main

media-worker/Services/MinimaxTtsService.cs

@@ -0,0 +1,132 @@
+using System.Net.Http.Headers;
+using System.Text;
+using System.Text.Json;
+using Microsoft.Extensions.Logging;
+using Microsoft.Extensions.Options;
+using SaasMediaWorker.Configuration;
+using SaasMediaWorker.Models;
+
+namespace SaasMediaWorker.Services;
+
+/// <summary>
+/// Minimax TTS API Client — Metin → Ses dönüşümü.
+/// </summary>
+public class MinimaxTtsService
+{
+    private readonly HttpClient _httpClient;
+    private readonly ILogger<MinimaxTtsService> _logger;
+    private readonly ApiSettings _settings;
+
+    public MinimaxTtsService(
+        HttpClient httpClient,
+        ILogger<MinimaxTtsService> logger,
+        IOptions<ApiSettings> settings)
+    {
+        _httpClient = httpClient;
+        _logger = logger;
+        _settings = settings.Value;
+
+        _httpClient.BaseAddress = new Uri("https://api.minimax.chat/v1/");
+        _httpClient.DefaultRequestHeaders.Authorization = new AuthenticationHeaderValue("Bearer", _settings.MinimaxApiKey);
+        _httpClient.Timeout = TimeSpan.FromMinutes(2);
+    }
+
+    /// <summary>
+    /// Bir sahnenin narration metnini sese çevirir ve dosyaya kaydeder.
+    /// </summary>
+    public async Task<GeneratedMediaFile> GenerateNarrationAsync(
+        ScenePayload scene,
+        string outputDirectory,
+        string voiceStyle,
+        CancellationToken ct)
+    {
+        _logger.LogInformation(
+            "🎙️ Minimax TTS üretimi — Sahne {Order}: \"{Text}\"",
+            scene.Order,
+            scene.NarrationText[..Math.Min(60, scene.NarrationText.Length)]);
+
+        // Varsayılan voiceStyle kullan veya fallback
+        var voiceId = string.IsNullOrWhiteSpace(voiceStyle) ? _settings.MinimaxTtsVoiceId : voiceStyle;
+
+        var requestBody = new
+        {
+            model = "speech-01-turbo",
+            text = scene.NarrationText,
+            voice_setting = new
+            {
+                voice_id = voiceId,
+                speed = 1.0,
+                vol = 1.0,
+                pitch = 0
+            },
+            audio_setting = new
+            {
+                sample_rate = 32000,
+                bitrate = 128000,
+                format = "mp3",
+                channel = 1
+            }
+        };
+
+        var content = new StringContent(
+            JsonSerializer.Serialize(requestBody),
+            Encoding.UTF8,
+            "application/json");
+
+        var response = await _httpClient.PostAsync("t2a_v2", content, ct);
+
+        response.EnsureSuccessStatusCode();
+
+        // Minimax T2A V2 returns JSON with data.audio containing hex string
+        var responseString = await response.Content.ReadAsStringAsync(ct);
+        var jsonResponse = JsonSerializer.Deserialize<JsonElement>(responseString);
+        
+        if (jsonResponse.TryGetProperty("data", out var dataElement) && dataElement.TryGetProperty("audio", out var audioHex))
+        {
+            var hexString = audioHex.GetString() ?? "";
+            byte[] audioBytes = ConvertHexStringToByteArray(hexString);
+            
+            var outputPath = Path.Combine(outputDirectory, $"scene_{scene.Order:D2}_narration.mp3");
+            await File.WriteAllBytesAsync(outputPath, audioBytes, ct);
+
+            var fileInfo = new FileInfo(outputPath);
+
+            _logger.LogInformation(
+                "Minimax TTS tamamlandı — Sahne {Order}: {Size} bytes",
+                scene.Order, fileInfo.Length);
+
+            return new GeneratedMediaFile
+            {
+                SceneId = scene.Id,
+                SceneOrder = scene.Order,
+                Type = MediaFileType.AudioNarration,
+                LocalPath = outputPath,
+                FileSizeBytes = fileInfo.Length,
+                DurationSeconds = scene.Duration,
+                MimeType = "audio/mpeg",
+                AiProvider = "minimax"
+            };
+        }
+        else
+        {
+            throw new Exception("Minimax API response invalid: " + responseString);
+        }
+    }
+
+    private static byte[] ConvertHexStringToByteArray(string hexString)
+    {
+        if (hexString.Length % 2 != 0)
+        {
+            throw new ArgumentException("Hex string must have an even length.");
+        }
+
+        byte[] data = new byte[hexString.Length / 2];
+        for (int index = 0; index < data.Length; index++)
+        {
+            string byteValue = hexString.Substring(index * 2, 2);
+            data[index] = byte.Parse(byteValue, System.Globalization.NumberStyles.HexNumber, System.Globalization.CultureInfo.InvariantCulture);
+        }
+
+        return data;
+    }
+}

media-worker/Services/RemotionService.cs

@@ -21,6 +21,7 @@ public class RemotionService
         List<GeneratedMediaFile> generatedMedia,
         string? musicPath,
         int targetDurationSeconds,
+        string visualEffect,
         CancellationToken ct)
     {
         _logger.LogInformation("🎬 Remotion render başlatılıyor — Project: {Id}", projectId);
@@ -32,6 +33,7 @@ public class RemotionService
         var props = new
         {
             musicPath = musicPath,
+            visualEffect = visualEffect,
             scenes = scenes.Select(s => new
             {
                 imagePath = s.ImagePath,

media-worker/Services/VideoRenderPipeline.cs

@@ -24,6 +24,7 @@ public class VideoRenderPipeline
     private readonly HiggsFieldService _higgsField;
     private readonly TtsService _tts;
     private readonly OpenAiTtsService _openAiTts;
+    private readonly MinimaxTtsService _minimaxTts;
     private readonly SunoMusicService _sunoMusic;
     private readonly AudioCraftService _audioCraft;
     private readonly RemotionService _remotion;
@@ -36,6 +37,7 @@ public class VideoRenderPipeline
         HiggsFieldService higgsField,
         TtsService tts,
         OpenAiTtsService openAiTts,
+        MinimaxTtsService minimaxTts,
         SunoMusicService sunoMusic,
         AudioCraftService audioCraft,
         RemotionService remotion,
@@ -47,6 +49,7 @@ public class VideoRenderPipeline
         _higgsField = higgsField;
         _tts = tts;
         _openAiTts = openAiTts;
+        _minimaxTts = minimaxTts;
         _sunoMusic = sunoMusic;
         _audioCraft = audioCraft;
         _remotion = remotion;
@@ -176,13 +179,48 @@ public class VideoRenderPipeline
             await progressCallback(70, "AMBIENT_GENERATION");
 
             // ═══════════════════════════════════════
-            // ADIM 5: REMOTION — Video render (Ken Burns + Audio Merge + Subtitles)
+            // ADIM 5: REMOTION — Segmented Video render + FFmpeg Merge
             // ═══════════════════════════════════════
-            _logger.LogInformation("🎬 Adım 5/6: Remotion render — Ken Burns + audio merge + subtitle");
-            await progressCallback(75, "MEDIA_MERGE");
+            _logger.LogInformation("🎬 Adım 5/6: Remotion Segmented Render — Ken Burns + audio merge + subtitle");
+            await progressCallback(70, "MEDIA_MERGE");
 
-            var finalLocalPath = await _remotion.RenderVideoAsync(
-                job.ProjectId, projectDir, scenes, allMediaFiles, musicFile?.LocalPath, job.TargetDuration, ct);
+            int chunkSize = 20; // 20 scenes per chunk to prevent OOM
+            var chunkPaths = new List<string>();
+            var chunkIndex = 0;
+
+            for (int i = 0; i < scenes.Count; i += chunkSize)
+            {
+                var chunkScenes = scenes.Skip(i).Take(chunkSize).ToList();
+                chunkIndex++;
+
+                _logger.LogInformation("Render Chunk {ChunkIndex} (Scenes {Start} to {End})", 
+                    chunkIndex, i + 1, i + chunkScenes.Count);
+
+                // Pass null for musicPath so Remotion doesn't add music to each chunk
+                var chunkPath = await _remotion.RenderVideoAsync(
+                    $"{job.ProjectId}_chunk_{chunkIndex}", 
+                    projectDir, 
+                    chunkScenes, 
+                    allMediaFiles, 
+                    null, // No music per chunk
+                    0, 
+                    job.VisualEffect, 
+                    ct);
+
+                chunkPaths.Add(chunkPath);
+
+                var progress = 70 + (int)(20.0 * (i + chunkScenes.Count) / scenes.Count);
+                await progressCallback(progress, "MEDIA_MERGE");
+            }
+
+            _logger.LogInformation("🎬 Chunklar birleştiriliyor ve müzik ekleniyor (FFmpeg)");
+            var finalLocalPath = await _ffmpeg.ConcatenateAndFinalize(
+                chunkPaths, 
+                musicFile?.LocalPath, 
+                projectDir, 
+                job.ProjectId, 
+                job.TargetDuration, 
+                ct);
 
             allMediaFiles.Add(new GeneratedMediaFile
             {
@@ -234,6 +272,10 @@ public class VideoRenderPipeline
         {
             result = await _openAiTts.GenerateNarrationAsync(scene, outputDir, voiceStyle, ct);
         }
+        else if (!string.IsNullOrEmpty(scene.TtsProvider) && scene.TtsProvider.Equals("minimax", StringComparison.OrdinalIgnoreCase))
+        {
+            result = await _minimaxTts.GenerateNarrationAsync(scene, outputDir, voiceStyle, ct);
+        }
         else
         {
             // Default: ElevenLabs