graph

A graph has two meanings, depending on context. In one context a graph represent the set of nodes that are evaluated by the evaluation engine. In the coding context, the graph represents a wiring function that describes the relationship between nodes.

We use decorators to mark functions to indicate the intent of the function.

In this case a graph looks as follows:

@graph
def my_graph():
    ...

All function signatures in hgraph are required to be typed using type-hints. For example:

@graph
def my_graph(ts: TS[int], v: str) -> TS[str]:
    ...

In this example we take in two inputs and return a time-series value of type string.

Code in the graph is only ever evaluated at wiring time (when the graph is built) and prior to evaluation. This describes the relationships between nodes.

A simple example is below:

from hgraph import graph, const, debug_print

@graph
def my_graph():
    c = const("World")
    debug_print("Hello", c)

In this scenario we are constructing two nodes, the const which will tick one at the start of the evaluation and the output of the const this output is then passed to the input of debug_print (another node). This produces a graph with two nodes and one edge.