithMany("SnapshotRules")
.HasForeignKey("RuleId")
.OnDelete(DeleteBehavior.Cascade);
b.HasOne("FilterLists.Data.Entities.Snapshot", "Snapshot")
.WithMany("SnapshotRules")
.HasForeignKey("SnapshotId")
.OnDelete(DeleteBehavior.Cascade);
});
modelBuilder.Entity("FilterLists.Data.Entities.Junctions.SoftwareSyntax", b =>
{
b.HasOne("FilterLists.Data.Entities.Software", "Software")
.WithMany("SoftwareSyntaxes")
.HasForeignKey("SoftwareId")
.OnDelete(DeleteBehavior.Cascade);
b.HasOne("FilterLists.Data.Entities.Syntax", "Syntax")
.WithMany("SoftwareSyntaxes")
.HasForeignKey("SyntaxId")
.OnDelete(DeleteBehavior.Cascade);
});
modelBuilder.Entity("FilterLists.Data.Entities.Snapshot", b =>
{
b.HasOne("FilterLists.Data.Entities.FilterList", "FilterList")
.WithMany("Snapshots")
.HasForeignKey("FilterListId")
.OnDelete(DeleteBehavior.Cascade);
});
#pragma warning restore 612, 618
}
}
}

View file

@ -0,0 +1,24 @@
using System;
using Microsoft.EntityFrameworkCore.Migrations;
namespace FilterLists.Api.Migrations
{
public partial class AddWaybackTimestamp : Migration
{
protected override void Up(MigrationBuilder migrationBuilder)
{
migrationBuilder.AddColumn<DateTime>(
"WaybackTimestamp",
"snapshots",
"TIMESTAMP",
nullable: true);
}
protected override void Down(MigrationBuilder migrationBuilder)
{
migrationBuilder.DropColumn(
"WaybackTimestamp",
"snapshots");
}
}
}

View file

@ -421,6 +421,9 @@ protected override void BuildModel(ModelBuilder modelBuilder)
b.Property<bool>("WasSuccessful");
b.Property<DateTime?>("WaybackTimestamp")
.HasColumnType("TIMESTAMP");
b.HasKey("Id");
b.HasIndex("FilterListId");

View file

@ -1,4 +1,5 @@
using System.Collections.Generic;
using System;
using System.Collections.Generic;
using FilterLists.Data.Entities.Junctions;
namespace FilterLists.Data.Entities
@ -10,5 +11,6 @@ public class Snapshot : BaseEntity
public uint? HttpStatusCode { get; set; }
public bool WasSuccessful { get; set; }
public ICollection<SnapshotRule> SnapshotRules { get; set; }
public DateTime? WaybackTimestamp { get; set; }
}
}

View file

@ -15,6 +15,8 @@ public override void Configure(EntityTypeBuilder<Snapshot> entityTypeBuilder)
entityTypeBuilder.Property(x => x.HttpStatusCode)
.HasColumnType("SMALLINT UNSIGNED")
.HasDefaultValueSql("NULL");
entityTypeBuilder.Property(x => x.WaybackTimestamp)
.HasColumnType("TIMESTAMP");
}
}
}

View file

@ -8,5 +8,8 @@ public static class StringExtensions
public static bool IsValidHttpOrHttpsUrl(this string source) =>
Uri.TryCreate(source, UriKind.Absolute, out var uriResult) &&
new[] {Uri.UriSchemeHttps, Uri.UriSchemeHttp}.Contains(uriResult.Scheme);
public static int GetNthIndexOfChar(this string s, int n, char t) =>
s.TakeWhile(c => (n -= c == t ? 1 : 0) > 0).Count();
}
}

View file

@ -0,0 +1,12 @@
using JetBrains.Annotations;
using Newtonsoft.Json;
namespace FilterLists.Services.Snapshot.Models.Wayback
{
[UsedImplicitly]
public class ArchivedSnapshots
{
[JsonProperty("closest")]
public Closest Closest { get; set; }
}
}

View file

@ -0,0 +1,15 @@
using JetBrains.Annotations;
using Newtonsoft.Json;
namespace FilterLists.Services.Snapshot.Models.Wayback
{
[UsedImplicitly]
public class Availability
{
[JsonProperty("url")]
public string Url { get; set; }
[JsonProperty("archived_snapshots")]
public ArchivedSnapshots ArchivedSnapshots { get; set; }
}
}

View file

@ -0,0 +1,21 @@
using JetBrains.Annotations;
using Newtonsoft.Json;
namespace FilterLists.Services.Snapshot.Models.Wayback
{
[UsedImplicitly]
public class Closest
{
[JsonProperty("status")]
public string Status { get; set; }
[JsonProperty("available")]
public bool Available { get; set; }
[JsonProperty("url")]
public string Url { get; set; }
[JsonProperty("timestamp")]
public string Timestamp { get; set; }
}
}

View file

@ -9,6 +9,7 @@
using FilterLists.Data;
using FilterLists.Services.Extensions;
using FilterLists.Services.Snapshot.Models;
using JetBrains.Annotations;
using Microsoft.ApplicationInsights;
using MoreLinq;
@ -18,25 +19,36 @@ public class Snapshot
{
private const int BatchSize = 500;
private readonly FilterListsDbContext dbContext;
private readonly string listViewUrl;
private readonly Data.Entities.Snapshot snapEntity;
public readonly FilterListViewUrlDto List;
protected readonly Data.Entities.Snapshot SnapEntity;
private readonly TelemetryClient telemetryClient;
private readonly string uaString;
private HashSet<string> lines;
protected string ListUrl;
public Snapshot()
{
}
[UsedImplicitly]
public Snapshot(FilterListsDbContext dbContext, FilterListViewUrlDto list, string uaString)
{
this.dbContext = dbContext;
listViewUrl = list.ViewUrl;
snapEntity = new Data.Entities.Snapshot {FilterListId = list.Id};
List = list;
ListUrl = list.ViewUrl;
SnapEntity = new Data.Entities.Snapshot {FilterListId = list.Id};
this.uaString = uaString;
telemetryClient = new TelemetryClient();
}
public async Task TrySaveAsync()
public bool WasSuccessful => SnapEntity.WasSuccessful;
public virtual async Task TrySaveAsync() => await TrySaveAsyncBase();
protected async Task TrySaveAsyncBase()
{
await AddSnapEntity();
if (!listViewUrl.IsValidHttpOrHttpsUrl())
if (!ListUrl.IsValidHttpOrHttpsUrl())
return;
try
{
@ -50,7 +62,7 @@ public async Task TrySaveAsync()
private async Task AddSnapEntity()
{
dbContext.Snapshots.Add(snapEntity);
dbContext.Snapshots.Add(SnapEntity);
await dbContext.SaveChangesAsync();
}
@ -77,7 +89,7 @@ private async Task TryGetLines()
}
catch (WebException we)
{
snapEntity.HttpStatusCode = (uint)((HttpWebResponse)we.Response).StatusCode;
SnapEntity.HttpStatusCode = (uint)((HttpWebResponse)we.Response).StatusCode;
await dbContext.SaveChangesAsync();
TrackException(we);
}
@ -88,8 +100,8 @@ private async Task GetLines()
using (var httpClient = new HttpClient())
{
httpClient.DefaultRequestHeaders.UserAgent.ParseAdd(uaString);
var response = await httpClient.GetAsync(listViewUrl, HttpCompletionOption.ResponseHeadersRead);
snapEntity.HttpStatusCode = (uint)response.StatusCode;
var response = await httpClient.GetAsync(ListUrl, HttpCompletionOption.ResponseHeadersRead);
SnapEntity.HttpStatusCode = (uint)response.StatusCode;
response.EnsureSuccessStatusCode();
lines = new HashSet<string>();
using (var stream = await response.Content.ReadAsStreamAsync())
@ -109,7 +121,7 @@ private async Task SaveInBatches()
}
private IEnumerable<SnapshotBatch> CreateBatches() =>
lines.Batch(BatchSize).Select(b => new SnapshotBatch(dbContext, b, snapEntity));
lines.Batch(BatchSize).Select(b => new SnapshotBatch(dbContext, b, SnapEntity));
private static async Task SaveBatches(IEnumerable<SnapshotBatch> batches)
{
@ -119,7 +131,7 @@ private static async Task SaveBatches(IEnumerable<SnapshotBatch> batches)
private async Task SetSuccessful()
{
snapEntity.WasSuccessful = true;
SnapEntity.WasSuccessful = true;
await dbContext.SaveChangesAsync();
}

View file

@ -13,6 +13,7 @@ namespace FilterLists.Services.Snapshot
public class SnapshotService : Service
{
private readonly DateTime yesterday = DateTime.UtcNow.AddDays(-1);
private string uaString;
public SnapshotService(FilterListsDbContext dbContext, IConfigurationProvider mapConfig)
: base(dbContext, mapConfig)
@ -21,10 +22,11 @@ public SnapshotService(FilterListsDbContext dbContext, IConfigurationProvider ma
public async Task CaptureAsync(int batchSize)
{
uaString = await UserAgentService.GetMostPopularString();
var lists = await GetListsToCapture(batchSize);
var uaString = await UserAgentService.GetMostPopularString();
var snaps = CreateSnaps(lists, uaString);
await SaveSnaps(snaps);
var snaps = await CreateAndSaveSnaps<Snapshot>(lists);
var listsToRetry = snaps.Where(s => !s.WasSuccessful).Select(s => s.List);
await CreateAndSaveSnaps<SnapshotWayback>(listsToRetry);
}
private async Task<IEnumerable<FilterListViewUrlDto>> GetListsToCapture(int batchSize) =>
@ -49,8 +51,17 @@ await DbContext
.ProjectTo<FilterListViewUrlDto>(MapConfig)
.ToListAsync();
private IEnumerable<Snapshot> CreateSnaps(IEnumerable<FilterListViewUrlDto> lists, string uaString) =>
lists.Select(l => new Snapshot(DbContext, l, uaString));
private async Task<List<TSnap>> CreateAndSaveSnaps<TSnap>(IEnumerable<FilterListViewUrlDto> lists)
where TSnap : Snapshot, new()
{
var snaps = CreateSnaps<TSnap>(lists).ToList();
await SaveSnaps(snaps);
return snaps;
}
private IEnumerable<TSnap> CreateSnaps<TSnap>(IEnumerable<FilterListViewUrlDto> lists)
where TSnap : Snapshot, new() =>
lists.Select(l => Activator.CreateInstance(typeof(TSnap), DbContext, l, uaString) as TSnap);
private static async Task SaveSnaps(IEnumerable<Snapshot> snaps)
{

View file

@ -0,0 +1,72 @@
using System;
using System.Net.Http;
using System.Threading.Tasks;
using FilterLists.Data;
using FilterLists.Services.Extensions;
using FilterLists.Services.Snapshot.Models;
using FilterLists.Services.Snapshot.Models.Wayback;
using JetBrains.Annotations;
using Newtonsoft.Json;
namespace FilterLists.Services.Snapshot
{
public class SnapshotWayback : Snapshot
{
private const string WaybackAvailabilityApiUrlPrefix = "https://archive.org/wayback/available?url=";
public SnapshotWayback()
{
}
[UsedImplicitly]
public SnapshotWayback(FilterListsDbContext dbContext, FilterListViewUrlDto list, string uaString)
: base(dbContext, list, uaString)
{
}
public override async Task TrySaveAsync()
{
await UpdateWaybackData();
await TrySaveAsyncBase();
}
private async Task UpdateWaybackData()
{
var availability = await GetWaybackAvailability();
var closest = availability.ArchivedSnapshots?.Closest;
if (closest != null)
{
UpdateViewUrl(closest);
UpdateTimestamp(closest);
}
else
{
ListUrl = null;
}
}
private async Task<Availability> GetWaybackAvailability()
{
using (var client = new HttpClient())
{
var url = WaybackAvailabilityApiUrlPrefix + ListUrl;
var response = await client.GetAsync(url, HttpCompletionOption.ResponseHeadersRead);
var jsonString = await response.Content.ReadAsStringAsync();
return JsonConvert.DeserializeObject<Availability>(jsonString);
}
}
private void UpdateViewUrl(Closest closest)
{
var waybackUrl = closest.Url;
var indexOfUrlModification = waybackUrl.GetNthIndexOfChar(5, '/');
const string urlModification = "if_";
ListUrl = waybackUrl.Substring(0, indexOfUrlModification) + urlModification +
waybackUrl.Substring(indexOfUrlModification);
}
private void UpdateTimestamp(Closest closest) =>
SnapEntity.WaybackTimestamp =
DateTimeOffset.ParseExact(closest.Timestamp, "yyyyMMddHHmmss", null).UtcDateTime;
}
}