feat: get full evaluation back online

Very hacky, but this is enough to be finished with the graph IR for now.

crates/ir/src/lib.rs

@@ -27,6 +27,8 @@ pub fn from_ron(source: &str) -> ron::error::SpannedResult<GraphIr> {
 
 /// The toplevel representation of a whole pipeline.
 ///
+/// # DAGs
+///
 /// Pipelines may not be fully linear. They may branch out and recombine later on.
 /// As such, the representation for them which is currently used is a
 /// [**D**irected **A**cyclic **G**raph](https://en.wikipedia.org/wiki/Directed_acyclic_graph).
@@ -56,6 +58,27 @@ pub fn from_ron(source: &str) -> ron::error::SpannedResult<GraphIr> {
 /// So the vertices of the DAG are the **sockets**
 /// (which are either [`id::Input`] or [`id::Output`] depending on the direction),
 /// and each **socket** in turn belongs to an **instruction**.
+///
+/// # Usage
+///
+/// - If you want to build one from scratch,
+///   add a few helper methods like
+///     constructing an empty one,
+///     adding instructions and
+///     adding edges
+/// - If you want to construct one from an existing repr,
+///   maybe you want to use [`semi_human::GraphIr`].
+///
+/// # Storing additional data
+///
+/// Chances are the graph IR seems somewhat fit to put metadata in it.
+/// However, most likely you're interacting in context of some other system,
+/// and also want to manage and index that data on your own.
+///
+/// As such, consider using _secondary_ maps instead.
+/// That is, store in a data structure _you_ own a mapping
+/// from [`id`]s
+/// to whatever data you need.
 #[derive(Clone, Debug, Default, PartialEq, Eq, Deserialize, Serialize)]
 pub struct GraphIr {
     /// "Backbone" storage of all **instruction** IDs to