Tutorial on using tableau-api-lib and Tableau's REST API and Metadata API to pull underlying data sources across the entire server.

tableau-api-lib-metadata-underlying-data-sources.py

import pandas as pd

from tableau_api_lib import TableauServerConnection
from tableau_api_lib.utils.querying import get_sites_dataframe

SITE_NAMES = None  # if this variable is None then all sites will be queried

#  uncomment the line below if defining a subset of sites
# SITE_NAMES = ['SiteA', 'SiteB', 'SiteC']  # optional variable to define a subset of sites

DB_SERVER_QUERY = """
{
  databaseServers {
    name
    hostName
    port
    connectionType
    isEmbedded
  }
}"""

DATA_CONNECTIONS_QUERY = """
{
  databases {
    name
    __typename
    connectionType
  }
}"""

TS_CONFIG = {
        'my_env': {
                'server': 'https://YourTableauServer.com',
                'api_version': '<YOUR_API_VERSION>',
                'username': '<YOUR_USERNAME>',
                'password': '<YOUR_PASSWORD>',
                'site_name': '<YOUR_SITE_NAME>',
                'site_url': '<YOUR_SITE_CONTENT_URL>'
        }
}


conn = TableauServerConnection(config_json=TS_CONFIG, env='my_env')
conn.sign_in()

sites_df = get_sites_dataframe(conn)

if any(SITE_NAMES):
  sites_df = sites_df.loc[sites_df['name'].isin(SITE_NAMES)]
  
db_servers_df = pd.DataFrame()
data_connections_df = pd.DataFrame()

for index, site in sites_df.iterrows():
    conn.switch_site(site['contentUrl'])
    response_a = conn.metadata_graphql_query(DB_SERVER_QUERY)
    response_b = conn.metadata_graphql_query(DATA_CONNECTIONS_QUERY)
    site_db_servers_df = pd.DataFrame(response_a.json()['data']['databaseServers'])
    site_data_connections_df = pd.DataFrame(response_b.json()['data']['databases'])
    site_db_servers_df['site_name'] = site['name']
    site_data_connections_df['site_name'] = site['name']
    db_servers_df = db_servers_df.append(site_db_servers_df, sort=False, ignore_index=True)
    data_connections_df = data_connections_df.append(site_data_connections_df, sort=False, ignore_index=True)
    
combined_db_df = data_connections_df.merge(db_servers_df, how='left', on=['name', 'connectionType', 'site_name'])
combined_db_df_unique = combined_db_df.drop(columns=['name', 'site_name']).drop_duplicates()

conn.sign_out()