Metrics Layer using Hamilton

metrics_layer.py

### metrics_layer.py

import pandas as pd
from hamilton import driver

def workspaces(db: str, table: str='workspaces') -> pd.DataFrame:
    return query("SELECT * from {db}.{table}".format(db)) 

def count_workspaces(workspaces: pd.DataFrame, grain: str) -> pd.Series:
    """Pandas code that does...
    1. Group by grain
    2. Count(distinct)
    3. Return time-series of workspace counts
    """
    return # the pandas code described above

def count_active_workspaces(workspaces: pd.DataFrame, grain: str) -> pd.Series:
    """Pandas code that does...
    1. Filter to active
    2. Group by grain
    3. Return time-series of active workspace counts 
    """
    return # the pandas code described above

def activation_rate(count_workspaces: pd.Series, count_active_workspaces: pd.Series) -> pd.Series:
    return 100*count_workspaces/count_active_workspaces


### driver.py
driver = driver.Driver(config={'db' : ...})
daily_activation_rate = driver.execute(['activation_rate'], inputs={'grain': 'day'})
weekly_activation_rate = driver.execute(['activation_rate'], inputs={'grain': 'week'})
monthly_activation_rate = driver.execute(['activation_rate'], inputs={'grain': 'month'})
# There is another approach or two that put them all into the same execution context,
# But this was clearer to demonstrate