feat: #760 polish — metadata, ranked alternatives, dead code cleanup by amiable-dev · Pull Request #791 · amiable-dev/conductor

amiable-dev · 2026-03-29T17:48:45Z

Summary

ADR-022 Phase 2A: get_discovered_ports Tauri command + discoveredPorts store #740: Add metadata field to DiscoveredPort (manufacturer, vendor_id, product_id for HID; display_name for all)
P18: suggest_binding now returns ranked alternatives array (up to 2 fallback categories with lower confidence)
ADR-022 Phase 2C: BindingPills rename + unbound badge + protocol icons (ADR-019 A4.1) #742: Remove dead DeviceStatusPills.svelte wrapper + test (was unused shim around BindingPills)

Test plan

cargo clippy --workspace -- -D warnings passes
cargo test --workspace passes
5 suggest_binding tests pass with alternatives in response
DiscoveredPort serialization includes metadata when non-empty

🤖 Generated with Claude Code

…d code 1. #740: Add metadata field to DiscoveredPort struct - MIDI ports: display_name (stripped instance suffix) - HID ports: display_name, manufacturer, vendor_id, product_id 2. P18: Return ranked alternatives in suggest_binding - Primary suggestion + up to 2 alternatives with lower confidence - build_alternatives() generates fallback categories 3. #742: Remove dead DeviceStatusPills wrapper + test - Was a thin shim around BindingPills, not imported anywhere Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

This PR polishes ADR-022 discovery and binding-suggestion UX by enriching discovered port data with metadata, returning ranked alternative binding suggestions for suggest_binding, and removing an unused legacy UI wrapper.

Changes:

Add a metadata field to DiscoveredPort (best-effort display_name for MIDI; USB info for HID).
Extend conductor_suggest_binding MCP tool responses to include an alternatives array (up to 2).
Remove dead DeviceStatusPills wrapper component and its test.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 4 comments.

File	Description
conductor-gui/ui/src/lib/components/DeviceStatusPills.test.ts	Deleted unused tests for the removed wrapper component.
conductor-gui/ui/src/lib/components/DeviceStatusPills.svelte	Deleted unused wrapper around `BindingPills`.
conductor-gui/src-tauri/src/commands.rs	Added `DiscoveredPort.metadata` and populated it for MIDI/HID discovered ports.
conductor-daemon/src/daemon/mcp_tools.rs	Added `alternatives` to `conductor_suggest_binding` responses via a helper builder.

Copilot · 2026-03-29T17:53:10Z

conductor-daemon/src/daemon/mcp_tools.rs

@@ -1698,6 +1730,7 @@ impl McpToolExecutor {
                        "reasoning": suggestion.reasoning,
                        "method": "event_fingerprint",
                        "event_count": stats.note_count + stats.cc_count + stats.gamepad_count,
+                        "alternatives": alternatives,
                        "note": "Classification based on observed event patterns"


conductor_suggest_binding now returns an alternatives array, but the existing suggest-binding tests in this file only assert category/method/confidence and never validate alternatives is present/structured (or that confidences are lower than the primary). Please extend the current tests to cover the new alternatives field for both event_fingerprint and port_name_heuristic paths.

Copilot · 2026-03-29T17:53:11Z

conductor-gui/src-tauri/src/commands.rs

+        let mut meta = std::collections::HashMap::new();
+        // Strip trailing instance suffix (" #2", " Port 1")
+        let base = name
+            .trim_end_matches(|c: char| c.is_ascii_digit() || c == '#' || c == ' ')
+            .trim_end_matches("Port")
+            .trim();


port_metadata trims all trailing ASCII digits unconditionally. This will incorrectly strip model numbers from real device names that legitimately end with digits (e.g., "nanoKONTROL2", "APC40"), producing a wrong display_name. Consider only stripping digits when they match known instance-suffix patterns (e.g. " #" or " Port ") rather than any trailing digit.

Suggested change

let mut meta = std::collections::HashMap::new();

// Strip trailing instance suffix (" #2", " Port 1")

let base = name

.trim_end_matches(|c: char| c.is_ascii_digit() || c == '#' || c == ' ')

.trim_end_matches("Port")

.trim();

fn strip_instance_suffix(s: &str) -> &str {

let s = s.trim_end();

// Pattern 1: trailing " #<digits>", e.g. "APC MINI #2"

if let Some(pos) = s.rfind(" #") {

let suffix = &s[pos + 2..];

if !suffix.is_empty() && suffix.chars().all(|c| c.is_ascii_digit()) {

return s[..pos].trim_end();

}

}

// Pattern 2: trailing " Port <digits>", e.g. "APC MINI Port 1"

if let Some(pos) = s.rfind(" Port ") {

let suffix = &s[pos + " Port ".len()..];

if !suffix.is_empty() && suffix.chars().all(|c| c.is_ascii_digit()) {

return s[..pos].trim_end();

}

}

// No known instance suffix found; return the original (trimmed) string.

s

}

let mut meta = std::collections::HashMap::new();

// Strip trailing instance suffix (" #2", " Port 1") but keep model numbers.

let base = strip_instance_suffix(name);

Copilot · 2026-03-29T17:53:11Z

conductor-daemon/src/daemon/mcp_tools.rs

+            (PadController, "pads", "midi"),
+            (Keyboard, "keys", "midi"),
+            (FaderController, "faders", "midi"),
+            (EncoderController, "encoders", "midi"),
+            (GameController, "gamepad", "hid"),
+        ];
+        all.iter()
+            .filter(|(cat, _, _)| cat != primary_category)
+            .take(2)
+            .map(|(cat, alias, protocol)| {
+                let cat_value = serde_json::to_value(cat)
+                    .unwrap_or(serde_json::Value::String("Unknown".to_string()));
+                json!({
+                    "category": cat_value,
+                    "suggested_alias": alias,
+                    "suggested_protocol": protocol,


build_alternatives re-encodes the category→(alias,protocol) mapping that already exists in conductor_core::device_intelligence::fingerprint::suggest_binding. This duplication can drift over time (e.g., if new categories are added or aliases change) and would make alternatives inconsistent with primary suggestions. Consider centralizing the mapping in conductor-core (e.g., a helper on DeviceCategory) and reusing it here.

Suggested change

(PadController, "pads", "midi"),

(Keyboard, "keys", "midi"),

(FaderController, "faders", "midi"),

(EncoderController, "encoders", "midi"),

(GameController, "gamepad", "hid"),

];

all.iter()

.filter(|(cat, _, _)| cat != primary_category)

.take(2)

.map(|(cat, alias, protocol)| {

let cat_value = serde_json::to_value(cat)

.unwrap_or(serde_json::Value::String("Unknown".to_string()));

json!({

"category": cat_value,

"suggested_alias": alias,

"suggested_protocol": protocol,

PadController,

Keyboard,

FaderController,

EncoderController,

GameController,

];

all.iter()

.filter(|cat| *cat != primary_category)

.take(2)

.map(|cat| {

let cat_value = serde_json::to_value(cat)

.unwrap_or(serde_json::Value::String("Unknown".to_string()));

json!({

"category": cat_value,

Copilot · 2026-03-29T17:53:11Z

conductor-daemon/src/daemon/mcp_tools.rs

+    /// Build ranked alternative suggestions (P18).
+    /// Returns up to 2 alternatives with lower confidence.
+    fn build_alternatives(
+        primary_category: &DeviceCategory,
+        primary_confidence: f64,
+    ) -> Vec<serde_json::Value> {


The tool definition for conductor_suggest_binding (in get_tool_definitions) still describes the response as only containing category/confidence/alias/protocol/reasoning. Since this change adds an alternatives field to the JSON output, please update the tool description so MCP/LLM consumers know it exists and how to use it.

… description - Fix port_metadata to only strip " #N" / " Port N" suffixes (preserve model numbers) - build_alternatives uses core's compute_suggestion (no duplicated mapping) - Update tool description to mention alternatives array - Add test assertions for alternatives (present, lower confidence) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

Copilot · 2026-03-29T18:05:07Z

conductor-daemon/src/daemon/mcp_tools.rs

+        category_seeds
+            .iter()
+            .filter(|(_, cat)| cat != primary_category)
+            .take(2)
+            .map(|(seed_fn, _)| {
+                let mut stats = EventStats::new();
+                seed_fn(&mut stats);
+                let suggestion = compute_suggestion(&stats, port_name);
+                let cat_value = serde_json::to_value(&suggestion.category)
+                    .unwrap_or(serde_json::Value::String("Unknown".to_string()));
+                json!({
+                    "category": cat_value,
+                    "suggested_alias": suggestion.suggested_alias,
+                    "suggested_protocol": suggestion.suggested_protocol,
+                    "confidence": (primary_confidence * 0.3).min(0.2),


In build_alternatives, the .take(2) happens before any ranking/selection, so the last seed (GameController) can never be returned as an alternative (it’s always after PadController/Keyboard/FaderController, and only one category is filtered out). If the intent is “ranked alternatives”, consider generating all non-primary candidates, sorting them by a score (e.g., their own suggestion.confidence scaled below the primary), then taking the top 2.

Suggested change

category_seeds

.iter()

.filter(|(_, cat)| cat != primary_category)

.take(2)

.map(|(seed_fn, _)| {

let mut stats = EventStats::new();

seed_fn(&mut stats);

let suggestion = compute_suggestion(&stats, port_name);

let cat_value = serde_json::to_value(&suggestion.category)

.unwrap_or(serde_json::Value::String("Unknown".to_string()));

json!({

"category": cat_value,

"suggested_alias": suggestion.suggested_alias,

"suggested_protocol": suggestion.suggested_protocol,

"confidence": (primary_confidence * 0.3).min(0.2),

// Build suggestions for all non-primary categories first.

let mut candidates = Vec::new();

for (seed_fn, cat) in category_seeds.iter() {

if cat == primary_category {

continue;

}

let mut stats = EventStats::new();

seed_fn(&mut stats);

let suggestion = compute_suggestion(&stats, port_name);

candidates.push(suggestion);

}

// Sort candidates by their own confidence (highest first), then take top 2.

candidates.sort_by(|a, b| {

b.confidence

.partial_cmp(&a.confidence)

.unwrap_or(std::cmp::Ordering::Equal)

});

candidates

.into_iter()

.take(2)

.map(|suggestion| {

let cat_value = serde_json::to_value(&suggestion.category)

.unwrap_or(serde_json::Value::String("Unknown".to_string()));

let alt_confidence = (suggestion.confidence * primary_confidence)

.min(primary_confidence * 0.9);

json!({

"category": cat_value,

"suggested_alias": suggestion.suggested_alias,

"suggested_protocol": suggestion.suggested_protocol,

"confidence": alt_confidence,

Copilot · 2026-03-29T18:05:07Z

conductor-daemon/src/daemon/mcp_tools.rs

+                json!({
+                    "category": cat_value,
+                    "suggested_alias": suggestion.suggested_alias,
+                    "suggested_protocol": suggestion.suggested_protocol,
+                    "confidence": (primary_confidence * 0.3).min(0.2),
+                })


Alternative confidence is currently derived only from primary_confidence and is identical for every alternative. This also produces alternatives with confidence == 0.0 when the primary heuristic classification yields 0.0 (Unknown/no events), violating the “lower confidence than primary” expectation. Consider either omitting alternatives when primary_confidence is 0.0, or computing per-alternative confidence from each candidate’s own stats and clamping it to be strictly less than the primary.

- Alternatives now have rank-based decay (1st: 30%, 2nd: 15% of primary) - Min confidence 0.01 to avoid zero-confidence alternatives - Test verifies decreasing confidence across alternatives Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

Copilot · 2026-03-29T18:19:07Z

conductor-daemon/src/daemon/mcp_tools.rs

+            // P18: ranked alternatives
+            let alts = parsed["alternatives"]
+                .as_array()
+                .expect("should have alternatives");
+            assert!(alts.len() >= 1, "should have at least 1 alternative");
+            let mut prev_conf = parsed["confidence"].as_f64().unwrap();
+            for alt in alts {
+                assert!(alt["category"].is_string());
+                assert!(alt["suggested_alias"].is_string());
+                let alt_conf = alt["confidence"].as_f64().unwrap();
+                assert!(alt_conf > 0.0, "alternative confidence should be positive");
+                assert!(
+                    alt_conf <= prev_conf,
+                    "alternatives should have decreasing confidence"
+                );
+                prev_conf = alt_conf;
+            }


The new alternatives assertions only cover cases where the primary confidence is > 0 (e.g., names that match the heuristic seeds). It would be good to add a regression test for an “unknown/insufficient data” port name (primary confidence 0.0) to ensure alternatives are either omitted or never exceed the primary confidence.

Copilot · 2026-03-29T18:19:07Z

conductor-gui/src-tauri/src/commands.rs

+    /// Optional metadata (manufacturer, model, etc.) extracted from port name or USB info
+    #[serde(default, skip_serializing_if = "std::collections::HashMap::is_empty")]


metadata is described as “Optional” and has skip_serializing_if = is_empty, but get_discovered_ports always inserts at least display_name, so the map is never empty and the field will always be serialized. Either adjust the doc comment to reflect that metadata is always present (at least display_name), or change construction/typing so it can actually be omitted when empty.

Suggested change

/// Optional metadata (manufacturer, model, etc.) extracted from port name or USB info

#[serde(default, skip_serializing_if = "std::collections::HashMap::is_empty")]

/// Metadata (always includes at least `display_name`, and may contain manufacturer, model, etc.)

#[serde(default)]

- Fix DiscoveredPort.metadata doc: always present (not optional) - Remove skip_serializing_if since metadata always has display_name - Add test: unknown port name produces positive-confidence alternatives Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

Copilot · 2026-03-29T18:43:08Z

conductor-gui/src-tauri/src/commands.rs

+    /// Port metadata (always includes display_name; HID ports add manufacturer, vendor_id, product_id)
+    #[serde(default)]
+    pub metadata: std::collections::HashMap<String, String>,


The PR description/test plan says DiscoveredPort serialization includes metadata when non-empty, but metadata currently lacks skip_serializing_if, so an empty map would still serialize as {}. If the intent is conditional inclusion, add #[serde(default, skip_serializing_if = "std::collections::HashMap::is_empty")] (or adjust the description if metadata is always expected to be present).

Copilot · 2026-03-29T18:43:08Z

conductor-gui/src-tauri/src/commands.rs

+    /// Extract metadata from a MIDI port name (best-effort heuristic).
+    fn port_metadata(name: &str) -> std::collections::HashMap<String, String> {
+        let mut meta = std::collections::HashMap::new();
+        // Strip known instance suffixes: " #2", " Port 3" (OS-appended)
+        // Preserve model numbers that are part of the name (e.g., "nanoKONTROL2")
+        let base = if let Some((prefix, suffix)) = name.rsplit_once(" #") {
+            if suffix.chars().all(|c| c.is_ascii_digit()) {
+                prefix
+            } else {
+                name
+            }
+        } else if let Some((prefix, suffix)) = name.rsplit_once(" Port ") {
+            if suffix.chars().all(|c| c.is_ascii_digit()) {
+                prefix
+            } else {
+                name
+            }
+        } else {
+            name
+        };
+        meta.insert("display_name".to_string(), base.to_string());
+        meta


port_metadata introduces non-trivial name normalization logic (stripping " #N" / " Port N") and is part of the new API payload (metadata.display_name), but there are no unit tests covering these cases. Since this file already has a test module, consider extracting the normalization into a testable helper and adding cases like "Foo #2" → "Foo" and "Foo Port 3" → "Foo", plus a serialization test asserting metadata.display_name is present.

- Move port_metadata out of nested function for testability - 5 tests: model number preservation (nanoKONTROL2, APC40), instance suffix stripping (#2, Port 3), plain name passthrough Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

Copilot · 2026-03-29T18:51:46Z

conductor-daemon/src/daemon/mcp_tools.rs

+                // Decreasing confidence: 1st alt ~30%, 2nd alt ~15% of primary (min 0.01)
+                let decay = if rank == 0 { 0.3 } else { 0.15 };
+                let alt_conf = (primary_confidence * decay).clamp(0.01, 0.2);
+                json!({
+                    "category": cat_value,
+                    "suggested_alias": suggestion.suggested_alias,
+                    "suggested_protocol": suggestion.suggested_protocol,
+                    "confidence": alt_conf,
+                })
+            })


build_alternatives clamps alternative confidence to a minimum of 0.01. When the primary suggestion has confidence 0.0 (e.g., unknown port name with no heuristic seed events), this produces alternatives with higher confidence than the primary, contradicting the tool description/comments that alternatives are lower-confidence fallbacks. Consider either (a) allowing 0.0 alternative confidence when primary is 0.0, or (b) special-casing the unknown/zero-confidence case so alternatives remain strictly <= primary (and updating tests accordingly).

Copilot · 2026-03-29T18:51:46Z

conductor-daemon/src/daemon/mcp_tools.rs

+        // Synthetic EventStats that produce each category via classify()
+        #[allow(clippy::type_complexity)]
+        let category_seeds: &[(fn(&mut EventStats), DeviceCategory)] = &[
+            (
+                |s| {
+                    for n in 36..52 {
+                        s.record_note(n, 100);
+                    }
+                },
+                DeviceCategory::PadController,
+            ),
+            (
+                |s| {
+                    for n in 21..109 {
+                        s.record_note(n, 80);
+                    }
+                },
+                DeviceCategory::Keyboard,
+            ),
+            (
+                |s| {
+                    for cc in 0..8 {
+                        s.record_cc(cc);
+                    }
+                },
+                DeviceCategory::FaderController,
+            ),
+            (
+                |s| {
+                    for _ in 0..20 {
+                        s.record_gamepad();
+                    }
+                },
+                DeviceCategory::GameController,
+            ),


build_alternatives seeds synthetic EventStats using large loops (e.g., 88 note events for keyboard) even though classification can be triggered with far fewer events (e.g., 2 notes far apart for keyboard; 1 CC for fader; 1 gamepad event). Since this runs on every conductor_suggest_binding call, consider reducing the seeded events to the minimum needed to hit each category’s classify() thresholds to cut unnecessary work.

- Empty alternatives for Unknown ports (avoids higher-than-primary confidence) - Use minimal synthetic events (2 per category) instead of large loops - Update test to verify empty alternatives for unknown ports Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

Copilot · 2026-03-29T18:59:28Z

conductor-daemon/src/daemon/mcp_tools.rs

+        let categories = [
+            (DeviceCategory::PadController, 36u8, 40u8, false),
+            (DeviceCategory::Keyboard, 21, 100, false),
+            (DeviceCategory::FaderController, 0, 1, true),
+            (DeviceCategory::GameController, 0, 0, false),
+        ];


build_alternatives never considers DeviceCategory::EncoderController, even though it exists in conductor_core::device_intelligence::fingerprint::DeviceCategory. This makes it impossible for the MCP response to ever return an encoder alternative (and the current synthetic CC stats can only classify as FaderController). Consider adding EncoderController to the candidate list and generating stats that satisfy core’s encoder heuristic (high hit count on <=4 CCs).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

Copilot · 2026-03-29T19:05:25Z

conductor-gui/src-tauri/src/commands.rs

+            let mut metadata = std::collections::HashMap::new();
+            metadata.insert("display_name".to_string(), name.to_string());


HID metadata construction duplicates the display_name logic and bypasses port_metadata(), so any instance-suffix stripping behavior applied to MIDI ports won’t be applied to HID ports. Consider starting with let mut metadata = port_metadata(name); here and then inserting manufacturer/vendor_id/product_id to keep display_name normalization consistent across protocols and avoid drift.

Suggested change

let mut metadata = std::collections::HashMap::new();

metadata.insert("display_name".to_string(), name.to_string());

let mut metadata = port_metadata(name);

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

Copilot · 2026-03-29T19:10:31Z

conductor-daemon/src/daemon/mcp_tools.rs

+    #[tokio::test]
+    async fn test_suggest_binding_unknown_port_omits_alternatives() {
+        let executor = McpToolExecutor::new();
+        // Unknown port name — primary confidence is 0, so alternatives should be empty
+        let result = executor
+            .execute(


Test name says the tool "omits" alternatives, but the assertions expect an alternatives array to be present (empty). Rename the test to reflect the actual contract (e.g., "returns empty alternatives"), or change the response to truly omit the field if that's the intended behavior.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot AI review requested due to automatic review settings March 29, 2026 17:48

Copilot started reviewing on behalf of amiable-dev March 29, 2026 17:49 View session

Copilot AI reviewed Mar 29, 2026

View reviewed changes

amiable-dev requested a review from Copilot March 29, 2026 18:01

Copilot started reviewing on behalf of amiable-dev March 29, 2026 18:01 View session

Copilot AI reviewed Mar 29, 2026

View reviewed changes

amiable-dev and others added 2 commits March 29, 2026 19:07

fix: use clamp() instead of max().min() (clippy manual_clamp)

5e18f41

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

amiable-dev requested a review from Copilot March 29, 2026 18:15

Copilot started reviewing on behalf of amiable-dev March 29, 2026 18:15 View session

Copilot AI reviewed Mar 29, 2026

View reviewed changes

amiable-dev requested a review from Copilot March 29, 2026 18:38

Copilot started reviewing on behalf of amiable-dev March 29, 2026 18:39 View session

fix: cargo fmt

b33c845

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot AI reviewed Mar 29, 2026

View reviewed changes

amiable-dev requested a review from Copilot March 29, 2026 18:47

Copilot started reviewing on behalf of amiable-dev March 29, 2026 18:48 View session

Copilot AI reviewed Mar 29, 2026

View reviewed changes

amiable-dev requested a review from Copilot March 29, 2026 18:54

Copilot started reviewing on behalf of amiable-dev March 29, 2026 18:55 View session

Copilot AI reviewed Mar 29, 2026

View reviewed changes

fix: include EncoderController in alternatives seeds

72685b1

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

amiable-dev requested a review from Copilot March 29, 2026 19:01

Copilot started reviewing on behalf of amiable-dev March 29, 2026 19:02 View session

Copilot AI reviewed Mar 29, 2026

View reviewed changes

fix: HID metadata reuses port_metadata() for consistent display_name

0d83917

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

amiable-dev requested a review from Copilot March 29, 2026 19:07

Copilot started reviewing on behalf of amiable-dev March 29, 2026 19:08 View session

Copilot AI reviewed Mar 29, 2026

View reviewed changes

fix: rename test to match behavior (empty alternatives, not omitted)

1273f4e

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

amiable-dev merged commit d293d85 into main Mar 29, 2026
11 checks passed

-        let mut meta = std::collections::HashMap::new();
-        // Strip trailing instance suffix (" #2", " Port 1")
-        let base = name
-            .trim_end_matches(|c: char| c.is_ascii_digit() || c == '#' || c == ' ')
-            .trim_end_matches("Port")
-            .trim();
+        fn strip_instance_suffix(s: &str) -> &str {
+            let s = s.trim_end();
+            // Pattern 1: trailing " #<digits>", e.g. "APC MINI #2"
+            if let Some(pos) = s.rfind(" #") {
+                let suffix = &s[pos + 2..];
+                if !suffix.is_empty() && suffix.chars().all(|c| c.is_ascii_digit()) {
+                    return s[..pos].trim_end();
+                }
+            }
+            // Pattern 2: trailing " Port <digits>", e.g. "APC MINI Port 1"
+            if let Some(pos) = s.rfind(" Port ") {
+                let suffix = &s[pos + " Port ".len()..];
+                if !suffix.is_empty() && suffix.chars().all(|c| c.is_ascii_digit()) {
+                    return s[..pos].trim_end();
+                }
+            }
+            // No known instance suffix found; return the original (trimmed) string.
+            s
+        }
+        let mut meta = std::collections::HashMap::new();
+        // Strip trailing instance suffix (" #2", " Port 1") but keep model numbers.
+        let base = strip_instance_suffix(name);

		/// Optional metadata (manufacturer, model, etc.) extracted from port name or USB info
		#[serde(default, skip_serializing_if = "std::collections::HashMap::is_empty")]

		let mut metadata = std::collections::HashMap::new();
		metadata.insert("display_name".to_string(), name.to_string());

	let mut metadata = std::collections::HashMap::new();
	metadata.insert("display_name".to_string(), name.to_string());
	let mut metadata = port_metadata(name);

Conversation

amiable-dev commented Mar 29, 2026

Summary

Test plan

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 29, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!