1. Error when running docker-compose: Error while fetching server API version: HTTPConnection.request()

When installing and running docker-compose, the following error occurs:

Traceback (most recent call last):
  File "/usr/bin/docker-compose", line 33, in <module>
    sys.exit(load_entry_point('docker-compose==1.29.2', 'console_scripts', 'docker-compose')())
  File "/usr/lib/python3/dist-packages/compose/cli/main.py", line 81, in main
    command_func()
  File "/usr/lib/python3/dist-packages/compose/cli/main.py", line 200, in perform_command
    project = project_from_options('.', options)
  File "/usr/lib/python3/dist-packages/compose/cli/command.py", line 60, in project_from_options
    return get_project(
  File "/usr/lib/python3/dist-packages/compose/cli/command.py", line 152, in get_project
    client = get_client(
  File "/usr/lib/python3/dist-packages/compose/cli/docker_client.py", line 41, in get_client
    client = docker_client(
  File "/usr/lib/python3/dist-packages/compose/cli/docker_client.py", line 170, in docker_client
    client = APIClient(use_ssh_client=not use_paramiko_ssh, **kwargs)
  File "/usr/lib/python3/dist-packages/docker/api/client.py", line 197, in __init__
    self._version = self._retrieve_server_version()
  File "/usr/lib/python3/dist-packages/docker/api/client.py", line 221, in _retrieve_server_version
    raise DockerException(
docker.errors.DockerException: Error while fetching server API version: HTTPConnection.request() got an unexpected keyword argument 'chunked'

This is caused by an outdated version of Python's urllib. Upgrade urllib3 to 2.0.0:

pip install 'urllib3<2'
  1. Function Call Issue: Signature Failure

When calling functions like Weibo Hot Search or DALL·E 3, you may encounter the following error:

error with parse auth token: token signature is invalid: signature is invalid

This happens because the function call token imported into the database was generated locally. The local key will differ from your online key, so you need to regenerate the function token in the admin panel.

Go to the admin panel => "Edit Function" and regenerate the token.

Regenerate Token

  1. Inconsistent Function Calls

Large language models have inherent randomness. Even if you ask the same question multiple times, the answers may vary. Therefore, sometimes a function call may be triggered, while other times it may not. This is normal behavior. Based on testing, models like gpt-3.5-turbo, gpt-3.5-16k, and gpt-4 exhibit more stable function calling. The 1106 and 0125 series models are less stable in this regard.

  1. System Error: "System is Unavailable"

If you encounter the error "System is Unavailable," it usually indicates a backend API issue. Error logs are typically printed, so you can check the logs to identify the specific cause.

The error logs are located in the logs/app.log file in your deployment root directory (where the docker-compose.yaml file is located).

  1. Large Model Fails to Recognize Uploaded PDF Files

First, ensure the Tika container is running. Execute docker compose ps in the project directory to verify the Tika container's status.

geekai-api       geekai-api     2 weeks ago   Up 2 weeks   0.0.0.0:5678->5678/tcp, 0.0.0.0:9999->9999/tcp
geekai-mysql     geekai-mysql   2 weeks ago   Up 2 weeks   33060/tcp, 0.0.0.0:3307->3306/tcp
geekai-redis     geekai-redis   2 weeks ago   Up 2 weeks   0.0.0.0:6380->6379/tcp
geekai-web       geekai-web     2 weeks ago   Up 2 weeks   80/tcp, 443/tcp, 0.0.0.0:8080->8080/tcp
tika             tika           2 weeks ago   Up 2 weeks   0.0.0.0:9998->9998/tcp

Next, check the tika configuration in the config/config.toml file:

Listen = "0.0.0.0:5678"
StaticDir = "./static"
StaticUrl = "/static"
TikaHost = "http://tika:9998"

Most importantly, avoid using http://localhost:8080 to access your system. Instead, use http://IP:8080. Files are uploaded to the geekai-api container, and the Tika container cannot download and parse files via localhost. This is why the model might respond with errors like "I cannot read local files."

  1. Image-to-Image or Image-to-Video Task Fails

Image-to-Image and Image-to-Video tasks require third-party APIs to access the uploaded images. If you are testing locally, the images are stored locally, and third-party APIs cannot access them, leading to errors. There are two solutions:

  1. Avoid local storage. Configure third-party OSS services like Qiniu, Alibaba Cloud, or Tencent Cloud to upload images and provide access to the APIs.
  2. Deploy the system on an online server to ensure uploaded images are accessible to third-party APIs.