Unify EQBSL and CATHEXIS Rust crates into a shared workspace by Copilot · Pull Request #22 · Steake/EQBSL

Copilot · 2026-03-13T12:25:33Z

This branch had diverged from main in a way that dropped the CATHEXIS Rust crate while introducing a separate EQBSL crate. This update restores the CATHEXIS codepath from main and integrates both Rust codebases as a layered, non-duplicative system: EQBSL as the trust core, CATHEXIS as the semantic trust-handle layer built on top.

Workspace + crate topology
- Added a root Cargo workspace for rust/ and cathexis-rs/
- Restored cathexis-rs/ and the CATHEXIS manual from main
- Updated repository docs to describe the two-crate architecture clearly
Shared trust core
- Wired cathexis-rs to depend on the EQBSL crate via eqbsl = { path = "../rust" }
- Reused the shared EQBSL evidence/opinion mapping instead of keeping a second implementation in CATHEXIS
- Kept CATHEXIS focused on feature assembly, categorisation, and labeling
Integration layer
- Added a narrow adapter from eqbsl::BasicEmbedding into CATHEXIS TrustEmbedding
- Preserved CATHEXIS-facing types and APIs while moving shared trust math behind the EQBSL crate boundary
Docs + examples
- Updated the root README and CATHEXIS manual to reflect the layered Rust architecture
- Clarified that rust/ is the core trust/evidence library and cathexis-rs/ is the higher-level trust-handle pipeline
Focused hardening
- Added small unit tests around CATHEXIS activation helpers, labeling, embedding conversion, and summary construction to pin the integration behavior

Example of the new layering:

# cathexis-rs/Cargo.toml
[dependencies]
eqbsl = { path = "../rust" }

// cathexis-rs/src/core.rs
pub fn to_opinion(&self, base_rate: f64) -> Opinion {
    calculate_opinion(self.r, self.s, self.k, base_rate).into()
}

📱 Kick off Copilot coding agent tasks wherever you are with GitHub Mobile, available on iOS and Android.

superconductor-for-github · 2026-03-13T12:25:37Z

✅ @Copilot — Superconductor finished — View implementation

Standing by for instructions.

Co-authored-by: Steake <530040+Steake@users.noreply.github.com>

Copilot

Pull request overview

This PR restores and integrates the CATHEXIS Rust crate alongside the existing EQBSL Rust crate by introducing a root Cargo workspace and documenting a layered architecture (EQBSL core → CATHEXIS semantic trust-handle layer).

Changes:

Added a root Cargo workspace including rust/ (eqbsl) and cathexis-rs/ (cathexis), with updated ignore rules and a consolidated lockfile.
Introduced the cathexis-rs crate (pipeline, categoriser, labeling, feature model) wired to the EQBSL core via a path dependency and an embedding adapter.
Updated repository documentation to describe and link the two-crate architecture; added basic examples and tests for the CATHEXIS integration surface.

Reviewed changes

Copilot reviewed 15 out of 17 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
`Cargo.toml`	Defines the new workspace members (`rust`, `cathexis-rs`).
`Cargo.lock`	New workspace lockfile capturing dependency graph for both crates.
`.gitignore`	Ignores workspace-level and per-crate Rust build outputs.
`README.md`	Documents the layered Rust architecture and workspace build/test commands.
`rust/README.md`	Clarifies `rust/` as the shared trust core consumed by `cathexis-rs`.
`docs/CATHEXIS-Manual.md`	Adds CATHEXIS architecture + usage manual.
`cathexis-rs/Cargo.toml`	Adds CATHEXIS crate manifest + dependency on `eqbsl` via path.
`cathexis-rs/README.md`	Adds crate-level usage and module overview.
`cathexis-rs/src/*`	Implements CATHEXIS pipeline, categoriser, labeling, features, and EQBSL adapter types.
`cathexis-rs/examples/basic.rs`	Adds a runnable demo for the CATHEXIS pipeline.
`cathexis-rs/tests/pipeline_test.rs`	Adds an integration-style test covering the pipeline flow.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

cathexis-rs/src/pipeline.rs

+    pub fn batch_process(&mut self) -> Result<(), String> {
+        let nodes = self.graph.get_nodes();
+        let mut category_assignments: HashMap<String, usize> = HashMap::new();
+        let mut category_features: HashMap<usize, Vec<FeatureState>> = HashMap::new();
+
+        // 1. Compute features and assign categories
+        for agent_id in &nodes {
+            let features = self.graph.compute_features(agent_id)?;
+            let category_id = self.categoriser.predict(&features)?;
+
+            category_assignments.insert(agent_id.clone(), category_id);
+            category_features.entry(category_id).or_default().push(features);
+        }
+
+        // 2. Build summaries and generate labels
+        for (category_id, features_list) in category_features {
+            // Build summary (simplified implementation)
+            let summary = self.build_category_summary(category_id, &features_list);
+
+            // Generate or refresh the current category label from the summary.
+            let label_info = self.labeler.generate_label(&summary)?;
+            self.category_labels.insert(category_id, label_info);
+        }


cathexis-rs/src/pipeline.rs

+
+        // Convert probs to hard category
+        let category_id = self.categoriser.predict(&features)?;
+


cathexis-rs/src/categoriser.rs

+    fn predict(&self, features: &FeatureState) -> Result<usize, String> {
+        let probs = self.forward(features)?;
+        probs
+            .iter()
+            .enumerate()
+            .max_by(|(_, a), (_, b)| a.partial_cmp(b).unwrap_or(std::cmp::Ordering::Equal))
+            .map(|(index, _)| index)
+            .ok_or_else(|| "Empty probability vector".to_string())
+    }


cathexis-rs/src/categoriser.rs

+fn softmax(x: &Array1<f64>) -> Array1<f64> {
+    let max = x.iter().fold(f64::NEG_INFINITY, |a, &b| a.max(b));
+    let exp = x.mapv(|v| (v - max).exp());
+    let sum = exp.sum();


cathexis-rs/src/core.rs

+        Self { r, s, k }
+    }
+
+    /// Maps evidence to a Subjective Logic opinion (Equation 3).
+    /// b = r / (r + s + K)
+    /// d = s / (r + s + K)
+    /// u = K / (r + s + K)
+    pub fn to_opinion(&self, base_rate: f64) -> Opinion {


cathexis-rs/Cargo.toml

+serde_json = "1.0"
+thiserror = "1.0"
+rand = "0.8"
+ndarray = { version = "0.15", features = ["serde"] } # Used for categoriser weights, activations, and probability vectors.
+uuid = { version = "1.0", features = ["v4", "serde"] }
+chrono = { version = "0.4", features = ["serde"] }


cathexis-rs/src/pipeline.rs

+        let nodes = self.graph.get_nodes();
+        let mut category_assignments: HashMap<String, usize> = HashMap::new();
+        let mut category_features: HashMap<usize, Vec<FeatureState>> = HashMap::new();
+
+        // 1. Compute features and assign categories
+        for agent_id in &nodes {
+            let features = self.graph.compute_features(agent_id)?;
+            let category_id = self.categoriser.predict(&features)?;
+
+            category_assignments.insert(agent_id.clone(), category_id);
+            category_features.entry(category_id).or_default().push(features);
+        }


Steake · 2026-03-13T12:46:38Z

@copilot apply changes based on the comments in this thread

Co-authored-by: Steake <530040+Steake@users.noreply.github.com>

Copilot · 2026-03-13T12:52:20Z

@copilot apply changes based on the comments in this thread

Applied all review feedback in commit 29980eb:

pipeline.rs: Clear category_labels at start of batch_process to avoid stale labels
pipeline.rs: Use argmax on probs_array instead of calling predict twice for consistency
pipeline.rs: Remove unused category_assignments variable
categoriser.rs: Handle NaN probabilities in predict by returning an error
categoriser.rs: Validate inputs are finite in softmax and return Result
core.rs: Validate k > 0 and r,s >= 0 in Evidence::new, returning Result
Cargo.toml: Remove unused dependencies (serde_json, thiserror, rand, uuid, chrono)

All 19 workspace tests pass.

Initial plan

48feaa9

Copilot AI assigned Copilot and Steake Mar 13, 2026

Copilot AI mentioned this pull request Mar 13, 2026

Implement Rust crate version of EQBSL library with docs, examples, and tests #17

Open

Copilot started work on behalf of Steake March 13, 2026 12:25 View session

Copilot AI and others added 2 commits March 13, 2026 12:32

Integrate cathexis and EQBSL Rust crates

8f56914

Co-authored-by: Steake <530040+Steake@users.noreply.github.com>

Finalize Rust crate integration with cathexis

1a4ea2b

Co-authored-by: Steake <530040+Steake@users.noreply.github.com>

Copilot AI changed the title ~~[WIP] [Superconductor Ticket] Implement Rust crate version of EQBSL library~~ Unify EQBSL and CATHEXIS Rust crates into a shared workspace Mar 13, 2026

Copilot AI requested a review from Steake March 13, 2026 12:39

Copilot finished work on behalf of Steake March 13, 2026 12:39

Steake marked this pull request as ready for review March 13, 2026 12:40

Copilot AI review requested due to automatic review settings March 13, 2026 12:40

Copilot started reviewing on behalf of Steake March 13, 2026 12:41 View session

Copilot AI reviewed Mar 13, 2026

View reviewed changes

Copilot started work on behalf of Steake March 13, 2026 12:46 View session

Apply code review feedback: fix robustness issues in CATHEXIS crate

29980eb

Co-authored-by: Steake <530040+Steake@users.noreply.github.com>

Copilot finished work on behalf of Steake March 13, 2026 12:53

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Unify EQBSL and CATHEXIS Rust crates into a shared workspace#22

Unify EQBSL and CATHEXIS Rust crates into a shared workspace#22
Copilot wants to merge 4 commits intoimplement-rust-crate-version-of-eqbsl-library-with-docs-examples-and-tests-Wn9PR9JC77fhfrom
copilot/sub-pr-17-again

Copilot AI commented Mar 13, 2026 •

edited

Loading

Uh oh!

superconductor-for-github bot commented Mar 13, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Steake commented Mar 13, 2026

Uh oh!

Copilot AI commented Mar 13, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants


		// Convert probs to hard category
		let category_id = self.categoriser.predict(&features)?;

Conversation

Copilot AI commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

superconductor-for-github bot commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Steake commented Mar 13, 2026

Uh oh!

Copilot AI commented Mar 13, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Copilot AI commented Mar 13, 2026 •

edited

Loading

superconductor-for-github bot commented Mar 13, 2026 •

edited

Loading