Full support of Windows 10 (instead of WSL)

[Redrum624]

Original thread issue #637

Summary
I was able to have a running version of OpenDevin on Windows 10 (not WSL) before the change to Makefile. I switched to WSL since I thought I had no other choice, but whatever I did finished like the issue #561 (ws://localhost:3001/ws). The package manager for Windows Chocolatey that comes with Node.js has a "make" installer (https://community.chocolatey.org/packages/make) so I tried to set it up straight on Windows 10 (which would be better than WSL imho). "uvloop" does not work on Windows, but I read that fastapi works without it on Windows, so I removed the dependency, ran "make build" and was able to run "make start-backend" + "make start-frontend" without errors. I tried to run it with my old LM Studio config and got it to work.

LM Studio:
[2024-04-03 01:24:08.701] [INFO] [LM STUDIO SERVER] Supported endpoints:
[2024-04-03 01:24:08.701] [INFO] [LM STUDIO SERVER] -> GET http://localhost:1234/v1/models
[2024-04-03 01:24:08.702] [INFO] [LM STUDIO SERVER] -> POST http://localhost:1234/v1/chat/completions
[2024-04-03 01:24:08.702] [INFO] [LM STUDIO SERVER] -> POST http://localhost:1234/v1/completions

config.toml:
LLM_BASE_URL="http://localhost:1234/v1/"
LLM_API_KEY="lm-studio"
LLM_EMBEDDING_MODEL="local"
WORKSPACE_DIR="./workspace"

It interacts with the console too.

Motivation
Using Windows 10 instead of WSL and full integration of LM Studio

Technical Design
Windows 10 + "choco install make" + LM Studio

[Redrum624]

Can't execute Python a script though.

It seems like they used Bash for the terminal (like Ubuntu on WSL) because of "/usr/bin/env: 'python3\r': No such file or directory".

Would it be possible to install Python 3.11 with the dependencies to make it callable from the terminal or change the path to the one on PATH to access Python?

[Redrum624]

Would it be possible to change the shell of the terminal depending on the os like so:
const shell = os.platform() === 'win32' ? 'powershell.exe' : 'bash';

[Redrum624]

Dampfinchen suggested to make it work without Docker and on Windows soon.

[Robertc30]

i still get an error....
[pipenv.exceptions.InstallError]: × Getting requirements to build wheel did not run successfully.
[pipenv.exceptions.InstallError]: │ exit code: 1
[pipenv.exceptions.InstallError]: ╰─> See above for output.
[pipenv.exceptions.InstallError]:
[pipenv.exceptions.InstallError]: note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Couldn't install package: {}
Package installation failed...
trying to figure it out but if anyone has an answer im bout that!

[Redrum624]

try removing uvloop from Pipfile.lock

[flavioTocca]

don't you specify the model in the config? Which model do you select in the web UI setting? if you don't select any the BE stops working, I tried putting the model I have in the config but it didn't work. other not local models have no problems
LLM_MODEL="TheBloke/Llama-2-7B-Chat-GGUF/llama-2-7b-chat.Q8_0.gguf"

[Redrum624]

I had to use @python -m pipenv run uvicorn opendevin.server.listen:app --port $(BACKEND_PORT) in make start-frontend instead of @cd frontend && BACKEND_HOST=$(BACKEND_HOST) FRONTEND_PORT=$(FRONTEND_PORT) npm run start which was the old config

[flavioTocca]

I solve the problem by adding this --host 0.0.0.0 to the make file here

start-backend:
@echo "Starting the backend..."
@python3 -m pipenv run uvicorn opendevin.server.listen:app --port $(BACKEND_PORT) --host 0.0.0.0

however now he sends the first step to LM studio but it remains stuck there it does not return the response to devin and the LM studio logs

lmstudio-server-log.txt

[Crypto-78]

I solve the problem by adding this --host 0.0.0.0 to the make file here

start-backend: @echo "Starting the backend..." @python3 -m pipenv run uvicorn opendevin.server.listen:app --port $(BACKEND_PORT) --host 0.0.0.0

however now he sends the first step to LM studio but it remains stuck there it does not return the response to devin and the LM studio logs

lmstudio-server-log.txt

Ciao Flavio, mi aiuteresti ad installarlo su Windows? Io non ci sono riuscito, mi potresti postare una breve guida per favore?

[flavioTocca]

Hi, I doubt I'll be able to make a guide because I too have had many problems and in fact I haven't yet found a way to make it work with a local model, I don't know what your problem is, if you managed to emulate Linux with WLS, check that python on linux is equal to or higher than 3.11, after that these are just dependency problems my advice is to execute the make build commands in phases and not all at once, for example to download the node modules for the front end part not I relied on make build but I entered the folder and started npm install by hand, another thing windows sees python as python3 so the commands that call python must be replaced for example "python -m pip install pipenv" in "python3 -m pip install pipenv", a guide would be very long but if you have some specific problem and I can help you I'm happy to do so, in the photo I modified a file to give priority to the config rather than the webui select

[risingsunomi]

Will be writing up a powershell script later today to get this running. Would love feedback on testing it when done. I got OpenDevin up and working with OpenAI on Windows 11 via powershell and just installing python3.12 and LTS NPM.

One thing for the environment variables to temp add them using the toml file I will need to look at but was about to add them for the agent and front in via using $env:ENV_VAR_NAME before bring the front and back ends up manually

There is also an issue with changing the execution policy since pnpm is unsigned on windows and then also you need windows admin to run corepack enable

Here are the basic steps I did but will be more detailed in the powershell script

Before git clone of repo

1. python -m venv opendevin_env make sure >=python3.11
2. cd opendevin_env
3. .\Scripts\Activate.ps1
4. git clone repo in opendevin_env folder

In OpenDevin folder in opendevin_env

1. docker pull ghcr.io/opendevin/sandbox
2. pip install poetry
3. poetry install --without evaluation
4. poetry run pre-commit install --config ./dev_config/python/.pre-commit-config.yaml
5. cd frontend; node ./scripts/detect-node-version.js
6. Check if npm corepack is installed and if not npm install -g corepack
7. Run corepack enable in a administrator powershell to do permitted operation
8. cd frontend if not already in frontend
9. For running pnpm you have to change the execution policy Set-ExecutionPolicy RemoteSigned
10. pnpm install; pnpm run make-i18n

To run front end
cd frontend; $env:BACKEND_HOST="127.0.0.1:3000"; $env:FRONTEND_PORT=3001; pnpm run start

To run backend with OpenAI
$env:LLM_MODEL="gpt-3.5-turbo"; $env:LLM_API_KEY="..."; $env:LLM_EMBEDDING_MODEL="openai"; $env:WORKSPACE_DIR="./workspace"; poetry run uvicorn opendevin.server.listen:app --port 3000

[Redrum624]

I'll test your powershell script no problem whenever it's ready!

[risingsunomi]

I have created a separate repo for it but was able to get it running with these scripts. Let me know your thoughts and any issues

https://github.com/risingsunomi/OpenDevinPowerShell