feat(core): complete the internal representation (imports, model file, model manager) by yash-kumarx · Pull Request #12 · accordproject/concerto-rust

yash-kumarx · 2026-06-20T18:51:15Z

Closes #1264, closes #1269

Finishes the internal representation, on top of #9.

Changes

Import: typed model imports (wildcard ns.*, named, and aliased).
ModelFile: owns one namespace's declarations and imports, indexes them, and
resolves names it can see locally (primitives, local types, named imports).
ModelManager: loads model files and resolves types across namespaces
(incl. wildcard imports), walks the inheritance chain to collect a type's
properties, and checks assignability. Preloads the concerto@1.0.0 system model.
Strict AST handling consistent with feat(core): scaffold workspace and add the typed leaf model #9 (rejects missing $class and non-array
imports/declarations).

Flags

A shared from_json trait (plus a derive macro in concerto-macros) is a
sensible follow-up, as discussed; kept as methods here.

Related Issues

Author Checklist

Ensure you provide a DCO sign-off for your commits using the --signoff option of git commit.
Vital features and changes captured in unit and/or integration tests
Commits messages follow AP format
Extend the documentation, if necessary
Merging to main from fork:yash/introspection-rest

Type a model's imports (wildcard, named, aliased), and group one namespace's declarations and imports into a ModelFile that resolves the names it can see locally. Signed-off-by: Yash Kumar <yashr17042006@gmail.com>

Resolve types across namespaces, walk the inheritance chain to collect a type's properties, and check assignability. Preload the concerto@1.0.0 system model that every user type extends. Signed-off-by: Yash Kumar <yashr17042006@gmail.com>

ekarademir · 2026-06-23T09:24:49Z

Ah the issue references are not working by just number references, because they are on the main repo :( you can add direct links.

ekarademir

I've reviewed about half way to model file. Left some comments that would simplify your life. You can address them until I finish the rest.

ekarademir · 2026-06-23T09:34:08Z

+    Type {
+        /// The namespace the type is imported from.
+        namespace: String,
+        /// The imported type's short name.


It's not the shortname it's the type.

Fixed that, doc now reads "The name of the imported type."

ekarademir · 2026-06-23T09:34:26Z

+    Types {
+        /// The namespace the types are imported from.
+        namespace: String,
+        /// The imported short type names.


The same here, they are types

ekarademir · 2026-06-23T09:40:35Z

+
+    /// `true` for a wildcard (`ns.*`) import, which can only be resolved by
+    /// consulting the declarations of the imported namespace.
+    pub fn is_wildcard(&self) -> bool {


With the new version of concerto, wildcard imports are throwing an error https://github.com/accordproject/concerto/blob/542b630a18b3febaffd67f94a366d4b6175fa79b/packages/concerto-core/src/introspect/modelfile.ts#L745

Maybe we can also do that instead of trying to make them work.

ekarademir · 2026-06-23T09:42:07Z

+        match self {
+            Self::All { .. } => None,
+            Self::Type { namespace, name } if name == short => Some(qualify(namespace, name)),
+            Self::Type { .. } => None,


I think this corresponds to non-strict mode. Which is another thing we don't allow with v4 https://github.com/accordproject/concerto/blob/542b630a18b3febaffd67f94a366d4b6175fa79b/packages/concerto-core/src/introspect/modelfile.ts#L700

Looks like there was some code that got in for non-strict code with the previous PR that allows non-strict mode

concerto-rust/concerto-core/src/model_util.rs

Line 66 in 2b69b2e

pub version: Option<String>,

You can remove them as well. Or in a separate PR.

The non-strict path in our code was the wildcard resolution, so removing that makes resolution strict: only primitives, local declarations, and named imports resolve now

Looks like there was some code that got in for non-strict code with the previous PR that allows non-strict mode

concerto-rust/concerto-core/src/model_util.rs

Line 66 in 2b69b2e

pub version: Option<String>,

You can remove them as well. Or in a separate PR.

Versions are mandatory in v4, so I'll strip the version-insensitive matching in a separate PR.

Concerto v4 disallows wildcard imports and non-strict resolution. Reject `import ns.*` while parsing (this was the non-strict path) and drop the wildcard handling in Import and the model manager. Also fix the import field docs to say 'type' rather than 'short name'. Signed-off-by: Yash Kumar <yashr17042006@gmail.com>

yash-kumarx · 2026-06-24T07:09:12Z

I've addressed all of these in 4e1e484, thanks

ekarademir · 2026-07-06T08:05:26Z

+    }
+
+    /// True if this is the built-in `concerto` system namespace.
+    pub fn is_system(&self) -> bool {


this is a bit ambiguous, it could be more clear

Suggested change

pub fn is_system(&self) -> bool {

pub fn is_system_namespace(&self) -> bool {

ekarademir · 2026-07-06T08:05:55Z

+    /// Resolves a short name from what this file declares or imports: the
+    /// primitives, its own declarations, and its named imports. Returns `None`
+    /// if the name is none of those.
+    pub fn resolve_local(&self, short: &str) -> Option<String> {


This could also be more clear

Suggested change

pub fn resolve_local(&self, short: &str) -> Option<String> {

pub fn resolve_local_type(&self, short: &str) -> Option<String> {

ekarademir · 2026-07-06T08:14:32Z

+}
+
+/// The `concerto@1.0.0` system model, as a JSON AST.
+pub fn root_model_ast() -> serde_json::Value {


This is not correct. All it's doing is serialising some JSON types. Instead it should directly use the JSON file here https://github.com/accordproject/concerto-metamodel/blob/main/lib/metamodel.json (or better, the raw version https://raw.githubusercontent.com/accordproject/concerto-metamodel/refs/heads/main/lib/metamodel.json)

This could be part of the concerto-metamodel crate, instead of core.

Vendored the canonical JSON instead of building the AST in code.
One note: metamodel.json defines concerto.metamodel@1.0.0, the metamodel types themselves. The file that defines concerto@1.0.0 is rootmodel.json in concerto-core, so that's the one vendored here, unchanged.

Moving it into the concerto-metamodel crate would conflict with the build script PR right now, but I can do that as a follow-up once both merge. Preloading the metamodel namespace itself I'd defer to the validation work, where it's actually needed.

Signed-off-by: Yash Kumar <yashr17042006@gmail.com>

Loads concerto@1.0.0 from the rootmodel.json that ships with concerto-core instead of building the AST in code. Signed-off-by: Yash Kumar <yashr17042006@gmail.com>

yash-kumarx added 2 commits June 18, 2026 15:22

feat(core): add imports and the model file

3147a13

Type a model's imports (wildcard, named, aliased), and group one namespace's declarations and imports into a ModelFile that resolves the names it can see locally. Signed-off-by: Yash Kumar <yashr17042006@gmail.com>

ekarademir reviewed Jun 23, 2026

View reviewed changes

ekarademir requested changes Jul 6, 2026

View reviewed changes

yash-kumarx added 2 commits July 8, 2026 16:31

refactor(core): clarify model file method names per review

3f8d6e8

Signed-off-by: Yash Kumar <yashr17042006@gmail.com>

refactor(core): vendor the canonical root model json

d0921e6

Loads concerto@1.0.0 from the rootmodel.json that ships with concerto-core instead of building the AST in code. Signed-off-by: Yash Kumar <yashr17042006@gmail.com>

	pub fn is_system(&self) -> bool {
	pub fn is_system_namespace(&self) -> bool {

	pub fn resolve_local(&self, short: &str) -> Option<String> {
	pub fn resolve_local_type(&self, short: &str) -> Option<String> {

Uh oh!

Conversation

yash-kumarx commented Jun 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Changes

Flags

Related Issues

Author Checklist

Uh oh!

ekarademir commented Jun 23, 2026

Uh oh!

ekarademir left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

yash-kumarx commented Jun 24, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

yash-kumarx commented Jun 20, 2026 •

edited

Loading