altium-toolkit 1.1.3 → 1.1.23

package/README.md

@@ -18,23 +18,34 @@ browser or Node-based tools.
 
 ## Features
 
-- Parse standalone native `.SchDoc`, `.PcbDoc`, `.PcbLib`, `.PrjPcb`, and
-  `.IntLib` files from `ArrayBuffer`
-- Recover schematic records, PCB outlines, placements, PCB library footprints,
+- Parse standalone native `.SchDoc`, `.PcbDoc`, `.SchLib`, `.PcbLib`,
+  `.PrjPcb`, and `.IntLib` files from `ArrayBuffer`
+- Recover schematic records, PCB outlines, placements, schematic library
+  symbols, PCB library footprints,
   project document references, variants, parameters, primitives, embedded
   schematic images, component annotations from PrimitiveParameters/Text streams,
   PCB pad/via stack and hole-tolerance detail, via-protection sidecars, custom
   pad shape links, extended mask/paste sidecars, PCB union metadata, embedded
   PCB 3D payload metadata, PCB component provenance, differential-pair class
   joins, schematic directive semantics, barcode PCB text metadata, mechanical
-  layer pairs, pick-and-place coordinate modes, PCB dimensions, project
+  layer pairs, pick-and-place coordinate modes, PCB dimensions, embedded-board
+  panel placements, placement rooms, project
   class-generation policy, project-level design bundles, annotation mappings,
   effective variant views, schematic/PCB ownership sidecars, deterministic
   wirelist/netlist exports, library render manifests, library lookup indexes,
+  library catalog artifacts, project hierarchy reports,
+  schematic-library section keys, pin side streams, compressed storage assets,
   schematic project-parameter text resolution, PCB QA statistics, structured
   diagnostics, and embedded PCB/PcbLib font payloads with basic text metrics
 - Preserve raw PCB primitive records through a read-only record registry so
-  unsupported or partially decoded stream data remains inspectable
+  unsupported or partially decoded stream data remains inspectable; native OLE
+  stream inventories summarize known, unknown, consumed, and opaque streams
+- Build deterministic parser field-coverage matrix, raw-data preservation,
+  parameter-record inventory, parser value-verification, normalized
+  diagnostics, geometry-bounds, fixture-coverage, embedded-asset,
+  library-diff, library-QA lint, project-hierarchy, and static
+  library-catalog reports, classify PCB layer ids, and convert common Altium
+  length units for downstream QA tooling
 - Emit Circuit JSON arrays from parser roots, with non-serialized
   renderer-compatibility fields for existing consumers
 - Render semantically annotated schematic SVG, semantically annotated PCB SVG,
@@ -54,6 +65,14 @@ The package is published on npm as
 npm install altium-toolkit
 ```
 
+GitHub Packages releases are published as `@sunbox/altium-toolkit`. Configure
+the GitHub Packages registry for the `@sunbox` scope before installing:
+
+```bash
+npm config set @sunbox:registry https://npm.pkg.github.com
+npm install @sunbox/altium-toolkit
+```
+
 ## Usage
 
 ```js
@@ -90,11 +109,21 @@ import 'altium-toolkit/styles/altium-renderers.css'
 - [Normalized Model Schema](docs/schemas/altium_toolkit/normalized_model_a1.schema.json)
 - [Project Bundle Schema](docs/schemas/altium_toolkit/project_bundle_a1.schema.json)
 - [Netlist Schema](docs/schemas/altium_toolkit/netlist_a1.schema.json)
+- [Parser Diagnostics Schema](docs/schemas/altium_toolkit/parser_diagnostics_a1.schema.json)
+- [Parser Value Verification Schema](docs/schemas/altium_toolkit/parser_value_verification_a1.schema.json)
+- [Geometry Bounds Schema](docs/schemas/altium_toolkit/geometry_bounds_a1.schema.json)
+- [Fixture Coverage Matrix Schema](docs/schemas/altium_toolkit/fixture_coverage_matrix_a1.schema.json)
+- [Unsupported Features Schema](docs/schemas/altium_toolkit/unsupported_features_a1.schema.json)
 - [Testing](docs/testing.md)
 - [Scope](spec/library-scope.md)
 
 ## Examples
 
+- Read-only utility scripts:
+  `examples/inspect-board.mjs`, `examples/extract-bom.mjs`,
+  `examples/generate-pnp.mjs`, `examples/net-report.mjs`,
+  `examples/library-catalog.mjs`, `examples/validate-library.mjs`, and
+  `examples/corpus-smoke.mjs`
 - [Arduino Uno Altium example](examples/arduino-uno/) based on Mehdi
   KHALFALLAH's public
   [My-Arduino-UNO-Design](https://github.com/Mehdi-KHALFALLAH/My-Arduino-UNO-Design)

package/docs/api.md

@@ -28,23 +28,26 @@ import { AltiumParser } from 'altium-toolkit/parser'
 const circuitJson = AltiumParser.parseArrayBuffer(fileName, arrayBuffer)
 ```
 
-`fileName` is used to infer schematic, PCB document, PCB footprint-library, PCB
-project, or integrated-library parsing from the extension. The parser accepts
-native `.SchDoc`, `.PcbDoc`, `.PcbLib`, `.PrjPcb`, and `.IntLib` bytes as an
-`ArrayBuffer` and returns a Circuit JSON element array. The returned array
-carries non-serialized renderer-compatibility fields such as `kind`, `fileType`,
-`schematic`, `pcb`, `pcbLibrary`, `project`, `integratedLibrary`, `summary`,
+`fileName` is used to infer schematic, PCB document, schematic symbol-library,
+PCB footprint-library, PCB project, or integrated-library parsing from the
+extension. The parser accepts native `.SchDoc`, `.PcbDoc`, `.SchLib`,
+`.PcbLib`, `.PrjPcb`, and `.IntLib` bytes as an `ArrayBuffer` and returns a
+Circuit JSON element array. The returned array carries non-serialized
+renderer-compatibility fields such as `kind`, `fileType`, `schematic`, `pcb`,
+`schematicLibrary`, `pcbLibrary`, `project`, `integratedLibrary`, `summary`,
 `diagnostics`, and `bom` so existing renderers can consume parser output
 directly during the migration.
 
 PCB parsing reads the main primitive streams together with sidecar streams such
 as `PrimitiveParameters/Data`, `WideStrings6/Data`,
 `ExtendedPrimitiveInformation/Data`, `CustomShapes/Data`, `UnionNames/Data`,
-and `SmartUnions/Data`. Component parameters are joined by native primitive
+`SmartUnions/Data`, `EmbeddedBoards6/Data`, and `Rooms6/Data`. Component parameters are joined by native primitive
 unique id, modern `Texts6` designator records may resolve their display string
 through the wide-string table, custom pad geometry is linked to anchor pads, and
 smart-union memberships are attached to referenced primitives before the
-normalized component list and BOM are built.
+normalized component list and BOM are built. Embedded-board and room streams are
+promoted to read-only `pcb.embeddedBoards` and `pcb.rooms` collections when
+present.
 
 ```js
 import { CircuitJsonModelSchema } from 'altium-toolkit/parser'
@@ -58,21 +61,38 @@ Use `AltiumParser.parseArrayBufferToRendererModel(fileName, arrayBuffer)` when
 an integration still needs the legacy renderer model object. The
 `CircuitJsonModelAdapter` export also exposes `fromRendererModel()`,
 `toRendererModel()`, and `isCircuitJson()` for explicit conversions.
+Use `AltiumParser.tryParseArrayBuffer()` or
+`AltiumParser.tryParseArrayBufferToRendererModel()` when a batch integration
+needs a non-throwing `{ ok, model, diagnostics }` envelope.
 
 Specialized parser helpers are exported for lower-level integrations, including
-`IntLibStreamExtractor`, `PcbBoardRegionSemanticsParser`,
-`PcbComponentPrimitiveIndexer`, `PcbCustomPadShapeParser`,
-`PcbDimensionParser`, `PcbEmbeddedFontExtractor`,
+`AltiumUnits`, `IntLibStreamExtractor`, `FixtureCoverageMatrixBuilder`,
+`GeometryBoundsReportBuilder`,
+`ParameterCollection`, `ParameterRecordInventoryBuilder`,
+`ParserDiagnosticNormalizer`,
+`ParserFieldCoverageReportBuilder`, `ParserValueVerificationReportBuilder`,
+`ParserCompatibilityFuzzer`, `NativeStreamInventoryBuilder`,
+`RawDataPreservationReportBuilder`, `UnsupportedFeatureReportBuilder`,
+`EmbeddedAssetReportBuilder`, `LibraryDiffReportBuilder`,
+`LibraryInspectionReportBuilder`,
+`PcbBoardRegionSemanticsParser`, `PcbComponentPrimitiveIndexer`,
+`PcbCustomPadShapeParser`, `PcbDimensionParser`, `PcbEmbeddedFontExtractor`,
 `PcbExtendedPrimitiveInformationParser`, `PcbFontMetricsParser`,
-`LibraryRenderManifestBuilder`, `LibrarySearchIndex`,
-`PcbBomProfileBuilder`, `PcbLayerStackFidelityReportBuilder`,
-`PcbOwnershipGraphBuilder`, `PcbPadStackParser`, `PcbPickPlacePositionResolver`,
-`ProjectAnnotationParser`, `ProjectDesignBundleBuilder`,
+`LibraryRenderManifestBuilder`, `LibraryCatalogArtifactBuilder`,
+`LibrarySearchIndex`, `SchLibModelParser`, `SchLibStreamExtractor`,
+`PcbBomProfileBuilder`, `PcbClassReportBuilder`,
+`PcbInspectionReportBuilder`, `PcbLayerStackFidelityReportBuilder`,
+`PcbLayerGroups`, `PcbNetMembershipReportBuilder`,
+`PcbOwnershipGraphBuilder`, `PcbPadStackParser`,
+`PcbPickPlacePositionResolver`, `ProjectAnnotationParser`,
+`ProjectDesignBundleBuilder`, `ProjectHierarchyReportBuilder`,
 `ProjectNetlistExporter`, `ProjectVariantViewBuilder`,
 `PcbMechanicalLayerPairParser`, `PcbSpecialStringResolver`, `PcbUnionParser`,
 `PcbViaStackParser`, `PcbRuleParser`, `PcbRawRecordRegistry`,
-`PcbStatisticsBuilder`, `SchematicOwnershipGraphParser`, and
-`SchematicProjectParameterResolver`.
+`PcbStatisticsBuilder`, `SchematicCodeSymbolParser`,
+`SchematicOwnershipGraphParser`,
+`SchematicProjectParameterResolver`, `SchematicRecordStreamParser`, and
+`SchematicTextRunParser`.
 `PcbBoardRegionSemanticsParser` exposes the substack and bending-line
 normalization used by `.PcbDoc` models. `PcbComponentPrimitiveIndexer` exposes
 the native component-index grouping used to populate
@@ -83,6 +103,9 @@ parser-only Dimensions6 normalization used by `.PcbDoc` parsing.
 `SchematicOwnershipGraphParser` and `PcbOwnershipGraphBuilder` expose the
 read-only ownership sidecars that parser roots attach under
 `schematic.ownership` and `pcb.ownership`.
+`SchematicCodeSymbolParser` exposes auxiliary schematic code-symbol, entry,
+text, and marker records through the parser root's `schematic.codeSymbols`
+sidecar.
 `PcbPickPlacePositionResolver` exposes the component-origin and pad-anchor
 coordinate modes used by the normalized `pnp` model.
 `ProjectDesignBundleBuilder` composes separately parsed project, schematic, and
@@ -93,7 +116,11 @@ and net views. `ProjectAnnotationParser` parses read-only annotation mapping
 files, and `ProjectNetlistExporter` emits deterministic wirelist and richer
 JSON netlist contracts from normalized bundles. The JSON contract includes
 schematic source sheets, graphical elements, aliases, terminal endpoints,
-hierarchy paths, and PCB net-table provenance when present.
+hierarchy paths, and PCB net-table provenance when present. Alias arrays include
+explicit schematic labels and additive unnamed-net candidates when available.
+`ProjectHierarchyReportBuilder` emits a deterministic sheet hierarchy report
+from parsed project and schematic models, including roots, child sheet links,
+missing-sheet diagnostics, cycles, repeated references, and sheet-entry names.
 `PcbMechanicalLayerPairParser`
 exposes the mechanical-layer flip map used by `.PcbDoc` parsing. The font
 helpers expose the same
@@ -105,11 +132,107 @@ extractors.
 normalization. `PcbLayerStackFidelityReportBuilder` classifies layer-stack
 source evidence and unsupported native-regeneration limits for deterministic
 QA/reporting.
-`LibraryRenderManifestBuilder` and `LibrarySearchIndex` expose deterministic
-SchLib/PcbLib render/export manifests plus exact, keyword, and fuzzy lookup
-helpers. `PcbStatisticsBuilder` emits board QA summaries used by `.PcbDoc`
-models. `SchematicProjectParameterResolver` resolves dot-prefixed and
+`PcbLayerGroups` classifies legacy PCB layer ids into stable groups such as
+copper, overlay, paste, solder mask, mechanical, drill, and multi-layer for
+filtering and reporting. It also exposes deterministic colors and draw
+priorities for report and renderer-side layer ordering. `AltiumUnits` converts
+common Altium lengths between mil, millimeter, inch, and raw fixed-point
+coordinate units.
+`ParserFieldCoverageReportBuilder` builds a deterministic report of observed,
+mapped, missing, and unsupported native fields from explicit source records or
+parser roots that carry source-record sidecars. Reports include a compact matrix
+view with per-family coverage status and mapped-field ratios. Field matching is
+case-insensitive, while report rows preserve observed source spelling.
+`ParameterRecordInventoryBuilder` scans raw pipe/backtick parameter records into
+delimiter-aware field rows, duplicate-key rollups, UTF-8 marker counts, and
+typed scalar hints without changing parser semantics.
+`ParameterCollection` exposes duplicate-preserving, case-insensitive typed
+reads over parsed parameter fields for integrations that need local lookup
+helpers without writer behavior.
+`ParserValueVerificationReportBuilder` compares curated expected path/value
+assertions against parser output, producing fixture-gate reports that classify
+missing paths separately from mismatched values.
+`ParserDiagnosticNormalizer` converts string, `Error`, and object diagnostics
+into a shared envelope with stable `code`, `severity`, `message`, source stream,
+record index, typed error kind, field, and context metadata. Typed parser error
+classes include `AltiumParseError`, `AltiumCorruptFileError`, and
+`AltiumUnsupportedFeatureError`.
+`GeometryBoundsReportBuilder` emits `altium-toolkit.geometry-bounds.a1` reports
+with deterministic axis-aligned bounds for common parsed schematic and PCB
+primitive families. `FixtureCoverageMatrixBuilder` emits
+`altium-toolkit.fixture-coverage-matrix.a1` reports from synthetic fixture
+manifests, including required coverage/contract gaps and native-asset policy
+status.
+`RawDataPreservationReportBuilder` summarizes preserved raw primitive records
+unknown records, and opaque schematic/library records without copying raw
+payload bytes into the report. When parser roots carry native stream
+inventories, it also reports known/unknown stream counts, unconsumed stream
+counts, and stream byte totals separately from raw-record payload bytes.
+`UnsupportedFeatureReportBuilder` emits
+`altium-toolkit.unsupported-features.a1` summaries of unsupported record
+families, unsupported or unparsed raw records, opaque preserved schematic rows,
+and unsupported diagnostics across parser roots.
+`NativeStreamInventoryBuilder` emits metadata-only OLE stream rows with byte
+length, checksum, known/unknown classification, and consumed status.
+`SchematicRecordStreamParser` exposes framed schematic stream parsing with
+opaque-frame preservation for lower-level extractors.
+`ParserCompatibilityFuzzer` runs deterministic malformed,
+wrong-reader, and sparse-input cases against parser entrypoints for parser QA.
+`EmbeddedAssetReportBuilder` emits unified embedded-asset inventories across
+parser roots, and `LibraryDiffReportBuilder` compares parsed symbol and
+footprint libraries by name, counts, and parameters.
+`LibraryInspectionReportBuilder` composes library inventory and QA findings
+into one stable artifact. `SchLibStreamExtractor` and `SchLibModelParser`
+expose native schematic-symbol library recovery, including section keys,
+file-header font metadata, pin side streams, compressed storage assets, and
+implementation child rows where available.
+`LibraryRenderManifestBuilder`, `LibraryCatalogArtifactBuilder`, and
+`LibrarySearchIndex` expose deterministic SchLib/PcbLib render/export
+manifests, static catalog artifacts, search metadata, plus exact, keyword, and
+fuzzy lookup helpers. `PcbStatisticsBuilder`, `PcbNetMembershipReportBuilder`,
+`PcbClassReportBuilder`, and `PcbInspectionReportBuilder` emit board QA,
+net-ownership, class-membership, and combined inspection artifacts for
+`.PcbDoc` models. `SchematicProjectParameterResolver` resolves dot-prefixed and
 equals-prefixed schematic special strings for parser and SVG integrations.
+`SchematicTextRunParser` parses schematic backslash suffix markers into display
+text plus overline run metadata reused by pin and text rendering.
+
+## Library Exporters
+
+```js
+import {
+    SourceComponentClient,
+    SourceComponentBundleNormalizer,
+    SourceBundleExporter,
+    AltiumSchLibExporter,
+    AltiumPcbLibExporter,
+    AltiumLibraryBatchExporter
+} from 'altium-toolkit/parser'
+```
+
+The exporter surface is local-first and host-controlled:
+
+- `SourceComponentClient` performs component search, component fetch, model
+  asset fetch, retry, and response validation through an injected `fetcher`.
+  It does not use global `fetch` implicitly.
+- `SourceComponentBundleNormalizer.normalize(raw)` converts provider-specific
+  component responses into a deterministic bundle with `symbol`, `footprint`,
+  `models`, `metadata`, `sourceJson`, and diagnostics fields.
+- `SourceBundleExporter.export(bundle)` emits deterministic raw source bundle
+  entries: `manifest.json`, `source/source.json`, and optional `models/*`
+  assets.
+- `AltiumSchLibExporter.export(bundles)` and
+  `AltiumPcbLibExporter.export(bundles)` write compact OLE-backed `.SchLib`
+  and `.PcbLib` byte arrays. The `.PcbLib` writer includes generated library
+  streams plus STEP/WRL model payload streams when the normalized bundle
+  contains model assets.
+- `AltiumLibraryBatchExporter` orchestrates id lists, search-and-export,
+  per-component source/SchLib/PcbLib outputs, merged library outputs,
+  append/skip manifests, progress events, continue-on-error diagnostics, and
+  checkpoint state.
+
+Hosts are responsible for choosing and configuring any outbound component
+source. Tests use repo-owned fake responses only.
 
 ## Netlist Query
 
@@ -148,7 +271,8 @@ import {
     PcbSvgRenderer,
     PcbSideResolvedRenderModel,
     preparePcbSideResolvedRenderModel,
-    BomTableRenderer
+    BomTableRenderer,
+    PcbLayerGroups
 } from 'altium-toolkit/renderers'
 ```
 
@@ -165,6 +289,9 @@ import {
   `side: 'back'` to project bottom components, documentation layers, copper
   primitives, vias, and pad stack geometry into the top-facing render surface.
 - `BomTableRenderer.render(rows)` returns grouped BOM table markup.
+- `PcbLayerGroups` is also exported from the renderer entrypoint for
+  layer-filter and layer-visibility code that should not depend on parser
+  internals.
 
 Renderer output is deterministic string markup. The library does not attach DOM
 events or mutate a host document.
@@ -214,3 +341,24 @@ import {
 
 The library intentionally does not create Three.js objects, canvases, controls,
 or event listeners.
+
+## Read-only CLI Examples
+
+The `examples/` directory includes small Node.js scripts for common
+non-interactive workflows:
+
+```bash
+node examples/inspect-board.mjs board.PcbDoc --json
+node examples/inspect-schematic.mjs design.SchDoc --view all --json
+node examples/extract-bom.mjs design.SchDoc
+node examples/generate-pnp.mjs board.PcbDoc
+node examples/net-report.mjs design.SchDoc --json
+node examples/library-catalog.mjs footprints.PcbLib
+node examples/validate-library.mjs footprints.PcbLib --json
+node examples/corpus-smoke.mjs local-corpus --coverage --json
+```
+
+Each script reads one input file and writes text, CSV, or JSON to stdout. The
+corpus smoke script reads a caller-provided directory and can include aggregate
+parser coverage and field-gap counters. The scripts are examples of library
+usage, not installed package binaries.

package/docs/model-format.md

@@ -34,6 +34,24 @@ Circuit JSON array. `JSON.stringify(result)` serializes only the Circuit JSON
 elements, including custom `altium_toolkit_*` sidecar elements; compatibility
 fields are intentionally omitted from serialized JSON.
 
+## Source Export Bundle
+
+`SourceComponentBundleNormalizer` produces the exporter input contract used by
+the source bundle, `.SchLib`, and `.PcbLib` writers:
+
+- `id` and `name`: stable component identity
+- `metadata`: provider metadata copied into deterministic plain-object form
+- `symbol`: schematic symbol name, pins, primitives, and raw source object
+- `footprint`: PCB footprint name, primitive families, and raw source object
+- `models`: model id, file name, format, bytes/text, and optional source URL
+- `sourceJson`: the original raw response retained for reproducible exports
+- `diagnostics`: warnings for incomplete source data
+
+`SourceBundleExporter.export()` serializes the original source response and a
+manifest that lists included model assets. It does not fetch network resources;
+callers provide already-normalized model bytes or use `SourceComponentClient`
+before exporting.
+
 ## Renderer Compatibility Fields
 
 For compatibility, `AltiumParser.parseArrayBuffer()` attaches the previous
@@ -46,13 +64,14 @@ the object form can call
 
 - `schema`: normalized model schema id, currently
   `urn:altium-toolkit:normalized-model:a1`
-- `kind`: `schematic`, `pcb`, `pcb-library`, `project`,
+- `kind`: `schematic`, `pcb`, `schematic-library`, `pcb-library`, `project`,
   `integrated-library`, or `design-bundle`
-- `fileType`: `SchDoc`, `PcbDoc`, `PcbLib`, `PrjPcb`, `IntLib`, or
+- `fileType`: `SchDoc`, `PcbDoc`, `SchLib`, `PcbLib`, `PrjPcb`, `IntLib`, or
   `ProjectDesignBundle`
 - `fileName`: original file name passed to the parser
 - `diagnostics`: parser warnings and recovery notes. Each diagnostic carries a
-  machine-readable `code` plus `severity` and `message`.
+  machine-readable `code`, `severity`, and `message`; entries may also carry
+  source stream, storage, record-index, field, and context metadata.
 - `bom`: grouped component metadata where available
 
 ## Schema Contracts
@@ -65,6 +84,102 @@ The serialized parser return value follows the upstream
 [`tscircuit/circuit-json`](https://github.com/tscircuit/circuit-json) element
 array convention.
 
+## Parser Utility Reports
+
+`ParserFieldCoverageReportBuilder.build()` emits
+`altium-toolkit.parser-field-coverage.a1` reports. Rows are grouped by parser
+domain and primitive family, with observed native fields split into mapped,
+missing, and unsupported sets. The report also carries a compact matrix with
+per-family coverage status and mapped-field ratios. Source field matching is
+case-insensitive while observed field names preserve their original spelling.
+It is intended for fixture coverage and parser QA; it does not mutate parser
+output or require native source files in tests.
+
+`ParameterRecordInventoryBuilder.build()` emits
+`altium-toolkit.parameter-record-inventory.a1` reports. Rows expose raw keys,
+normalized keys, values, delimiter type, backtick nesting level, duplicate
+occurrence numbers, UTF-8 key markers, and simple boolean/integer/number hints.
+Duplicate-key rollups include first and last values so consumers can audit
+parser policies without changing parsed model output.
+`ParameterCollection.parse()` provides the corresponding local read helper for
+duplicate-preserving, case-insensitive string, integer, number, boolean, code,
+and coordinate-style parameter access.
+
+`ParserValueVerificationReportBuilder.build()` emits
+`altium-toolkit.parser-value-verification.a1` reports for curated fixture
+manifests. Each assertion targets one explicit model path and expected value,
+then reports pass, missing-path, or mismatch status. This complements field
+coverage by checking recovered values rather than only field presence.
+
+`ParserDiagnosticNormalizer.buildReport()` emits
+`altium-toolkit.parser-diagnostics.a1` reports from string, `Error`, or object
+diagnostics. The normalized envelope uses stable codes, `info`/`warning`/`error`
+severities, messages, and optional source-stream, storage, record-index, field,
+typed error-kind, field, and context keys.
+
+`GeometryBoundsReportBuilder.build()` emits
+`altium-toolkit.geometry-bounds.a1` reports from parsed schematic and PCB
+document models. Rows identify document, domain, primitive family, primitive
+index, status, and rounded axis-aligned bounds; the summary carries the union
+bounds and missing-bounds count for QA tooling.
+
+`FixtureCoverageMatrixBuilder.build()` emits
+`altium-toolkit.fixture-coverage-matrix.a1` reports from synthetic fixture
+manifests. Coverage rows and contract rows list fixture keys, required flags,
+and missing gaps, while the policy section records the manifest asset policy
+and native-asset count.
+
+`RawDataPreservationReportBuilder.build()` emits
+`altium-toolkit.raw-data-preservation.a1` reports. It summarizes preserved raw
+primitive records, unknown records, and opaque schematic/library records by
+count, parse state, support state, and byte length without repeating the base64
+payloads already carried by parser models. When parser roots include
+`nativeStreams`, the report adds native-stream counts and byte totals without
+folding those stream bytes into the raw-record payload total.
+
+`UnsupportedFeatureReportBuilder.build()` emits
+`altium-toolkit.unsupported-features.a1` reports. It collects unsupported
+schematic record type summaries, unsupported or unparsed raw records, opaque
+preserved records, and unsupported diagnostics so parser coverage gaps can be
+reviewed separately from general raw-data preservation.
+
+`NativeStreamInventoryBuilder.buildFromStreams()` emits
+`altium-toolkit.native-stream-inventory.a1` reports for OLE stream maps. Parser
+roots expose the same metadata under `schematic.nativeStreams`,
+`pcb.nativeStreams`, `schematicLibrary.nativeStreams`, and
+`pcbLibrary.nativeStreams` when the source file was parsed from a compound
+document. Rows include stream path, storage, leaf name, byte length, checksum,
+known/unknown classification, and consumed status.
+
+`EmbeddedAssetReportBuilder.build()` emits
+`altium-toolkit.embedded-assets.a1` reports that normalize schematic images,
+embedded file inventories, PCB font/model payloads, and integrated-library
+source entries into one stable asset table. `LibraryDiffReportBuilder.build()`
+emits `altium-toolkit.library.diff.a1` reports comparing parsed symbol and
+footprint libraries by item name, counts, and parameter values.
+`LibraryInspectionReportBuilder.build()` emits
+`altium-toolkit.library.inspection.a1` reports that combine library inventory
+rows with duplicate, stale-link, missing-model, lint, and merge-plan QA
+summaries.
+
+`PcbNetMembershipReportBuilder.build()` emits
+`altium-toolkit.pcb.net-membership.a1` reports that count observed copper
+ownership by net across pads, tracks, arcs, vias, fills, regions, and polygons.
+Declared-but-empty nets, observed-but-undeclared nets, unowned primitives, and
+possible unrouted pad-only nets are split into deterministic lists.
+`PcbClassReportBuilder.build()` emits
+`altium-toolkit.pcb.class-report.a1` reports that summarize PCB classes by
+kind, enabled state, member resolution, empty classes, and unresolved members.
+`PcbInspectionReportBuilder.build()` emits `altium-toolkit.pcb.inspection.a1`
+reports that compose board statistics, primitive counts, design-rule counts,
+diagnostics, net membership, class membership, and route-analysis summaries
+into one inspection artifact.
+
+`PcbLayerGroups` provides stable layer-group names, deterministic display
+colors, and draw priorities for legacy PCB layer ids. `AltiumUnits` provides
+deterministic mil, millimeter, inch, and raw coordinate conversions for report
+output.
+
 ## Schematic Fields
 
 Schematic documents include recovered `schematic` data with sheet metadata,
@@ -79,9 +194,18 @@ pie-chart primitives are exposed as first-class `schematic.beziers` and
 `schematic.pies` arrays for deterministic SVG rendering. Rounded rectangles are
 exposed through `schematic.roundedRectangles`, and IEEE drawing symbols are
 exposed through `schematic.ieeeSymbols` with a stable `symbolName`. The registry
-also names schematic families such as notes, compile masks, harness records,
-blankets, and hyperlinks so consumers can inspect supported parser coverage
-without stringly typed local maps.
+also names schematic families such as notes, compile masks, harness records, and
+blankets so consumers can inspect supported parser coverage without stringly
+typed local maps. Hyperlink records are exposed as first-class
+`schematic.hyperlinks` entries with URL, display text, location, font/color, and
+owner/source metadata.
+
+`schematic.qa.fieldCoverage` provides a parser-development coverage sidecar for
+additive native fields that are not in the toolkit's current schematic field
+catalog. It groups unrecognized field names by record type and stable record
+keys without emitting top-level diagnostics. Known schematic fields are matched
+case-insensitively, so uppercase native keys do not inflate the unrecognized
+field counts.
 
 Record-28 text frames are preserved both as drawable note text and as a
 read-only `schematic.textFrames` contract with frame rectangle, alignment,
@@ -97,9 +221,14 @@ distinguish hidden directive metadata from visible sheet text.
 
 `schematic.ownership` is a read-only sidecar built from raw record
 `OwnerIndex` and `IndexInSheet` values. It exposes stable record keys,
-`childrenByParentKey`, `parentsByChildKey`, and `recordsByIndexInSheet` so
-consumers can inspect component, sheet-symbol, and directive children without
-reimplementing owner-index lookup rules.
+`childrenByParentKey`, `parentsByChildKey`, `recordsByIndexInSheet`, and a
+nested `hierarchy` projection so consumers can inspect component, sheet-symbol,
+and directive children without reimplementing owner-index lookup rules.
+
+`schematic.codeSymbols` is an optional read-only sidecar for auxiliary
+code-symbol style records. It preserves block geometry, entry pins, title/source
+text rows, routine and memory metadata, and marker points without affecting
+schematic SVG rendering or netlist merging.
 
 Schematic project parameters and special strings can be resolved without
 mutating source parser models through `SchematicProjectParameterResolver`.
@@ -115,10 +244,17 @@ family, and matching top-level warning diagnostic.
 
 The normalized schematic net model is single-sheet. `schematic.nets` is built
 from the wires, labels, ports, pins, junctions, bus entries, and sheet entries
-present in the parsed `.SchDoc`. Project-level hierarchy, repeated channels,
-variants, and cross-sheet compilation metadata are preserved through the
-`.PrjPcb` parser, but this schema does not currently emit a compiled
+present in the parsed `.SchDoc`. Nets with explicit labels keep those labels as
+their canonical `name`; unnamed nets keep their stable `UnknownNetN` name and
+may add `autoName`, `autoNameSource`, and `aliasCandidates` from connected
+component pins for display/search use. Project-level hierarchy, repeated
+channels, variants, and cross-sheet compilation metadata are preserved through
+the `.PrjPcb` parser, but this schema does not currently emit a compiled
 multi-sheet design netlist.
+`schematic.connectivityQa` reports read-only connectivity findings, including
+implicit generated net names, dangling labels, orphan ports, unconnected pins,
+ambiguous junctions, and un-junctioned tee contacts where one wire endpoint
+touches another wire interior without an authored junction.
 
 Embedded schematic images preserve the raw record geometry and expose
 browser-facing payload metadata. When an embedded stream contains a native
@@ -126,6 +262,14 @@ PNG/JPEG/GIF/SVG/WebP payload alongside a preview, `mimeType` and `dataBase64`
 refer to the native payload while `sourceMimeType` records the preview format.
 Alpha-bearing 32-bit BMP previews are converted to PNG and marked with
 `hasAlpha` so SVG renderers can display transparency deterministically.
+When a schematic OLE container exposes preview metadata, `schematic.thumbnails`
+contains PNG thumbnail sidecars with `kind`, dimensions, `sourceStream`,
+`pixelFormat`, `mimeType`, and `dataBase64` fields.
+
+When a native framed schematic stream contains non-property frames,
+`schematic.opaqueRecords` preserves compact metadata for those unmodelled
+payloads. Each row carries source stream context, frame type, record index, byte
+length, and base64 payload data for downstream read-only diagnostics.
 
 ## PCB Fields
 
@@ -200,6 +344,11 @@ dielectric thickness, dielectric constant, and dissipation factor, plus
 aggregate material and role counts. The `planning` section summarizes keepout
 regions, room-related rules and names, board-region flex/rigid counts, locked
 3D regions, bending-line counts, and board-region layer-stack usage.
+Native panel and placement streams are exposed as `pcb.embeddedBoards` and
+`pcb.rooms` when present. Embedded-board rows preserve source document path,
+placement layer, rotation, mirroring, array counts and spacing, unique id, and
+selected policy flags. Room rows preserve name, unique id, members, and simple
+bounds when available.
 
 Decoded pad primitives preserve raw `padFlags` plus named tenting and testpoint
 flags. Pad shape codes are kept as raw `shapeTop` / `shapeMid` / `shapeBottom`
@@ -283,6 +432,19 @@ callers that do not need the full geometry.
 
 ## PCB Library Fields
 
+Schematic symbol libraries include recovered `schematicLibrary` data with a
+library header, file-header metadata, section keys, stream names, ordered
+`symbols`, lookup `indexes`, embedded file inventory, opaque native frames,
+render manifest, and library QA report. Each symbol exposes source storage,
+pins, parts, parameters, implementation rows, primitive summaries, and
+referenced embedded assets where available. Pin rows preserve side-stream
+metadata such as fractional location/length, descriptions, package length,
+symbol line width, text placement, and pin-function hints when those streams are
+present.
+`schematicLibrary.indexes.symbolsByName` provides read-only symbol lookup and
+search metadata, while `schematicLibrary.renderManifest` lists deterministic
+symbol render outputs and embedded asset descriptors.
+
 PCB footprint libraries include recovered `pcbLibrary` data with library header
 properties, optional SectionKeys mappings, ComponentParamsTOC entries, and an
 ordered `footprints` list. Each footprint exposes its source storage name,
@@ -311,8 +473,13 @@ preserved versus stripped parameter names and stripped implementation keys.
 `LibraryRenderManifestBuilder.buildSchematicTemplateExtractionManifest()`
 summarizes template identity, owned records, fonts, title-block fields, and
 missing template parameters without generating template files.
-`LibrarySearchIndex` provides exact, keyword, and fuzzy symbol/footprint lookup
-helpers over parsed library read models.
+`LibraryCatalogArtifactBuilder.build()` emits
+`altium-toolkit.library.catalog.a1` artifacts for parsed schematic and PCB
+libraries. The artifact includes deterministic entries, QA issue badges,
+search-index rows, preview SVG keys from render manifests, and a static HTML
+catalog string with no server or runtime interaction. `LibrarySearchIndex`
+provides exact, keyword, and fuzzy symbol/footprint lookup helpers over parsed
+library read models.
 
 ## Project Fields
 
@@ -336,9 +503,13 @@ should combine project metadata with separately parsed schematic documents.
 `ProjectDesignBundleBuilder.build({ projectModel, documentModels,
 annotationModels })` composes already parsed project, schematic, PCB, and
 annotation models into a `design-bundle` payload. The bundle exposes `project`,
-`variants`, `sheets`, `components`, `schematic_hierarchy`, `pnp`, `nets`,
-`annotations`, and `indexes` so multi-document consumers can use one normalized
-JSON object above single-document parser output. Passing `variantName` adds
+`variants`, `sheets`, `components`, `schematic_hierarchy`,
+`schematicHierarchyReport`, `pnp`, `nets`, `annotations`, and `indexes` so
+multi-document consumers can use one normalized JSON object above
+single-document parser output. `ProjectHierarchyReportBuilder.build()` emits
+the richer hierarchy report directly from parsed project and schematic models,
+including child sheet links, missing sheets, cycles, repeated references,
+roots, and sheet-entry names. Passing `variantName` adds
 `effectiveVariant`, which applies DNP rows, alternate fitted rows, parameter
 overrides, and annotation designator mappings to BOM, PnP, component, and net
 views without mutating the source parser models. `ProjectNetlistExporter` emits
@@ -347,7 +518,8 @@ effective variant view. The wirelist remains a compact line-oriented
 `component.pin` view. The JSON netlist also carries aliases, auto-named flags,
 schematic source sheets, graphical source elements, terminal endpoints,
 hierarchy paths, and PCB net-table provenance when the bundle includes those
-details.
+details. Alias lists include explicit schematic labels and additive unnamed-net
+alias candidates when present.
 
 ## SVG And 3D Contracts
 
@@ -392,5 +564,5 @@ Parser fixes may add detail, but existing field names and shapes should stay
 compatible unless a new schema id explicitly documents a model migration.
 Focused machine-readable schemas are available under
 `docs/schemas/altium_toolkit/` for the normalized root plus focused project,
-netlist, SVG, PCB review, layer-stack, Draftsman, library, and CI/reporting
-contracts.
+netlist, SVG, PCB review, layer-stack, Draftsman, library, parser QA,
+inspection, unsupported-feature, and CI/reporting contracts.