EAI-deepface/main.py

from flask import Flask, request, jsonify
from upload_file import OSSUploader
import os
from video_handler import process_video, merge_videos
from speech_synthesis_factory import TtsClientWrapper
import azure.cognitiveservices.speech as speechsdk
import time


app = Flask(__name__)

oss_uploader = OSSUploader()


@app.route('/tts', methods=['POST'])
def tts_process():
    print(request.get_json())
    data = request.get_json()
    if 'code' not in data:
        return jsonify({"error": "Missing 'code' in request data"}), 400
    if 'text' not in data:
        return jsonify({"error": "Missing 'text' in request data"}), 400
    tts_wrapper = TtsClientWrapper()
    code = data['code']
    text = data['text']
    synthesizer = tts_wrapper.getTtsClient(code)
    speech_synthesis_result = synthesizer.speak_text_async(text).get()
    if speech_synthesis_result.reason == speechsdk.ResultReason.SynthesizingAudioCompleted:
        audio_data_stream = speechsdk.AudioDataStream(speech_synthesis_result)
        timestamp = int(time.time())
        temp_file_path = f'tts_{timestamp}.wav'
        audio_data_stream.save_to_wav_file(temp_file_path)
        video_path = oss_uploader.upload_video_to_oss_local(temp_file_path)
        # 删除临时文件
        if os.path.exists(temp_file_path):
            os.remove(temp_file_path)
        print(video_path)
        return video_path
    else:
        return jsonify({"error": f"Speech synthesis failed: {speech_synthesis_result.error_details}"})


@app.route('/deepface', methods=['POST'])
def deepface_process():
    print(request.get_json())
    video_paths = []
    try:
        data = request.get_json()
        if 'video_urls' not in data:
            return jsonify({"error": "Missing 'video_urls' in request data"}), 400
        video_urls = data['video_urls']
        # 分别处理每个视频URL
        emotion_dicts = []
        for url in video_urls:
            emotion_dict, video_path = process_video(url.split("?")[0])
            emotion_dicts.append(emotion_dict)
            video_paths.append(video_path)

        # 合并标注好的视频
        merged_video_path = merge_videos(video_paths)
        # 合并情绪字典
        merged_emotion_dict = {}
        for d in emotion_dicts:
            for emotion, count in d.items():
                if emotion in merged_emotion_dict:
                    merged_emotion_dict[emotion] += count
                else:
                    merged_emotion_dict[emotion] = count

        # 计算总帧数
        total_frames = sum(sum(d.values()) for d in emotion_dicts)

        # 计算每种情绪的百分比并保留整数部分
        emotion_percentage_dict = {}
        for emotion, count in merged_emotion_dict.items():
            percentage = int((count / total_frames) * 100)
            emotion_percentage_dict[emotion] = percentage

        merged_video_path = oss_uploader.upload_video_to_oss_local(merged_video_path)
        response = {
            "emotion_percentage_dict": emotion_percentage_dict,
            "merged_video_path": merged_video_path
        }
        return jsonify(response)
    finally:
        # 删除中间产生的标注视频
        for video_path in video_paths:
            if os.path.exists(video_path):
                os.remove(video_path)


if __name__ == '__main__':
    app.run(host='0.0.0.0', port=2002, debug=True)