NeoAnd/evolution-api
1
0
Fork 0
mirror of https://github.com/EvolutionAPI/evolution-api.git synced 2025-12-20 12:22:21 -06:00
Code Issues Actions Packages Projects Releases Wiki Activity

refactor(chatbot): integrate OpenAI service into chatbot implementations

Browse Source 
- Updated various chatbot services (Typebot, Dify, EvolutionBot, Flowise, N8n) to include the OpenAI service for audio transcription capabilities.
- Modified constructors to accept OpenaiService as a dependency, enhancing the ability to transcribe audio messages directly within each service.
- Refactored the handling of `keywordFinish` in multiple controllers and services, changing its type from an array to a string for consistency and simplifying logic.
- Removed redundant audio transcription logic from the base service, centralizing it within the OpenAI service to improve maintainability and reduce code duplication.

This commit focuses on enhancing the chatbot services by integrating OpenAI's transcription capabilities, improving code structure, and ensuring consistent handling of session keywords.
This commit is contained in:
Davidson Gomes
2025-05-21 22:17:10 -03:00
parent 9cedf31eed
commit 6a0fc19702
10 changed files with 124 additions and 90 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
src/api/integrations/chatbot/base-chatbot.service.ts
View File

@@ -2,11 +2,9 @@ import { InstanceDto } from '@api/dto/instance.dto';
import { PrismaRepository } from '@api/repository/repository.service';
import { WAMonitoringService } from '@api/services/monitor.service';
import { Integration } from '@api/types/wa.types';
import { ConfigService, Language } from '@config/env.config';
import { ConfigService } from '@config/env.config';
import { Logger } from '@config/logger.config';
import { IntegrationSession } from '@prisma/client';
import axios from 'axios';
import FormData from 'form-data';
/**
* Base class for all chatbot service implementations
@@ -73,46 +71,6 @@ export abstract class BaseChatbotService<BotType = any, SettingsType = any> {
return null;
}
/**
* Transcribes audio to text using OpenAI's Whisper API
*/
protected async speechToText(audioBuffer: Buffer): Promise<string | null> {
if (!this.configService) {
this.logger.error('ConfigService not available for speech-to-text transcription');
return null;
}
try {
// Try to get the API key from process.env directly since ConfigService might not access it correctly
const apiKey = this.configService.get<any>('OPENAI')?.API_KEY || process.env.OPENAI_API_KEY;
if (!apiKey) {
this.logger.error('No OpenAI API key set for Whisper transcription');
return null;
}
const lang = this.configService.get<Language>('LANGUAGE').includes('pt')
? 'pt'
: this.configService.get<Language>('LANGUAGE');
const formData = new FormData();
formData.append('file', audioBuffer, 'audio.ogg');
formData.append('model', 'whisper-1');
formData.append('language', lang);
const response = await axios.post('https://api.openai.com/v1/audio/transcriptions', formData, {
headers: {
...formData.getHeaders(),
Authorization: `Bearer ${apiKey}`,
},
});
return response?.data?.text || null;
} catch (err) {
this.logger.error(`Whisper transcription failed: ${err}`);
return null;
}
}
/**
* Create a new chatbot session
*/
@@ -174,12 +132,9 @@ export abstract class BaseChatbotService<BotType = any, SettingsType = any> {
}
// For existing sessions, keywords might indicate the conversation should end
const keywordFinish = (settings as any)?.keywordFinish || [];
const keywordFinish = (settings as any)?.keywordFinish || '';
const normalizedContent = content.toLowerCase().trim();
if (
keywordFinish.length > 0 &&
keywordFinish.some((keyword: string) => normalizedContent === keyword.toLowerCase())
) {
if (keywordFinish.length > 0 && normalizedContent === keywordFinish.toLowerCase()) {
// Update session to closed and return
await this.prismaRepository.integrationSession.update({
where: {