Docker 1.0.16 image does not work with Postgres

[mieslep]

Bug Description

Running docker run -d langflowai/langflow:1.0.16 works just fine, but if you specify -e LANGFLOW_DATABASE_URL=postgres://...

you get errors:

Can't load plugin:                
sqlalchemy.dialects:postgres      

import_dbapi                      
    import psycopg2               
ModuleNotFoundError: No module    
named 'psycopg2'                  

Reproduction

Start a Docker instance for v1.0.16 that connects to a Postgres database using the LANGFLOW_DATABASE_URL environment vairable.

Expected behavior

Container starts and Langflow service runs as normal :)

Who can help?

No response

Operating System

Langflow Docker image

Langflow Version

1.0.16

Python Version

3.12

Screenshot

No response

Flow File

No response

[dosubot]

I found a similar open issue: No PostgreSQL driver in docker image for 1.0.16, which was last updated on August 22, 2024. The suggested solution is to modify the pyproject.toml file to include psycopg[binary]:

[tool.poetry.dependencies]
psycopg = { version = "3.1.9", extras = ["binary"] }

This change ensures that the binary version of the psycopg package is used, which can help with compatibility and performance.

Additionally, to resolve the issue with the Docker image, you can create a custom Dockerfile that installs the psycopg2 module. Here is an example:

1. Create a Dockerfile for the Langflow service:

   FROM langflowai/langflow:1.0-alpha
   
   # Install psycopg2
   RUN pip install psycopg2-binary
   
   # Copy the rest of your application code if needed
   # COPY . /app
   
   # Set the working directory
   WORKDIR /app
   
   # Expose the necessary port
   EXPOSE 7860
   
   # Define the entry point
   CMD ["python", "-m", "langflow", "run"]

2. Modify the docker-compose.yml to use the custom Dockerfile:

   version: "3.8"
   
   services:
     langflow:
       build:
         context: .
         dockerfile: Dockerfile
       ports:
         - "7860:7860"
       depends_on:
         - postgres
       environment:
         - LANGFLOW_DATABASE_URL=postgresql://langflow:langflow@postgres:5432/langflow
         - LANGFLOW_CONFIG_DIR=app/langflow
       volumes:
         - langflow-data:/app/langflow
   
     postgres:
       image: postgres:16
       environment:
         POSTGRES_USER: langflow
         POSTGRES_PASSWORD: langflow
         POSTGRES_DB: langflow
       ports:
         - "5432:5432"
       volumes:
         - langflow-postgres:/var/lib/postgresql/data
   
   volumes:
     langflow-postgres:
     langflow-data:

By building a custom Docker image that includes the psycopg2 module, you should be able to connect to the PostgreSQL database without encountering the ModuleNotFoundError ^[1][2].

To continue talking to Dosu, mention @dosu.

[italojohnny]

It seems the issue is related to PR #3292.

For now, if you can manually install the psycopg2 dependency in your container, I believe you should be able to continue your work until the problem is resolved