app.py

import os
import gradio as gr
import requests
import joblib
import numpy as np
import librosa
from huggingface_hub import hf_hub_download
from deepface import DeepFace

# --- 配置：Hugging Face Inference API 文本分析 ---
HF_API_TOKEN = os.getenv("Hhf_ZfPJkShFNeheFJQZwkZFEOCkYKCBwNhLAw")
if HF_API_TOKEN is None:
    print("警告：未检测到 HF_API_TOKEN，文字分析可能失败或限流。")
# 选用公开存在的中文情感分类模型 ID
HF_TEXT_MODEL = "uer/roberta-base-finetuned-dianping-chinese"
HF_API_URL = f"https://api-inference.huggingface.co/models/{HF_TEXT_MODEL}"
HEADERS = {"Authorization": f"Bearer {HF_API_TOKEN}"} if HF_API_TOKEN else {}

def call_text_api(text: str):
    if not text or text.strip() == "":
        return {}
    payload = {"inputs": text}
    try:
        res = requests.post(HF_API_URL, headers=HEADERS, json=payload, timeout=15)
        res.raise_for_status()
        data = res.json()
        result = {}
        if isinstance(data, list):
            for item in data:
                label = item.get("label", "")
                score = item.get("score", 0.0)
                result[label] = float(score)
        else:
            # 如果返回不同结构，可根据实际调整
            print("call_text_api 返回格式未预期:", data)
            return {}
        return result
    except Exception as e:
        print("call_text_api error:", e)
        return {"中性": 1.0}

# --- 语音情绪分析 SVM 模型加载 ---
USE_VOICE = True
svm_model = None
if USE_VOICE:
    try:
        print("下载并加载语音 SVM 模型...")
        model_path = hf_hub_download(repo_id="GCLing/emotion-svm-model", filename="svm_emotion_model.joblib")
        svm_model = joblib.load(model_path)
        print("SVM 模型加载完成")
    except Exception as e:
        print("语音 SVM 模型加载失败，禁用语音模块:", e)
        USE_VOICE = False

def extract_feature(signal: np.ndarray, sr: int) -> np.ndarray:
    mfcc = librosa.feature.mfcc(y=signal, sr=sr, n_mfcc=13)
    feat = np.concatenate([np.mean(mfcc, axis=1), np.var(mfcc, axis=1)])
    return feat

def predict_voice(audio_path: str):
    if not USE_VOICE or svm_model is None:
        return {"error": 1.0}
    if not audio_path:
        return {}
    try:
        signal, sr = librosa.load(audio_path, sr=None)
        feat = extract_feature(signal, sr)
        probs = svm_model.predict_proba([feat])[0]
        labels = svm_model.classes_
        return {labels[i]: float(probs[i]) for i in range(len(labels))}
    except Exception as e:
        print("predict_voice error:", e)
        return {}

# --- 臉部情緒分析，使用 DeepFace 分析上傳或拍照圖片 ---
def predict_face(img: np.ndarray):
    # img 為 numpy array，或 None
    if img is None:
        return {}
    try:
        res = DeepFace.analyze(img, actions=["emotion"], detector_backend="opencv")
        if isinstance(res, list):
            first = res[0] if res else {}
            emo = first.get("emotion", {}) if isinstance(first, dict) else {}
        else:
            emo = res.get("emotion", {}) if isinstance(res, dict) else {}
        emo_fixed = {k: float(v) for k, v in emo.items()}
        return emo_fixed
    except Exception as e:
        print("DeepFace.analyze error:", e)
        return {}

# --- Gradio 界面 ---
def build_interface():
    with gr.Blocks() as demo:
        gr.Markdown("## 多模態情緒分析（簡化版：上傳/拍照人臉 + 語音 + 文字）")
        with gr.Tabs():
            # 臉部 Tab：上傳或拍照
            with gr.TabItem("臉部情緒"):
                gr.Markdown("### 臉部情緒 分析 (上傳或拍照圖片)")
                with gr.Row():
                    # sources=["upload"] 在手機上點上傳可調出相機拍照
                    face_input = gr.Image(sources=["upload"], type="numpy", label="上傳或拍照圖片")
                    face_out = gr.Label(label="情緒分布")
                face_input.change(fn=predict_face, inputs=face_input, outputs=face_out)

            # 語音 Tab
            with gr.TabItem("語音情緒"):
                gr.Markdown("### 語音情緒 分析 (錄音並上傳)")
                if USE_VOICE:
                    with gr.Row():
                        audio_input = gr.Audio(source="microphone", streaming=False, type="filepath", label="錄音")
                        voice_out = gr.Label(label="語音情緒結果")
                    audio_input.change(fn=predict_voice, inputs=audio_input, outputs=voice_out)
                else:
                    gr.Markdown("語音模塊不可用。")

            # 文字 Tab
            with gr.TabItem("文字情緒"):
                gr.Markdown("### 文字情緒 分析 (Hugging Face Inference API)")
                with gr.Row():
                    text_input = gr.Textbox(lines=3, placeholder="請輸入中文文字…")
                    text_out = gr.Label(label="文字情緒結果")
                text_input.submit(fn=call_text_api, inputs=text_input, outputs=text_out)

        return demo

if __name__ == "__main__":
    demo = build_interface()
    # share=True 可生成临时公开链接；部署到 Spaces 时无需此参数
    demo.launch()