There are various functions available in the pre-installed sm_udf library for common use cases.
Data Access
Database
sm_udf.app_builder.odbc.get_sql_engine()
Returns a sqlalchemy.engine.Engine that is set up to connect to the database.
Usage:
from sm_udf.app_builder.odbc import get_sql_engine()
engine = get_sql_engine()
with engine.connect() as conn:
conn.execute(sqlalchemy_query)Sight Machine SDK
sm_udf.app_builder.smsdk.get_smsdk_client()
Returns an smsdk.client.Client that is set up to connect to the Sight Machine API.
Usage:
from sm_udf.app_builder.smsdk import get_smsdk_client
client = get_smsdk_client()Output Formatting
Graphs & Charts
sm_udf.app_builder.util.setup_plotly_theme()
Sets up a custom theme for Plotly that matches the look and feel of other Sight Machine applications and dashboards.
Usage:
from sm_udf.app_builder.util import setup_plotly_theme
setup_plotly_theme()Final Output
sm_udf.app_builder.util.return_app_builder_results(results)
Formats your results for final output. This is required to be the final line of the final cell in your notebook, and results must be a Python dictionary containing only JSON-serializable values.
Usage:
from sm_udf.app_builder.util import return_app_builder_results
return_app_builder_results(response)Encode Page Linking for Markdown Panels
sm_udf.app_builder.util.create_markdown_encoded_data_link()
Help encode the link_mapping to pass to data-link for creating a link element in a markdown panel.
Usage:
from sm_udf.app_builder.util import create_markdown_encoded_data_link
link_encoded = create_markdown_encoded_data_link(link_mapping)
markdown = f"""
<span data-link="{link_encoded}">View Chart</span>
"""
markdown_panel = {
"id": "markdown-panel",
"viz_type": "markdown",
"viz_props": {
"markdown": markdown
}
}