Base: Access Tables via Python

You can access the databases and tables within the Base module using Python from your local machine. Once retrieved as e.g. pandas object, the data can be processed further. In this tutorial, we will describe how you could create a Python script which will retrieve the data and visualize it using Dash framework. The script will contain the following parts:

• Importing dependencies and variables.

• Function to fetch the data from Base table.

• Creating and running the Dash app.

Importing dependencies and variables

This part of the code imports the dependencies which have to be installed on your machine (possibly with pip). Furthermore, it imports the variables API_KEY and PROJECT_ID from the file named config.

from dash import Dash, html, dcc, callback, Output, Input
import plotly.express as px

from config import API_KEY, PROJECT_ID
import requests
import snowflake.connector
import pandas as pd

Function to fetch the data from Base table

We will be creating a function called fetch_data to obtain the data from Base table. It can be broken into several logically separated parts:

• Retrieving the token to access the Base table together with other variables using API.

• Establishing the connection using the token.

• SQL query itself. In this particular example, we are extracting values from two tables Demo_Ingesting_Metrics and BB_PROJECT_PIPELINE_EXECUTIONS_DETAIL. The table Demo_Ingesting_Metrics contains various metrics from DRAGEN analyses (e.g. the number of bases with quality at least 30 Q30_BASES) and metadata in the column ica which needs to be flattened to access the value Execution_reference. Both tables are joined on this Execution_reference value.

• Fetching the data using the connection and the SQL query.

Here is the corresponding snippet:

def fetch_data():
    # Your data fetching and processing code here
    # retrieving the Base oauth token
    url = 'https://ica.illumina.com/ica/rest/api/projects/' + PROJECT_ID +  '/base:connectionDetails'

    # set the API headers
    headers = {
                'X-API-Key': API_KEY,
                'accept': 'application/vnd.illumina.v3+json'
                }

    response = requests.post(url, headers=headers)
    ctx = snowflake.connector.connect(
        account=response.json()['dnsName'].split('.snowflakecomputing.com')[0],
        authenticator='oauth',
        token=response.json()['accessToken'], 
        database=response.json()['databaseName'],
        role=response.json()['roleName'],
        warehouse=response.json()['warehouseName']
    )
    cur = ctx.cursor()
    sql = '''
    WITH flattened_Demo_Ingesting_Metrics AS (
        SELECT 
            flattened.value::STRING AS execution_reference_Demo_Ingesting_Metrics,
            t1.SAMPLEID,
            t1.VARIANTS_TOTAL_PASS,
            t1.VARIANTS_SNPS_PASS,
            t1.Q30_BASES,
            t1.READS_WITH_MAPQ_3040_PCT
        FROM 
            Demo_Ingesting_Metrics t1,
            LATERAL FLATTEN(input => t1.ica) AS flattened
        WHERE 
            flattened.key = 'Execution_reference'
    ) SELECT 
        f.execution_reference_Demo_Ingesting_Metrics,
        f.SAMPLEID,
        f.VARIANTS_TOTAL_PASS,
        f.VARIANTS_SNPS_PASS,
        t2."EXECUTION_REFERENCE",
        t2.END_DATE,
        f.Q30_BASES,
        f.READS_WITH_MAPQ_3040_PCT
    FROM 
        flattened_Demo_Ingesting_Metrics f
    JOIN 
        BB_PROJECT_PIPELINE_EXECUTIONS_DETAIL t2
    ON 
        f.execution_reference_Demo_Ingesting_Metrics = t2."EXECUTION_REFERENCE";
    '''

    cur.execute(sql)
    data = cur.fetch_pandas_all()
    return data

df = fetch_data()

Creating and running the Dash app

Once the data is fetched, it is visualized in an app. In this particular example, a scatter plot is presented with END_DATE as x axis and the choice of the customer from the dropdown as y axis.

app = Dash(__name__)
#server = app.server


app.layout = html.Div([
    html.H1("My Dash Dashboard"),
    
    html.Div([
        html.Label("Select X-axis:"),
        dcc.Dropdown(
            id='x-axis-dropdown',
            options=[{'label': col, 'value': col} for col in df.columns],
            value=df.columns[5]  # default value
        ),
        html.Label("Select Y-axis:"),
        dcc.Dropdown(
            id='y-axis-dropdown',
            options=[{'label': col, 'value': col} for col in df.columns],
            value=df.columns[2]  # default value
        ),
    ]),
    
    dcc.Graph(id='scatterplot')
])


@callback(
    Output('scatterplot', 'figure'),
    Input('y-axis-dropdown', 'value')
)
def update_graph(value):
    return px.scatter(df, x='END_DATE', y=value, hover_name='SAMPLEID')

if __name__ == '__main__':
    app.run(debug=True)

Now we can create a single Python script called dashboard.py by concatenating the snippets and running it. The dashboard will be accessible in the browser on your machine.