Streamlit本地环境文件上传至后端卡住问题排查

原Streamlit应用与LLM对接功能正常，能返回正确响应。但在添加Chroma DB及文本分块、拆分功能后，前端文件上传功能失效：上传文件时卡在“Uploading to backend”环节，即便设置了300秒超时限制，仍会触发超时错误。

Streamlit前端代码 #

import streamlit as st
import requests

# Backend URL
BACKEND_URL = "http://127.0.0.1:8000"

st.title("⚖️ AI Legal Contract Analyzer")

# File Upload
uploaded_file = st.file_uploader("Upload a legal document (PDF, DOCX, TXT)", type=["pdf", "docx", "txt"])

if uploaded_file is not None:
    st.info(f"📄 Uploading file: {uploaded_file.name} ({uploaded_file.size} bytes)")

    # Detect MIME type
    if uploaded_file.type:
        mime_type = uploaded_file.type
    else:
        # fallback by extension
        if uploaded_file.name.endswith(".pdf"):
            mime_type = "application/pdf"
        elif uploaded_file.name.endswith(".docx"):
            mime_type = "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
        else:
            mime_type = "text/plain"

    # Show progress
    progress_bar = st.progress(0)
    status_text = st.empty()

    try:
        status_text.text("🔄 Processing file...")
        progress_bar.progress(25)

        # Read file bytes
        file_bytes = uploaded_file.read()
        files = {"file": (uploaded_file.name, file_bytes, mime_type)}

        status_text.text("📤 Uploading to backend...")
        progress_bar.progress(50)

        response = requests.post(f"{BACKEND_URL}/parse", files=files, timeout=300)

        progress_bar.progress(75)

        if response.status_code == 200:
            result = response.json()
            progress_bar.progress(100)
            status_text.text("✅ Processing complete!")

            st.success(f"File '{uploaded_file.name}' parsed successfully!")
            st.info("📊 **Processing Results:**")
            st.write(f"- **Chunks created:** {result.get('chunks_stored', 'N/A')}")

        else:
            progress_bar.progress(0)
            status_text.text("❌ Processing failed")
            st.error(f"Failed to parse the file. Status: {response.status_code}")
            if response.text:
                st.error(f"Error details: {response.text}")

    except requests.exceptions.Timeout:
        progress_bar.progress(0)
        status_text.text("⏰ Request timed out")
        st.error("⏰ The file processing took too long.")

    except requests.exceptions.ConnectionError:
        progress_bar.progress(0)
        status_text.text("🔌 Connection error")
        st.error("🔌 Could not connect to the backend server.")

    except Exception as e:
        progress_bar.progress(0)
        status_text.text("❌ Unexpected error")
        st.error(f"❌ An unexpected error occurred: {str(e)}")


# Question Answering
st.subheader("Ask a Question")
user_query = st.text_input("Enter your question about the document")

if st.button("Get Answer"):
    if user_query.strip() == "":
        st.warning("Please enter a question.")
    else:
        response = requests.post(f"{BACKEND_URL}/query", json={"query": user_query})
        if response.status_code == 200:
            st.write("### Answer:")
            st.write(response.json().get("answer"))
        else:
            st.error("Error fetching answer from backend.")

报错现象 #

• 上传进程卡在“Uploading to backend”阶段，无法推进到后续步骤
• 最终触发超时错误提示

排查与解决方案建议 #

1. 定位后端/parse接口性能瓶颈 #

• 在后端代码的文本分块、嵌入生成、Chroma DB写入等关键步骤添加日志，记录每个环节的耗时，定位具体慢操作
• 针对大文件优化分块策略：减小单块文本长度、采用并行处理分块，降低整体处理时间
• 改用异步处理+前端轮询模式：后端先接收文件存储，返回任务ID；前端定期调用查询接口获取处理状态，避免长连接超时

2. 优化后端资源与配置 #

• 若使用本地嵌入模型，优先采用量化版本减小内存占用，或切换为云嵌入API降低本地计算压力
• 将Chroma DB从默认内存模式切换为持久化存储（如SQLite或PostgreSQL），避免内存过载导致写入缓慢
• 调整后端框架（如FastAPI）的超时设置，确保后端处理时间不被框架提前中断

3. 前端请求优化 #

• 替换同步POST请求为异步流程：先上传文件到后端临时存储，再触发处理任务，前端通过轮询获取结果
• 添加前端重试机制，针对临时网络波动或短时间超时自动重试请求

4. 网络与基础环境检查 #

• 用curl直接调用后端/parse接口测试，排除Streamlit前端的请求问题
• 确认后端服务端口未被占用、防火墙未拦截请求，确保前后端通信正常

内容的提问来源于stack exchange，提问作者Mirza Mahad Baig