Module node_helpers 

Helper functions for building graph nodes with minimal boilerplate.

These utilities eliminate the manual JSON ↔ typed conversion that makes Rust examples verbose compared to Python’s langchain-core.

Functions

ask_json

Ask the LLM a single prompt and get back a parsed JSON value.

extract_messages

Extract typed messages from a graph state JSON, with an optional system prompt prepended.

get_i64

Get a field from state as i64, defaulting to 0.

get_str

Get a field from state as a string, defaulting to “”.

invoke_llm

Invoke an LLM and return a state update.

invoke_llm_with_config

Invoke an LLM with a custom config and return a state update.

llm_response_to_json

Convert a model response into a state update JSON.

parse_json_response

Strip markdown code fences (```json … ```) and parse the inner JSON.

print_result

Print the last AI message from an invoke / ainvoke result.

print_result_with_options

Like print_result but with explicit control over thinking display.

print_stream

Stream graph execution and print tokens to stdout in real-time.

print_stream_with_options

Like print_stream but with explicit control over thinking display.

response_text

Extract the assistant’s text reply from an invoke_llm / stream_llm result.

stream_llm

Stream LLM tokens via StreamWriter and return the final state update.