From 4ccfaad15fb487834cfb8dd6cad321efeaf83e55 Mon Sep 17 00:00:00 2001
From: Gervasio Marchand <github@gervas.io>
Date: Sat, 17 Dec 2022 17:27:04 -0800
Subject: [PATCH] Make it possible to scrape followed tags from the instance
 (#6)

---
 src/Config.cs                   | 16 +++++++++++++---
 src/GetMoarFediverse.csproj     |  4 ++++
 src/MastodonConnectionHelper.cs | 23 +++++++++++++++++++++++
 src/Program.cs                  | 23 +++++++++++++++++------
 4 files changed, 57 insertions(+), 9 deletions(-)
 create mode 100644 src/MastodonConnectionHelper.cs

diff --git a/src/Config.cs b/src/Config.cs
index 4ac1dcb..c499880 100644
--- a/src/Config.cs
+++ b/src/Config.cs
@@ -10,15 +10,18 @@ public class Config
     public string ImportedPath { get; }
     public string FakeRelayUrl { get; }
     public string FakeRelayApiKey { get; }
+    public string? MastodonPostgresConnectionString { get; }
     public ImmutableArray<string> Tags { get; }
     public ImmutableArray<SiteData> Sites { get; }
 
 
-    private Config(string importedPath, string fakeRelayUrl, string fakeRelayApiKey, ImmutableArray<string> tags, ImmutableArray<SiteData> sites)
+    private Config(string importedPath, string fakeRelayUrl, string fakeRelayApiKey, string? mastodonPostgresConnectionString,
+                    ImmutableArray<string> tags, ImmutableArray<SiteData> sites)
     {
         ImportedPath = importedPath;
         FakeRelayUrl = fakeRelayUrl;
         FakeRelayApiKey = fakeRelayApiKey;
+        MastodonPostgresConnectionString = mastodonPostgresConnectionString;
         Tags = tags;
         Sites = sites;
     }
@@ -30,20 +33,27 @@ public class Config
             return;
         }
 
-        var data = JsonSerializer.Deserialize<ConfigData>(File.ReadAllText(path), JsonContext.Default.ConfigData);
+        var data = JsonSerializer.Deserialize(File.ReadAllText(path), JsonContext.Default.ConfigData);
         
         var importedPath = Path.Join(Path.GetDirectoryName(path), "imported.txt");
         var apiKey = string.IsNullOrEmpty(data.FakeRelayApiKey)
             ? Environment.GetEnvironmentVariable("FAKERELAY_APIKEY")
             : data.FakeRelayApiKey;
 
-        Instance = new Config(importedPath, data.FakeRelayUrl, apiKey, data.Tags.ToImmutableArray(), data.ImmutableSites);
+        if (apiKey == null)
+        {
+            throw new Exception("The api key is missing");
+        }
+
+        Instance = new Config(importedPath, data.FakeRelayUrl, apiKey, data.MastodonPostgresConnectionString,
+            data.Tags.ToImmutableArray(), data.ImmutableSites);
     }
 
     public class ConfigData
     {
         public string FakeRelayUrl { get; set; }
         public string? FakeRelayApiKey { get; set; }
+        public string? MastodonPostgresConnectionString { get; set; }
         public string[] Tags { get; set; }
         public InternalSiteData[]? Sites { get; set; }
 
diff --git a/src/GetMoarFediverse.csproj b/src/GetMoarFediverse.csproj
index 40c60dd..3221ac5 100644
--- a/src/GetMoarFediverse.csproj
+++ b/src/GetMoarFediverse.csproj
@@ -7,4 +7,8 @@
     <Nullable>enable</Nullable>
   </PropertyGroup>
 
+  <ItemGroup>
+    <PackageReference Include="Npgsql" Version="7.0.1" />
+  </ItemGroup>
+
 </Project>
diff --git a/src/MastodonConnectionHelper.cs b/src/MastodonConnectionHelper.cs
new file mode 100644
index 0000000..2eab0e8
--- /dev/null
+++ b/src/MastodonConnectionHelper.cs
@@ -0,0 +1,23 @@
+using Npgsql;
+
+namespace GetMoarFediverse;
+
+public static class MastodonConnectionHelper
+{
+    public static async Task<List<string>> GetFollowedTagsAsync()
+    {
+        var res = new List<string>();
+        
+        await using var conn = new NpgsqlConnection(Config.Instance.MastodonPostgresConnectionString);
+        await conn.OpenAsync();
+
+        await using (var cmd = new NpgsqlCommand("SELECT DISTINCT tags.name FROM tag_follows JOIN tags ON tag_id = tags.id ORDER BY tags.name ASC;", conn))
+        await using (var reader = await cmd.ExecuteReaderAsync())
+        {
+            while (await reader.ReadAsync())
+                res.Add(reader.GetString(0));
+        }
+
+        return res;
+    }
+}
diff --git a/src/Program.cs b/src/Program.cs
index 6325b72..a8057da 100644
--- a/src/Program.cs
+++ b/src/Program.cs
@@ -26,11 +26,22 @@ var importedList = File.ReadAllLines(importedPath).ToList();
 var imported = importedList.ToHashSet();
 var statusesToLoadBag = new ConcurrentBag<string>();
 
-var sitesTags = Config.Instance.Sites
-    .SelectMany(s => Config.Instance.Tags.Select(tag => (s.Host, tag)))
-    .Concat(Config.Instance.Sites.SelectMany(s => s.SiteSpecificTags.Select(tag => (s.Host, tag))))
-    .OrderBy(t => t.tag)
-    .ToList();
+List<(string host, string tag)> sitesTags;
+if (string.IsNullOrEmpty(Config.Instance.MastodonPostgresConnectionString))
+{
+    sitesTags = Config.Instance.Sites
+        .SelectMany(s => Config.Instance.Tags.Select(tag => (s.Host, tag)))
+        .Concat(Config.Instance.Sites.SelectMany(s => s.SiteSpecificTags.Select(tag => (s.Host, tag))))
+        .OrderBy(t => t.tag)
+        .ToList();
+}
+else
+{
+    var tags = await MastodonConnectionHelper.GetFollowedTagsAsync();
+    sitesTags = Config.Instance.Sites
+        .SelectMany(s => tags.Select(t => (s.Host, t)))
+        .ToList();
+}
 
 ParallelOptions parallelOptions = new()
 {
@@ -54,7 +65,7 @@ await Parallel.ForEachAsync(sitesTags, parallelOptions, async (st, _) =>
     }
 
     var json = await response.Content.ReadAsStringAsync();
-    var data = JsonSerializer.Deserialize<TagResponse>(json, CamelCaseJsonContext.Default.TagResponse);
+    var data = JsonSerializer.Deserialize(json, CamelCaseJsonContext.Default.TagResponse);
 
     foreach (var statusLink in data.OrderedItems.Where(i=>!imported.Contains(i)))
     {