0. dify环境变量修改

# 外部访问
APP_WEB_URL=http://10.0.2.218:8088
FILES_URL=http://10.0.2.218:8088

EXPOSE_NGINX_PORT=8088
EXPOSE_NGINX_SSL_PORT=1443

# 文件上传限制
UPLOAD_FILE_SIZE_LIMIT=500
UPLOAD_FILE_BATCH_LIMIT=50
NGINX_CLIENT_MAX_BODY_SIZE=500M

1. 批量添加分段

把excel里面的Q/A内容批量插入知识库，代码的方式插入有知识库,其中问题整理.xlsxexcel文件格式为问题、答案两列:格式如下:
问题1 答案1
问题2 答案2

import requests
import json
import pandas as pd


# pip install pandas openpyxl

def send_post_request(document_id="9e1538d6-1e14-4543-b49f-e890154b0737", question="你好", answer="你好！"):
    # 请求URL
    url = "http://ip:port/console/api/datasets/075ecd79-b1a5-4d2e-a466-2012c0dda14c/documents/9e1538d6-1e14-4543-b49f-e890154b0737/segment"

    # 请求头
    token = "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ1c2VyX2lkIjoiMzhmMzE1M2YtZDhiOS00MTgxLWI0ZmItZjU2MzdkMGU0Mzk5IiwiZXhwIjoxNzYwNDI2NjU1LCJpc3MiOiJTRUxGX0hPU1RFRCIsInN1YiI6IkNvbnNvbGUgQVBJIFBhc3Nwb3J0In0.F62P8qvUvC_bAxc_9KjQNcFN3ucyiB3Bi_fZL_Q_KKA"

    headers = {
        # 2. 确保Bearer后面有空格，格式正确
        "Authorization": f"Bearer {token}",
        "Content-Type": "application/json"
    }

    # 请求数据
    data = {
        "id": "5d640c23-1495-4751-aea7-554fc588df31",
        "position": 5,
        "document_id": document_id,
        "content": question,
        "sign_content": "c",
        "answer": answer,
        "word_count": 2,
        "tokens": 1,
        "keywords": None,
        "index_node_id": "f029fc4b-6014-4373-a89a-6faa4f700f78",
        "index_node_hash": "2f7cfb4fa6fef769c6daff15772f3b1193b19b20d7a1534ca6d2542658ec69e5",
        "hit_count": 0,
        "enabled": True,
        "disabled_at": None,
        "disabled_by": None,
        "status": "completed",
        "created_by": "38f3153f-d8b9-4181-b4fb-f5637d0e4399",
        "created_at": 1760422745,
        "updated_at": 1760422745,
        "updated_by": None,
        "indexing_at": 1760422745,
        "completed_at": 1760422745,
        "error": None,
        "stopped_at": None,
        "child_chunks": []
    }

    try:
        # 发送POST请求
        response = requests.post(url, headers=headers, json=data)

        # 检查响应状态码
        response.raise_for_status()

        # 打印响应内容
        print("请求成功！")
        print("响应状态码:", response.status_code)
        print("响应内容:", json.dumps(response.json(), indent=2, ensure_ascii=False))

        return response.json()

    except requests.exceptions.RequestException as e:
        print(f"请求失败: {e}")
        return None


def batch_insert():
    # 读取Excel文件
    file_path = '/home/geekplusa/ai/projects/data/问题整理.xlsx'
    df = pd.read_excel(file_path, sheet_name='内部知识库')

    # 初始化结果列表
    qa_list = []

    # 遍历每一行
    for index, row in df.iterrows():
        question = row.iloc[0]  # 第二列是问题
        answer = row.iloc[1]  # 第三列是答案

        send_post_request(question=question, answer=answer)
        print(f"问题: {question}, 答案: {answer}")

if __name__ == "__main__":
    batch_insert()