Langchain-Chatchat 往向量库上传文件报错：TypeError: string indices must be integers, not 'str'

问题描述 / Problem Description 在知识库上传csv文件时发生报错

复现问题的步骤 / Steps to Reproduce

执行 'python startup -a'
点击 '知识库管理'
滚动到 '上传文件.'
问题出现 /报错 预期的结果 / Expected Result 成功的传入到faiss向量库中

实际结果 / Actual Result 报错信息：： 2024-05-07 09:45:16,936 - utils.py[line:95] - ERROR: ConnectError: error when post /knowledge_base/search_docs: [Errno 111] Connection refused 2024-05-07 09:45:16,936 - utils.py[line:95] - ERROR: ConnectError: error when post /knowledge_base/search_docs: [Errno 111] Connection refused 2024-05-07 09:45:16,937 - utils.py[line:95] - ERROR: ConnectError: error when post /knowledge_base/search_docs: [Errno 111] Connection refused 2024-05-07 09:45:16.937 Uncaught app exception Traceback (most recent call last): File "/root/miniconda3/envs/lc/lib/python3.11/site-packages/streamlit/runtime/scriptrunner/script_runner.py", line 535, in _run_script exec(code, module.dict) File "/mnt/Langchain-Chatchat/webui.py", line 64, in pages[selected_page]["func"](api=api, is_lite=is_lite) File "/mnt/Langchain-Chatchat/webui_pages/knowledge_base/knowledge_base.py", line 330, in knowledge_base_page data = [ ^ File "/mnt/Langchain-Chatchat/webui_pages/knowledge_base/knowledge_base.py", line 331, in {"seq": i + 1, "id": x["id"], "page_content": x["page_content"], "source": x["metadata"].get("source"), ~^^^^^^ TypeError: string indices must be integers, not 'str'

环境信息 / Environment Information

langchain-ChatGLM 版本/commit 号：v2.10
是否使用 Docker 部署（是/否）：否
使用的模型（ChatGLM2-6B / Qwen-7B 等）：文心一言4.0
使用的 Embedding 模型（moka-ai/m3e-base 等）：bge-large-zh
使用的向量库类型 (faiss / milvus / pg_vector 等)： faiss
操作系统及版本 / Operating system and version: 阿里云服务器
Python 版本 / Python version:3.10
其他相关环境信息 / Other relevant environment information:

附加信息 / Additional Information 添加与问题相关的任何其他信息 / Add any other information related to the issue.

May 07 '24 02:05 b2383355038

The error you're encountering, TypeError: string indices must be integers, not 'str', suggests that the application is attempting to access elements of a string using string keys, which is only valid for dictionaries or lists. This typically happens when the data structure expected is a list of dictionaries (where each dictionary represents a row from the CSV file, and keys are column headers), but instead, a string or incorrect format is being processed.

To resolve this, ensure your CSV file is correctly parsed into a list of dictionaries. You can use the pandas library to easily convert a CSV file into this format:

import pandas as pd

# Assuming 'csv_file' is the path to your CSV file
df = pd.read_csv(csv_file)
records = df.to_dict(orient='records')

# 'records' is now a list of dictionaries, each representing a row from the CSV.

After parsing the CSV file into the correct format, you can iterate over records and access each column value by its header name, which should align with the expected data structure in your application and prevent the TypeError.

If the issue persists after ensuring the CSV is correctly parsed and the data structure is as expected, it might be necessary to review the specific steps or code handling the upload process for further discrepancies.

Details

Sources

About Dosu This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

To continue the conversation, mention @dosu.