python-drain

Python Drain

Demonstrates queue draining and non-blocking receive -- useful for clearing buffered messages or polling without blocking.

Architecture

timer (10ms) --> send_data --> data --> receive_data_with_sleep

Nodes

send_data (send_data.py) -- Sends 100 messages containing uint64 nanosecond timestamps at 10ms intervals. The high-frequency sender fills the receiver's input queue while the receiver sleeps.

receive_data_with_sleep (receive_data.py) -- Sleeps 1 second (allowing ~100 messages to queue up), then calls node.drain() to consume all pending events at once. After draining, loops 100 times calling node.try_recv() for non-blocking receive with error handling.

Key API Methods

• node.drain() -- returns all queued events immediately without blocking
• node.try_recv() -- returns one event if available, raises error if empty (non-blocking)
• queue_size: 100 -- YAML setting that controls input buffer capacity

Run

adora run dataflow.yaml

What This Demonstrates

Feature	Where
node.drain() for batch consumption	Receiver
node.try_recv() for non-blocking poll	Receiver
queue_size for input buffer sizing	YAML (both nodes)
High-frequency producer with slow consumer	Sender vs receiver