Update app.py

app.py

@@ -2,18 +2,26 @@ import gradio as gr
 import os
 import json
 import pandas as pd
+import random
 from datasets import load_dataset
 from huggingface_hub import HfApi
 
 dataset = load_dataset("intersteller2887/Turing-test-dataset", split="train")
 
+all_data_audio_paths = [
+    item['audio']['path'] for item in dataset
+    if item['audio']['path'].endswith(".wav") and "/data/" in item['audio']['path'].replace("\\", "/")
+]
+
+sample1_audio_path = next((p for p in all_data_audio_paths if p.endswith("sample1.wav")), None)
+
 # ==============================================================================
 # 数据定义 (Data Definition)
 # ==============================================================================
 DIMENSIONS_DATA = [
     {
         "title": "语义和语用特征",
-        "audio": "data/sample1.wav",
+        "audio": sample1_audio_path,
         "desc": "这是“语义和语用特征”维度的文本描述示例。",
         "sub_dims": [
             "记忆一致性：回应者是否能够正确并正确并延续并记忆并延续对话信息？是否存在对上下文的误解或不自洽？", "逻辑连贯性：回应者在语义与对话结构上保持前后一致、合乎逻辑？是否存在前后矛盾的情况？",
@@ -33,7 +41,7 @@ DIMENSIONS_DATA = [
     },
     {
         "title": "非生理性副语言特征",
-        "audio": "data/sample1.wav",
+        "audio": sample1_audio_path,
         "desc": "这是“非生理性副语言特征”维度的文本描述示例。",
         "sub_dims": [
             "节奏：回应者是否存在自然的停顿？语速是否存在自然、流畅的变化？", "语调：在表达疑问、惊讶、强调时，回应者的音调是否会自然上扬或下降？是否表现出符合语境的变化？",
@@ -48,7 +56,7 @@ DIMENSIONS_DATA = [
     },
     {
         "title": "生理性副语言特征",
-        "audio": "data/sample1.wav",
+        "audio": sample1_audio_path,
         "desc": "这是“生理性副语言特征”维度的文本描述示例。",
         "sub_dims": [
             "微生理杂音：回应中是否出现如呼吸声、口水音、气泡音等无意识发声？这些发声是否自然地穿插在恰当的语流节奏当中？",
@@ -62,7 +70,7 @@ DIMENSIONS_DATA = [
     },
     {
         "title": "机械人格",
-        "audio": "data/sample1.wav",
+        "audio": sample1_audio_path,
         "desc": "这是“机械人格”维度的文本描述示例。",
         "sub_dims": [
             "谄媚现象：回应者是否频繁地赞同用户、重复用户的说法、不断表示感谢或道歉？是否存在“无论用户说什么都肯定或支持”的语气模式？",
@@ -75,7 +83,7 @@ DIMENSIONS_DATA = [
     },
     {
         "title": "情感表达",
-        "audio": "data/sample1.wav",
+        "audio": sample1_audio_path,
         "desc": "这是“情感表达”维度的文本描述示例。",
         "sub_dims": [
             "语义层面：回应者的语言内容是否体现出符合上下文的情绪反应？是否表达了人类对某些情境应有的情感态度？",
@@ -87,12 +95,23 @@ DIMENSIONS_DATA = [
                         """
     }
 ]
+
 DIMENSION_TITLES = [d["title"] for d in DIMENSIONS_DATA]
+
+random.seed()
+selected_audio_paths = random.sample(all_data_audio_paths, 5)
+
 QUESTION_SET = [
+    {"audio": path, "desc": f"这是音频文件 {os.path.basename(path)} 的描述"} 
+    for path in selected_audio_paths
+]
+
+"""QUESTION_SET = [
     {"audio": "data/Ses02F_impro01.wav", "desc": "这是第一个测试文件的描述",},
     {"audio": "data/Ses02F_impro02.wav", "desc": "这是第二个测试文件的描述",},
     {"audio": "data/Ses02F_impro03.wav", "desc": "这是第三个测试文件的描述",},
-]
+]"""
+
 MAX_SUB_DIMS = max(len(d['sub_dims']) for d in DIMENSIONS_DATA)
 
 # ==============================================================================
@@ -462,8 +481,8 @@ with gr.Blocks(theme=gr.themes.Soft(), css=".gradio-container {max-width: 960px
 # 程序入口 (Entry Point)
 # ==============================================================================
 if __name__ == "__main__":
-    if not os.path.exists("audio"):
-        os.makedirs("audio")
+    """if not os.path.exists("audio"):
+        os.makedirs("audio")"""
     # A quick check to see if we're in a deployed Space, to avoid local errors.
     if "SPACE_ID" in os.environ:
         print("Running in a Hugging Face Space, checking for audio files...")