Async vs Sync Usage in GraphBit Python API¶

GraphBit's Python bindings provide both synchronous (blocking) and asynchronous (non-blocking) interfaces for key operations, allowing you to choose the best fit for your application—whether it's a quick script, a web server, or a data pipeline.

Overview¶

Synchronous (Sync) functions block the current thread until the operation completes. They are simple to use in scripts and REPLs.
Asynchronous (Async) functions return immediately and run in the background, allowing your program to do other work or handle many tasks concurrently. Async is ideal for web servers, pipelines, and high-throughput applications.

Supported Sync and Async Functions¶

Module¶

Function	Type
`init`	Sync
`version`	Sync
`get_system_info`	Sync
`health_check`	Sync
`configure_runtime`	Sync
`shutdown`	Sync

LLM Client¶

Function	Type
`complete`	Sync
`get_stats`	Sync
`complete_async`	Async
`complete_stream`	Async
`complete_batch`	Async
`warmup`	Async

Embedding Client¶

Function	Type
`embed`	Sync
`embed_many`	Sync
`similarity`	Sync

Workflow¶

Function	Type
`add_node`	Sync
`connect`	Sync
`validate`	Sync

Workflow Executor¶

Function	Type
`configure`	Sync
`get_stats`	Sync
`reset_stats`	Sync
`get_execution_mode`	Sync
`execute`	Async
`run_async`	Async

Workflow Result¶

Function	Type
`is_success`	Sync
`is_failed`	Sync
`state`	Sync
`execution_time_ms`	Sync
`get_variable`	Sync
`get_all_variables`	Sync
`variables`	Sync
`get_node_output`	Sync
`get_all_node_outputs`	Sync

Usage Examples¶

Synchronous Usage¶

import os

from graphbit import LlmConfig, LlmClient

config = LlmConfig.openai(os.getenv("OPENAI_API_KEY"))

client = LlmClient(config)
result = client.complete("Hello, world!")
print(result)

Asynchronous Usage¶

import os
import asyncio

from graphbit import LlmConfig, LlmClient

config = LlmConfig.openai(os.getenv("OPENAI_API_KEY"))

client = LlmClient(config)

async def main():
    result = await client.complete_async("Hello, async world!")
    print(result)

    # Streaming completion
    stream_result = await client.complete_stream("Stream this!")
    print(stream_result)

asyncio.run(main())

This documentation demonstrates sync and async functions supported by GraphBit’s Python API, enabling you to select the best approach for your use case, whether you need straightforward synchronous calls or high-performance asynchronous workflows.