How to use File Server correctly?

[luylucas10]

I've been reviewing the documentation, but it seems like there's still some work to be done for my specific use case. I've set up the file server directory as follows:

# Constants
ALLOWED_DIRECTORIES = [
    str(pathlib.Path("/home/<my-user>/projects").resolve())
]  # 👈 Replace with your paths

I'm trying to grant access to all my source code files to my models, but I've noticed that sometimes the listing works, while other times it doesn't return any files or code that doesn't exist. I'm getting a bit frustrated and would love some guidance on how to resolve this issue.

Is there a specific model or configuration that I should use to get it working as expected? I'm using Ollama, Open-webui in Docker, and running the file server from the terminal. I've managed to connect to the fileserver from OWUI, but I'd like to get the listing feature working consistently.

This is the fileserver log, just for example:

INFO:     127.0.0.1:40092 - "POST /read_file HTTP/1.1" 403 Forbidden
INFO:     127.0.0.1:48684 - "GET / HTTP/1.1" 404 Not Found
INFO:     127.0.0.1:48684 - "GET /favicon.ico HTTP/1.1" 404 Not Found
INFO:     127.0.0.1:58484 - "GET /openapi.json HTTP/1.1" 200 OK
INFO:     127.0.0.1:47382 - "POST /list_directory HTTP/1.1" 200 OK
INFO:     127.0.0.1:36820 - "GET /docs HTTP/1.1" 200 OK
INFO:     127.0.0.1:36820 - "GET /openapi.json HTTP/1.1" 200 OK
INFO:     127.0.0.1:58802 - "POST /list_directory HTTP/1.1" 400 Bad Request
INFO:     127.0.0.1:49462 - "GET /openapi.json HTTP/1.1" 200 OK
INFO:     127.0.0.1:40266 - "GET /openapi.json HTTP/1.1" 200 OK
INFO:     192.168.0.4:52818 - "OPTIONS /openapi.json HTTP/1.1" 200 OK
INFO:     192.168.0.4:52818 - "GET /openapi.json HTTP/1.1" 200 OK
INFO:     192.168.0.4:52818 - "GET /openapi.json HTTP/1.1" 200 OK
INFO:     192.168.0.4:52818 - "GET /openapi.json HTTP/1.1" 200 OK
INFO:     192.168.0.4:52818 - "GET /openapi.json HTTP/1.1" 200 OK
INFO:     192.168.0.4:42912 - "OPTIONS /list_allowed_directories HTTP/1.1" 200 OK
INFO:     192.168.0.4:42912 - "GET /list_allowed_directories HTTP/1.1" 200 OK
INFO:     192.168.0.4:48800 - "GET /list_allowed_directories HTTP/1.1" 200 OK
INFO:     192.168.0.4:36636 - "OPTIONS /directory_tree HTTP/1.1" 200 OK
INFO:     192.168.0.4:36636 - "POST /directory_tree HTTP/1.1" 200 OK

[luylucas10]

I'm trying to create a model that can generate code for my project based on a specific directory. I've been considering using the Git Server, but I'm not sure if it's the best approach. One option is to synchronize the directory as a knowledge base, but I'd need to constantly update it and also consider individualizing each project's directories for better performance.

Can anyone suggest an alternative or guidance on how to move forward?

[rwachters]

Maybe you could try adjusting the system prompt?.. in order to provide some guidance for the LLM.. like they suggest for the memory server: https://github.com/modelcontextprotocol/servers/tree/main/src/memory

Follow these steps for each interaction:

1. User Identification:
   - You should assume that you are interacting with default_user
   - If you have not identified default_user, proactively try to do so.

2. Memory Retrieval:
   - Always begin your chat by saying only "Remembering..." and retrieve all relevant information from your knowledge graph
   - Always refer to your knowledge graph as your "memory"

3. Memory
   - While conversing with the user, be attentive to any new information that falls into these categories:
     a) Basic Identity (age, gender, location, job title, education level, etc.)
     b) Behaviors (interests, habits, etc.)
     c) Preferences (communication style, preferred language, etc.)
     d) Goals (goals, targets, aspirations, etc.)
     e) Relationships (personal and professional relationships up to 3 degrees of separation)

4. Memory Update:
   - If any new information was gathered during the interaction, update your memory as follows:
     a) Create entities for recurring organizations, people, and significant events
     b) Connect them to the current entities using relations
     b) Store facts about them as observations

I tried this in LibreChat and it worked pretty well.. but I have to say, I'm still trying to work this out in open webui. The extra layer with mcpo -> openapi doesn't make it easier.