Just got started, and MetadataValue seems a bit verbose. Any efforts to simplify?

[benjameep]

I just got started on Dagster working through the introduction tutorial, and noticed that the method of adding a plot to the metadata is really verbose:

import base64
from io import BytesIO
import matplotlib.pyplot as plt
from dagster import MetadataValue

...

    # Make a bar chart of the top 25 words
    plt.figure(figsize=(10, 6))
    plt.bar(top_words.keys(), top_words.values())
    plt.xticks(rotation=45, ha="right")
    plt.title("Top 25 Words in Hacker News Titles")
    plt.tight_layout()
        
    # Convert the image to a saveable format
    buffer = BytesIO()
    plt.savefig(buffer, format="png")
    image_data = base64.b64encode(buffer.getvalue())

    # Convert the image to Markdown to preview it within Dagster
    md_content = f"![img](data:image/png;base64,{image_data.decode()})"

    # Attach the Markdown content as metadata to the asset
    context.add_output_metadata(metadata={"plot": MetadataValue.md(md_content)})

Maybe this isn't actually used that often, but are there any efforts, (third-party plugins or future dagster roadmap) to simplify this process?

I would love to see something more similar to streamlit's api for outputting content:

import streamlit as st
import numpy as np
import pandas as pd

chart_data = pd.DataFrame(
     np.random.randn(20, 3),
     columns=['a', 'b', 'c'])

# display markdown
st.markdown("## Here is our data:")

# display a pandas table
st.dataframe(chart_data)

# create a bar chart and display it
st.bar_chart(chart_data)

fig = plt.figure(figsize=(10, 6))
plt.bar(chart_data.index, chart_data.a)
plt.xticks(rotation=45, ha="right")
plt.title("Top 25 Words in Hacker News Titles")
plt.tight_layout()

# display a matplotlib plot
st.pyplt(fig)

Hoping my input as someone just starting out will more helpful than annoying 🙃

[sryza]

Just to make it uber-clear, this is the part that's cumbersome, right?

    # Convert the image to a saveable format
    buffer = BytesIO()
    plt.savefig(buffer, format="png")
    image_data = base64.b64encode(buffer.getvalue())

    # Convert the image to Markdown to preview it within Dagster
    md_content = f"![img](data:image/png;base64,{image_data.decode()})"

    # Attach the Markdown content as metadata to the asset
    context.add_output_metadata(metadata={"plot": MetadataValue.md(md_content)})

I agree that this is a lot of code. Would you be up for filing a github issue for making it simpler to include a plot in asset metadata?

[benjameep]

Sure I'll file a Github issue,

[benjameep]

Created #15468

[sryza]

🙏 🙏