Implement a custom SELECT with SAMPLE BY support

[jerrinot]

TODO:

• align to ... + offsets
• polish API
• tests, including both sqlalchemy 1.4 and 2.0 and in combination with other clauses (GROUP BY etc)
• documentation

Example usage:

from sqlalchemy import create_engine, MetaData, Table, Column
from questdb_connect import (
    Timestamp,
    Double,
    Symbol,
)
from sqlalchemy import func
from questdb_connect import select

engine = create_engine('questdb://admin:quest@localhost:8812/main')
metadata = MetaData()

# Define a table for sensor readings
sensors = Table(
    'sensors',
    metadata,
    Column('ts', Timestamp),
    Column('temperature', Double),
    Column('humidity', Double),
    Column('location', Symbol),
)

def main():
    metadata.create_all(engine)

    location_samples = select(
        sensors.c.ts,
        func.avg(sensors.c.temperature).label('avg_temp'),
        func.min(sensors.c.temperature).label('min_temp'),
        func.max(sensors.c.temperature).label('max_temp')
    ).where(
        sensors.c.location == 'warehouse'
    ).sample_by(1, 'd');

    with engine.connect() as conn:
        for row in conn.execute(location_samples).fetchall():
            print(f"Time: {row.ts}, Average Temp: {row.avg_temp}, Minimal Temp: {row.min_temp}, Maximal Temp: {row.max_temp}")

if __name__ == '__main__':
    main()

[jerrinot]

@nwoolmer I'm making this a non-draft even docs is still missing, I believe impl is now alright.

your review is very much appreciated!
some points:

1. sample_by() receives a lot of arguments. most of them are optional and given Python supports named parameters it's probably fine. wdyt? You certainly know the Python ecosystem and customers better than I do
2. not sure syntex for sampling intervals. right now the arguments are separated as value and unit. wdyt?
3. I did not implement any kind of a validation. for example FROM-TO does not support some FILL options, etc. so the extension could validate this on a client side. otoh: we would need to change this if/when we implement it on server-side.