{graph_notation}

{graph_spec}

Here is the python State class associated with this graph. The Node Functions all take this state as a parameter. We import that state from this file. It is named 'state_code.py'

{state_code}

Nodes are specified in markdown files in sections for each node.

{node_spec}

Here are some examples of Node Functions:

{node_code_example}

If a node needs input from a human (i.e. it needs to ask the human), do it like this:

{human_input_example}

If a node needs to "show" or "display" a value, use function "human_display" to accomplish this:

human_display(variable_to_display)

Please generate a node function for each node documented in the Node Specification included above. For each node, write the function in this format:

Node Function

def node_function_name(state: state_class_name):

reads: list state fields read

writes: list state fields we update

.. put each field read into a variable, and only use that variable .. node specific processing return .. updates to state specified as a dictionary, exactly matches the fields we write as listed in "# writes: ..."

When creating prompt strings, always use triple quoted f-strings so they can span multiple lines, and always put those prompt strings in a variable before using them.

We are only writing Node Functions that are specified in the NODE_SPECIFICATION above. Import the State class from 'state_code.py'. This State class is the one referenced in each Node Function.

We write exactly one function for each Node Function in our NODE_SPECIFICATION. For example, if the NODE_SPECIFICATION describes 3 Node Functions, our output is exactly 3 python functions, plus the "human_display" function shown below:

def human_display(val): print(val)

Save code in file node_code.py

{question}