Unable to run / Cannot get the application to run

[y1342474280]

When installing step by step via pip, dependency missing issues will occur. After switching to Docker installation and opening the web interface, it displays "something wrong".

[dahuilangda]

Hi, thanks for the report. Could you please attach a screenshot of the error or paste the full error message? It would help us troubleshoot this more quickly.

[y1342474280]

你好，谢谢你的报告。能否附上错误截图或粘贴完整错误信息？这会帮助我们更快排查问题。

Installing via pip results in missing dependencies. Specifically, the dependencies shown in the image were not installed (I verified this by comparing against the Dockerfile). It seems the installation guide in the README omitted these packages.

When running the image built from the Dockerfile with the following command:
docker run --rm -d --gpus all
-p 3000:3000 -p 8000:8000
-e http_proxy=""
-e https_proxy=""
-v $(pwd)/data:/app/data
-v $(pwd)/output:/app/output
-v $(pwd)/constants.py:/app/constants.py
biocheminsight

the following information appears. From the logs, it seems the Docker container started normally, but when I open the frontend at http://localhost:3000/, it shows an access error. After initial checking, I am not sure if this is related to not setting the API key for the large language model. However, I can run the previous steps successfully using the command-line mode with the Docker image. Lastly, my English is not very good, and the above content was translated with AI. I hope it does not cause any confusion for you. PS: The download speed from the Tsinghua mirror in the Dockerfile is extremely slow. Switching to the Alibaba Cloud mirror is much faster. I am not sure if this is just my personal network issue, but I would like to report it. Additionally, the command-line run failed due to a permission problem (the user inside the Docker container does not have permission to write to the /app/output directory). I fixed this issue with AI assistance, but I can no longer find the relevant logs.

[dahuilangda]

你好，感谢你的详细反馈！针对你提到的两个问题：

1. pip 安装依赖缺失

抱歉，README 中的安装步骤确实遗漏了部分依赖包（对比 Dockerfile 可以发现）。我已经更新了 README（#10），补充了以下缺失：

pip install Levenshtein mdutils google-generativeai tabulate python-multipart -i https://pypi.tuna.tsinghua.edu.cn/simple

建议先尝试本地安装（不用 Docker），按更新后的 README 一步步操作，这样更容易定位问题。

2. Docker 前端访问报错

你提到 Docker 容器启动后，前端 http://localhost:3000 显示 "something wrong"。这个问题很可能是因为 OCR 微服务没有启动，导致后端 API 调用失败。

BioChemInsight 在运行前 必须先启动至少一个 OCR 微服务：

方式 A：启动 PaddleOCR 服务（推荐）

cd DOCKER_PADDLE_OCR
docker compose build
docker compose up -d

服务默认监听 8010 端口，然后在 constants.py 中配置：

PADDLEOCR_SERVER_URL = 'http://localhost:8010'
DEFAULT_OCR_ENGINE = 'paddleocr'

方式 B：启动 DotsOCR 服务

cd DOCKER_DOTS_OCR
python download_model.py    # 先下载模型权重
docker compose pull
docker compose up -d

然后在 constants.py 中配置：

DOTSOCR_SERVER_URL = 'http://localhost:8001'
DEFAULT_OCR_ENGINE = 'dots_ocr'

注意：DotsOCR 需要约 30GB 显存（RTX 5090 级别）。

排查步骤

1. 确保 OCR 微服务已启动并正常运行
2. 用以下命令检查主容器日志，确认后端是否正常启动：

docker logs -f <container_id>

正常启动后应能看到：

INFO:     Started server process [7]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Uvicorn running on http://0.0.0.0:8000 (Press CTRL+C to quit)

3. 如果后端日志有报错，请贴出完整错误信息。

mkdir -p output && chmod 777 output

最后：推荐先用本地安装的方式（非 Docker）一步步跑通，这样更容易理解和排查每个环节的问题。有问题随时贴日志。