fix: resolve build errors and audio transcription issues across chatbot services

- Add YAML file loader to tsup.config.ts to fix build compilation errors
- Fix OpenAI speechToText method signature across all chatbot services
- Correct DifyService constructor parameter order in server.module.ts and channel.service.ts
- Add missing OpenAI service dependency injection to EvoaiService
- Standardize audio transcription logic in FlowiseService to match N8N implementation
- Fix speechToText calls in WhatsApp Baileys and Evolution channel services
- Ensure consistent error handling and parameter passing for audio processing

This resolves the "Cannot read properties of undefined" errors and ensures
all chatbot integrations (OpenAI, N8N, Flowise, EvoAI, Dify, EvolutionBot)
properly handle audio message transcription using OpenAI Whisper.

src/api/integrations/chatbot/evolutionBot/services/evolutionBot.service.ts

@@ -62,7 +62,7 @@ export class EvolutionBotService extends BaseChatbotService<EvolutionBot, Evolut
       if (this.isAudioMessage(content) && msg) {
         try {
           this.logger.debug(`[EvolutionBot] Downloading audio for Whisper transcription`);
-          const transcription = await this.openaiService.speechToText(msg);
+          const transcription = await this.openaiService.speechToText(msg, instance);
           if (transcription) {
             payload.query = transcription;
           } else {