Matplotlib add to Code Node

[Namec999]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
• Please do not modify this template :) and fill in all the required fields.

Provide a description of requested docs changes

Hello All,
i want to add to my workflow some libraries like Matplotlib, for chart creation

is there any how to, that can make me add libraries to my dify install, and then use them in the code node using the difyUI worflow

best

[dosubot[bot]]

To add libraries like Matplotlib to your Dify installation and use them in a code node, you would typically manage dependencies using a tool like Poetry, which Dify uses for its API service. Poetry manages dependencies in a pyproject.toml file, similar to a requirements.txt file. You can install the required dependencies by executing poetry install and activate the environment using poetry shell ^[1][2].

However, specific instructions for adding libraries directly to a Dify code node or workflow are not detailed in the available documentation. It's important to note that Dify's execution environments are sandboxed for security reasons, which may impose limitations on the use of certain libraries or functions ^[3].

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}

[Namec999]

i have my instance runing on docker containers, wich one i need to log in to manage my dependencies

and also when installed, do a simple "import ... from" needed inmy code node to use the necessary library ?

[dosubot[bot]]

To manage dependencies for your Dify instance running on Docker containers, you should log into the sandbox Docker container. This container is specifically set up for handling dependencies. Here's how you can do it:

1. Access the sandbox container:

   docker exec -it <sandbox_container_id> /bin/sh

2. Once inside the container, you can install the necessary library using pip:

   pip install <library_name>

After installing the library, you can use it in your code node with a simple import statement like this:

from <library_name> import <module>

Make sure to replace <sandbox_container_id>, <library_name>, and <module> with the appropriate values for your use case ^[1].

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}

[Namec999]

sounds good !

is this code can work inside a Code Node ?

import matplotlib.pyplot as plt

def main(arg1: str, arg2: str) -> dict:

# Sample data
x = [1, 2, 3, 4, 5]
y = [10, 20, 25, 40, 50]

# Create the plot
plt.plot(x, y, marker='o', linestyle='-', color='b', label='Sample Line')

# Adding labels and title
plt.xlabel('X-axis')
plt.ylabel('Y-axis')
plt.title('Sample Graph')

# Display a legend
plt.legend()

# Save the plot as an image file
plt.savefig('sample_graph.png')

return {
    "result": plt,
}

and also how to make the output variable an image ?

[dosubot[bot]]

The code you've provided will not work as expected within a Dify Code Node due to the sandboxed environment restrictions. Specifically, direct file system access, such as saving a plot to a file with plt.savefig('sample_graph.png'), is not allowed ^[1].

To handle image data as output in a Dify Code Node, you can convert the plot to a binary format (like a PNG image) and return it as a base64-encoded string. Here's an example of how you can modify your code to achieve this:

import matplotlib.pyplot as plt
import io
import base64

def main(arg1: str, arg2: str) -> dict:
    # Sample data
    x = [1, 2, 3, 4, 5]
    y = [10, 20, 25, 40, 50]

    # Create the plot
    plt.plot(x, y, marker='o', linestyle='-', color='b', label='Sample Line')

    # Adding labels and title
    plt.xlabel('X-axis')
    plt.ylabel('Y-axis')
    plt.title('Sample Graph')

    # Display a legend
    plt.legend()

    # Save the plot to a bytes buffer
    buf = io.BytesIO()
    plt.savefig(buf, format='png')
    buf.seek(0)

    # Encode the image to base64
    image_base64 = base64.b64encode(buf.read()).decode('utf-8')

    return {
        "result": image_base64,
    }

This code saves the plot to an in-memory buffer and then encodes it as a base64 string, which can be returned as an output variable. You can then handle this base64 string in your application to display or process the image.

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}

[Namec999]

thank you,

when using you code in code Node i got this error :
ModuleNotFoundError: No module named 'matplotlib'
error: exit status 255