script/show_conversation.py

import json
import os

def show_conversation_slice(file_path, target_number, num_surrounding=3):
    """
    Reads a JSON file (assumed to be a list of objects) and prints a specific item
    and its surrounding items.

    Args:
        file_path (str): The path to the JSON file.
        target_number (int): The 1-based index of the item to find.
        num_surrounding (int): The number of items to display before and after the target.
    """
    
    # 将1-based的索引转换为0-based的列表索引
    target_index = target_number - 1

    # 检查文件是否存在
    if not os.path.exists(file_path):
        print(f"错误：文件不存在。请检查路径是否正确：\n{file_path}")
        return

    try:
        # 打开并加载JSON文件
        with open(file_path, 'r', encoding='utf-8') as f:
            data = json.load(f)

        # 检查加载的数据是否为列表且包含足够多的元素
        if not isinstance(data, list) or len(data) < target_index + 1:
            print("错误：文件格式不正确或列表中的元素数量不足。")
            return

        # 计算要切片的起始和结束索引，确保它们在列表范围内
        start_index = max(0, target_index - num_surrounding)
        end_index = min(len(data), target_index + num_surrounding + 1)
        
        # 提取所需的对话切片
        conversations_slice = data[start_index:end_index]
        
        print(f"正在显示从第 {start_index + 1} 条到第 {end_index} 条，共 {len(conversations_slice)} 条对话。\n")
        
        # 遍历并打印切片，标记出目标对话
        for i, conv in enumerate(conversations_slice):
            current_index = start_index + i
            # 使用箭头标记出第3740条对话
            marker = "-->" if current_index == target_index else "   "
            print(f"{marker} 第 {current_index + 1} 条对话：")
            # 使用 json.dumps 格式化输出，使其更易读
            print(json.dumps(conv, indent=2, ensure_ascii=False))
            print("-" * 50) # 添加分隔线

    except json.JSONDecodeError:
        print(f"错误：文件 {file_path} 无法解析为有效的JSON格式。")
    except Exception as e:
        print(f"发生了一个意外错误：{e}")

# --- 配置 ---
# 请确保将这里的路径更改为你的实际文件路径
file_path = "/root/test/weitiao/data_processing_hsichen/data_process_bq/data/delta_score_all_sharegpt_byclosure.json"

# 要查找的对话序号（1-based）
target_conversation_number = 1173

# --- 运行脚本 ---
if __name__ == "__main__":
    show_conversation_slice(file_path, target_conversation_number)