data-modelling-sdk 2.0.9

# Changelog

All notable changes to this project will be documented in this file.

The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

## [2.0.9] - 2026-01-28

### Fixed

- **fix(export)**: Stable YAML key ordering for all exporters
  - Replaced manual `serde_yaml::Mapping` construction with struct-based serialization
  - ODCS, ODPS, and CADS exporters now produce consistent key ordering matching struct field definitions
  - Eliminates git diff noise from alphabetical key sorting (BTreeMap behavior)
  - Re-exporting the same data produces identical YAML output

### Changed

- **refactor(odcs)**: Simplified `ODCSExporter::export_table()` to use `ODCSContract::from_table()` converter
  - Removed ~1500 lines of manual Mapping construction code
  - All exports now use the same struct-based serialization path

- **refactor(odps)**: Migrated to struct-based serialization
  - Added `#[serde(rename_all = "camelCase")]` to all ODPS structs
  - Added `skip_serializing_if` for optional fields to produce clean YAML
  - Removed ~764 lines of manual Mapping construction

- **refactor(cads)**: Migrated to struct-based serialization
  - Added `skip_serializing_if` for optional fields
  - Removed ~771 lines of manual Mapping construction

### Added

- **feat(odcs)**: Enhanced Column to Property converter
  - `map_data_type_to_logical_type()` function for SQL to ODCS type mapping
  - `parse_struct_fields_from_data_type()` for expanding STRUCT type definitions
  - Proper handling of ARRAY<STRUCT<...>> nested types

- **feat(pdf)**: Comprehensive column details in PDF/Markdown export
  - Added logical type options (min/max length, pattern, format, precision, scale)
  - Added transformation metadata (source objects, logic, description)
  - Added relationships and authoritative definitions
  - Added column-level quality rules and tags
  - Added composite/secondary key information
  - Added partition key position and additional constraints

- **feat(workspace)**: Added `description` field to Workspace struct (GitHub #66)
  - Optional description field for workspace metadata
  - Displayed in UI and README generation

## [2.0.8] - 2026-01-14

### Fixed

- **chore**: Fix code formatting (cargo fmt)

## [2.0.7] - 2026-01-14

### Fixed

- **fix(odcl)**: Import model-level quality rules from ODCL Data Contract format (#64)
  - Fixed bug where `models.<name>.quality` rules were not being captured during ODCL import
  - Quality rules are now extracted from both root level and model level in Data Contract Specification format
  - Ensures all quality rules defined at the model level are preserved when converting ODCL to ODCS

### Added

- **test(odcl)**: Comprehensive E2E tests for ODCL import
  - 13 new tests verifying ODCL to ODCS conversion preserves all fields
  - Tests based on official datacontract-specification examples
  - Coverage for: model-level quality, field-level quality, metadata, servicelevels, terms, servers, tags, column descriptions, $ref resolution, and roundtrip export

## [2.0.6] - 2026-01-14

### Fixed

- **ci**: Increase crates.io indexing wait time to 60s to fix publish failures

## [2.0.5] - 2026-01-13

### Fixed

- **fix(odcs)**: Preserve customProperties at schema and column levels (#60)
  - Added schema-level `customProperties` extraction in `parse_odcl_v3_all()`
  - Updated `table_to_table_data()` to merge contract-level and schema-level customProperties
  - Custom properties are now preserved at all three ODCS levels: contract, schema, and column

### Added

- **test(odcs)**: Comprehensive E2E round-trip tests for ODCS v3.1.0
  - 51 tests covering all ODCS v3.1.0 specification fields
  - Tests for contract-level, schema-level, and property-level fields
  - Round-trip preservation tests via `import_contract`/`export_contract` v2 API
  - Builder API tests for programmatic contract construction

## [2.0.4] - 2026-01-13

### Added

- **feat(odcs)**: Native ODCS data structures for lossless round-trip
  - New `ODCSContract` struct representing full ODCS v3.1.0 contracts
  - New `SchemaObject` struct for schema-level metadata (physicalName, businessName, etc.)
  - New `Property` struct with nested support (`properties` for OBJECT, `items` for ARRAY)
  - Supporting types: `QualityRule`, `CustomProperty`, `AuthoritativeDefinition`, `Team`, `Server`, etc.
  - Converters: `From` trait implementations for Table/Column interop
  - `ODCSContract::to_tables()` and `ODCSContract::from_tables()` for backwards compatibility
  - `ODCSContract::to_table_data()` for UI rendering

- **feat(import)**: New `import_contract()` method for native ODCS parsing
  - Returns `ODCSContract` instead of flattened `Table`/`Column` types
  - Preserves all contract-level, schema-level, and property-level metadata
  - Supports nested properties (OBJECT/ARRAY) hierarchically
  - Multi-table contracts fully supported

- **feat(export)**: New `export_contract()` method for native ODCS serialization
  - Directly serializes `ODCSContract` to YAML (no reconstruction needed)
  - Zero data loss on round-trip
  - `export_contract_validated()` variant with optional schema validation

- **feat(wasm)**: V2 WASM bindings for native ODCS API
  - `parse_odcs_yaml_v2()` - Returns `ODCSContract` JSON
  - `export_odcs_yaml_v2()` - Takes `ODCSContract` JSON, returns YAML
  - `odcs_contract_to_tables()` - Convert contract to Table array
  - `tables_to_odcs_contract()` - Convert Table array to contract
  - `odcs_contract_to_table_data()` - Convert contract to TableData for UI

## [2.0.3] - 2026-01-12

### Fixed

- **fix(odcs)**: Full round-trip support for all column metadata fields
  - **Export**: Nested STRUCT/ARRAY columns now export all metadata fields matching top-level columns
  - **Import**: Parse all ODCS v3.1.0 column metadata fields during import
  - Fields supported: `businessName`, `physicalName`, `logicalTypeOptions`
  - Fields supported: `unique`, `primaryKey`, `primaryKeyPosition`
  - Fields supported: `partitioned`, `partitionKeyPosition`, `clustered`
  - Fields supported: `classification`, `criticalDataElement`, `encryptedName`
  - Fields supported: `transformSourceObjects`, `transformLogic`, `transformDescription`
  - Fields supported: `examples`, `authoritativeDefinitions`, `tags`, `customProperties`, `businessKey`
  - Fixes data loss for column metadata on round-trip import/export

## [2.0.2] - 2026-01-12

### Fixed

- **fix(odcs)**: Export all column-level metadata fields to ODCS YAML
  - Added export for: `businessName`, `physicalName`, `logicalTypeOptions`
  - Added export for: `primaryKeyPosition`, `partitionKeyPosition`, `unique`, `partitioned`
  - Added export for: `classification`, `criticalDataElement`, `encryptedName`
  - Added export for: `transformSourceObjects`, `transformLogic`, `transformDescription`
  - Added export for: `examples`, `authoritativeDefinitions`, `tags`
  - Added export for: `customProperties` (fixes data loss on round-trip)

## [2.0.1] - 2026-01-12

### Fixed

- **fix(ci)**: Restructure CI/CD workflows for proper publish order
  - Add `version-check.yml` as reusable workflow for version consistency
  - Add `release-sdk.yml` for crates.io publishing (core before sdk)
  - Update `release-cli.yml` to upload artifacts to GitHub release
  - Update `release-wasm.yml` to only run manually or when called
  - Update `publish.yml` as orchestrating wrapper
  - Fix crates.io publish error by publishing dependencies in order
  - Fix `pkg/package.json` not found error (generated during WASM build)
  - Support dev releases with prerelease flag

## [2.0.0] - 2026-01-11

### Fixed

- **fix(pdf)**: Long URLs and field names now wrap correctly in PDF tables (#50)
  - Text that exceeds column width is now broken with hyphen continuation markers
  - Prevents text from overlapping into adjacent columns
  - Works for URLs, long field names, and any text without natural break points

### Added

- **feat(staging)**: Real-time progress reporting for ingestion operations
  - `IngestProgress` with multi-bar display (files, records, bytes throughput)
  - `InferenceProgress` for schema inference operations
  - `Spinner` for indeterminate operations
  - `format_bytes()` and `format_number()` helper functions
  - Uses `indicatif` crate for terminal progress bars

- **feat(s3)**: AWS S3 ingestion support (feature: `s3`)
  - `S3Source` configuration with bucket, prefix, region, profile, endpoint
  - `S3Ingester` with streaming file discovery and download
  - Async file listing with pagination support
  - Memory-efficient streaming downloads
  - Integration with staging pipeline

- **feat(databricks)**: Databricks Unity Catalog Volumes ingestion (feature: `databricks`)
  - `UnityVolumeSource` configuration with host, token, catalog, schema, volume
  - `UnityVolumeIngester` using Databricks Files REST API
  - Secure token handling with `SecureCredentials` wrapper
  - File discovery and download from Unity Catalog Volumes

- **feat(security)**: Secure credential handling
  - `SecureCredentials` wrapper type preventing accidental Display/Debug logging
  - `redact_secret()` function showing only first N characters
  - `redact_secrets_in_string()` with regex-based detection for:
    - AWS access keys (AKIA...)
    - AWS secret keys
    - Bearer tokens
    - URL-embedded passwords
  - All credential types implement safe Display trait

- **feat(tests)**: Comprehensive integration and performance tests
  - 10 integration tests covering full staging pipeline
  - Performance benchmarks for 10K, 100K, and 1M records
  - Criterion benchmarks for staging operations

- **feat(pipeline)**: Full data pipeline orchestration (Phase 6)
  - `PipelineExecutor` for running multi-stage data pipelines
  - `PipelineConfig` with comprehensive configuration options
  - `PipelineStage` enum: Ingest, Infer, Refine, Map, Export
  - Checkpointing and resume support with `Checkpoint` struct
  - Dry-run mode for validation without execution
  - Stage timing and success/failure tracking
  - CLI command: `odm pipeline run` with 15+ options
  - CLI command: `odm pipeline status` for pipeline monitoring
  - 20 unit tests for pipeline functionality

- **feat(mapping)**: Schema-to-schema mapping with transformation generation (Phase 5)
  - `SchemaMatcher` with multiple matching algorithms
  - Match methods: Exact, CaseInsensitive, Fuzzy (Levenshtein), Semantic, LLM
  - `SchemaMapping` result with direct mappings, transformations, gaps, extras
  - Compatibility scoring (0.0-1.0)
  - `TransformType` variants: TypeCast, Rename, Merge, Split, FormatChange, Custom, Extract, Default
  - Transformation script generation: SQL, JQ, Python, PySpark
  - CLI command: `odm map` with fuzzy matching and transform output
  - 29 unit tests for mapping functionality

- **feat(llm-matcher)**: LLM-enhanced schema matching
  - `LlmSchemaMatcher` combining algorithmic and LLM matching
  - Retry logic with exponential backoff for LLM failures
  - Batching for large schemas exceeding prompt limits
  - Transform hint parsing to create `TransformMapping`s
  - Example value inclusion in prompts for better matching
  - Feature-gated under `llm` feature
  - 12 unit tests for LLM matching

- **feat(iceberg)**: Apache Iceberg integration for data lakehouse staging (Phase 3)
  - New `iceberg` feature flag for optional Iceberg support
  - Added `arrow = "55"` and `parquet = "55"` dependencies for data writing
  - Catalog abstraction supporting REST (Lakekeeper/Nessie/Polaris), S3 Tables, Unity Catalog, and Glue
  - `CatalogConfig` enum with serializable configuration for all catalog types
  - `IcebergCatalog` wrapper with async operations for namespace and table management
  - `IcebergTable` for raw JSON staging with time travel support
    - `append_records()` method converts JSON records to Arrow RecordBatch and writes Parquet files
    - `write_parquet_file()` helper with SNAPPY compression
  - `ingest_to_iceberg()` function for batch ingestion from local files to Iceberg tables
  - `to_raw_json_records()` conversion function for file-to-record transformation
  - Export module (`staging/export.rs`) with:
    - `ExportTarget` enum (Unity, Glue, S3Tables, Local)
    - `ExportConfig` and `ExportResult` structs
    - `export_to_catalog()` function with local filesystem export support
  - Snapshot listing and time travel queries by version or timestamp
  - Batch metadata storage in Iceberg table properties
  - New CLI commands:
    - `odm staging init --catalog <type>` - Initialize with Iceberg catalog (rest, s3-tables, unity, glue)
    - `odm staging history` - Show table version history
    - `odm staging query --version N` - Time travel queries by snapshot version
    - `odm staging query --timestamp <ISO8601>` - Time travel queries by timestamp
    - `odm staging export --target <unity|glue|s3-tables>` - Export to production catalogs
    - `odm staging view create` - Create typed views from inferred schemas

### Changed

- **refactor(workspace)**: Restructured monolithic crate into Cargo workspace with three sub-crates
  - `data-modelling-core` - Core SDK library with all business logic
  - `odm` - CLI binary (renamed from `data-modelling-cli`)
  - `data-modelling-wasm` - WASM bindings for browser applications
  - Root `data-modelling-sdk` crate remains for backward compatibility via re-exports
  - Uses Cargo workspace with `resolver = "2"` for proper feature unification
  - Added `wasm` feature to core crate for browser storage backend dependencies
  - Updated all CI/CD workflows for workspace structure

## [1.14.3] - 2026-01-11

### Added

- **feat(serde)**: Dual-case JSON field name support for WASM deserialization
  - All struct fields with `#[serde(rename_all = "camelCase")]` now accept both camelCase and snake_case during deserialization
  - Added `#[serde(alias = "snake_case")]` attributes to all multi-word field names
  - Serialization output remains camelCase for consistency
  - Updated structs: `Relationship`, `ForeignKeyDetails`, `ETLJobMetadata`, `VisualMetadata`, `Column`, `ForeignKey`, `PropertyRelationship`, `LogicalTypeOptions`, `AuthoritativeDefinition`, `Workspace`, `AssetReference`, `DomainReference`, `SystemReference`, `Table`, `SlaProperty`, `ContactDetails`, `CADSAsset`, `CADSDescription`, `CADSPricing`, `CADSRisk`, `CADSValidationProfile`, `CADSValidationProfileAppliesTo`, `KnowledgeArticle`, `KnowledgeIndex`, `KnowledgeIndexEntry`, `RelatedArticle`, `Decision`, `DecisionIndex`, `AssetLink`, `ComplianceAssessment`
  - Enables UI flexibility when passing JSON with either naming convention

## [1.14.3] - 2026-01-11

### Added

- **feat(serde)**: Dual-case JSON field name support for WASM deserialization
  - All struct fields with `#[serde(rename_all = "camelCase")]` now accept both camelCase and snake_case during deserialization
  - Added `#[serde(alias = "snake_case")]` attributes to all multi-word field names
  - Serialization output remains camelCase for consistency
  - Updated structs: `Relationship`, `ForeignKeyDetails`, `ETLJobMetadata`, `VisualMetadata`, `Column`, `ForeignKey`, `PropertyRelationship`, `LogicalTypeOptions`, `AuthoritativeDefinition`, `Workspace`, `AssetReference`, `DomainReference`, `SystemReference`, `Table`, `SlaProperty`, `ContactDetails`, `CADSAsset`, `CADSDescription`, `CADSPricing`, `CADSRisk`, `CADSValidationProfile`, `CADSValidationProfileAppliesTo`, `KnowledgeArticle`, `KnowledgeIndex`, `KnowledgeIndexEntry`, `RelatedArticle`, `Decision`, `DecisionIndex`, `AssetLink`, `ComplianceAssessment`
  - Enables UI flexibility when passing JSON with either naming convention

## [1.14.2] - 2026-01-11

### Added

- **feat(wasm)**: Native YAML format PDF/Markdown export bindings for all document types
  - `export_odcs_yaml_to_pdf` - Export ODCS YAML content directly to PDF
  - `export_odcs_yaml_to_markdown` - Export ODCS YAML content directly to Markdown
  - `export_odps_yaml_to_pdf` - Export ODPS YAML content directly to PDF
  - `export_odps_yaml_to_markdown` - Export ODPS YAML content directly to Markdown
  - `export_cads_yaml_to_pdf` - Export CADS YAML content directly to PDF
  - `export_cads_yaml_to_markdown` - Export CADS YAML content directly to Markdown
  - `export_decision_yaml_to_pdf` - Export Decision (ADR) YAML content directly to PDF
  - `export_decision_yaml_to_markdown` - Export Decision (ADR) YAML content directly to Markdown
  - `export_knowledge_yaml_to_pdf` - Export Knowledge Base YAML content directly to PDF
  - `export_knowledge_yaml_to_markdown` - Export Knowledge Base YAML content directly to Markdown
  - These functions accept raw YAML content (as stored in .odcs.yaml, .odps.yaml, .cads.yaml, .madr.yaml, .kb.yaml files)
  - Enables direct export from UI without requiring conversion to internal SDK structs

- **feat(models)**: Added `Table::from_table_data` for converting import data to Table model

## [1.14.1] - 2026-01-11

### Added

- **feat(wasm)**: ODCS/ODPS/CADS PDF and Markdown export bindings
  - `export_table_to_pdf` - Export ODCS Data Contract to PDF with optional branding
  - `export_odps_to_pdf` - Export ODPS Data Product to PDF with optional branding
  - `export_cads_to_pdf` - Export CADS Asset to PDF with optional branding
  - `export_table_to_markdown` - Export ODCS Data Contract to Markdown
  - `export_odps_to_markdown` - Export ODPS Data Product to Markdown
  - `export_cads_to_markdown` - Export CADS Asset to Markdown

- **feat(models)**: Added `Table::from_table_data` for converting import data to Table model

- **feat(cli)**: Extended PDF and Markdown export to support all contract types
  - `data-modelling-cli export pdf` now supports .odcs.yaml, .odps.yaml, .cads.yaml files
  - `data-modelling-cli export markdown` now supports .odcs.yaml, .odps.yaml, .cads.yaml files
  - Auto-detection of document type based on content

### Fixed

- **fix(npm)**: Corrected npm organization name to `@offenedatenmodellierung`
  - Fixed package name in release-cli.yml and release-wasm.yml workflows

## [1.14.0] - 2026-01-11

### Added

- **feat(relationship)**: Crow's feet notation cardinality support for ERD-style relationships
  - New `EndpointCardinality` enum with values: `zeroOrOne` (0..1), `exactlyOne` (1..1), `zeroOrMany` (0..*), `oneOrMany` (1..*)
  - New `sourceCardinality` and `targetCardinality` fields on `Relationship` struct
  - Enables precise endpoint cardinality modeling for data flow diagrams

- **feat(relationship)**: Data flow direction support
  - New `FlowDirection` enum with values: `sourceToTarget`, `targetToSource`, `bidirectional`
  - New `flowDirection` field on `Relationship` struct
  - Enables directional data flow modeling in relationship diagrams

- **feat(workspace)**: Extended workspace schema with full relationship metadata
  - Added `ForeignKeyDetails`, `ETLJobMetadata`, `SlaProperty`, `ContactDetails`, `VisualMetadata`, `ConnectionPoint` definitions
  - All relationship fields now available in workspace-schema.json

### Changed

- **refactor(serialization)**: Consistent camelCase serialization across all models
  - All structs now use `#[serde(rename_all = "camelCase")]` for JSON/YAML serialization
  - Updated: `Workspace`, `DomainReference`, `SystemReference`, `AssetReference`, `Relationship`, `ForeignKeyDetails`, `ETLJobMetadata`, `VisualMetadata`
  - Updated: `Cardinality`, `RelationshipType` enums now serialize as camelCase
  - Updated: `Decision`, `DecisionIndex`, `KnowledgeArticle`, `KnowledgeIndex` structs
  - Updated: `CADSAsset` and all nested CADS structs
  - Aligns with ODCS format conventions for consistent API responses

- **refactor(enums)**: Updated enum serialization to camelCase
  - `Cardinality`: `oneToOne`, `oneToMany`, `manyToOne`, `manyToMany`
  - `RelationshipType`: `dataFlow`, `dependency`, `foreignKey`, `etl`
  - `EndpointCardinality`: `zeroOrOne`, `exactlyOne`, `zeroOrMany`, `oneOrMany`
  - `FlowDirection`: `sourceToTarget`, `targetToSource`, `bidirectional`

- **refactor(schema)**: Updated workspace-schema.json to camelCase field names
  - All property names now use camelCase (e.g., `sourceTableId`, `targetTableId`, `createdAt`)
  - Relationship definition includes new cardinality and flow direction fields
  - Example updated to demonstrate crow's feet notation usage

### Fixed

- **fix(export)**: Added missing `CADSKind::DataPipeline` and `CADSKind::ETLProcess` match arms in CADS exporter
- **fix(export)**: Added missing `CADSDMNFormat::Json` match arm in CADS exporter

## [1.13.6] - 2026-01-08

### Added

- **feat(workspace)**: Add `view_positions` to DomainReference for per-view canvas node positioning (#37)
  - Supports view types: `operational`, `analytical`, `process`, `systems`
  - Each node (table, system, asset) stores x/y coordinates specific to its view context
  - Matches existing `DomainConfig.view_positions` structure for consistency

- **feat(relationship)**: Add `source_handle` and `target_handle` for edge connection points (#37)
  - New `ConnectionHandle` enum with 12 positions around node perimeter
  - Positions: `top-left`, `top-center`, `top-right`, `right-top`, `right-center`, `right-bottom`, `bottom-right`, `bottom-center`, `bottom-left`, `left-bottom`, `left-center`, `left-top`
  - Enables precise control over edge attachment points to prevent overlapping lines

## [1.13.5] - 2026-01-08

### Added

- **feat(import)**: Preserve all ODCS v3.1.0 contract-level fields during import (#35)
  - `TableData` struct now includes: `id`, `apiVersion`, `version`, `status`, `kind`, `domain`, `dataProduct`, `tenant`, `description`, `servers`, `team`, `support`, `roles`, `slaProperties`, `quality`, `price`, `tags`, `customProperties`, `authoritativeDefinitions`, `contractCreatedTs`, `odcsMetadata`
  - ODCS and ODCL importers extract all fields from source YAML
  - Other importers (Avro, JSON Schema, Protobuf, SQL) use defaults for ODCS-specific fields

- **feat(workspace)**: Add `table_ids` and `asset_ids` to SystemReference (#34)
  - Optional UUID arrays for explicit table-to-system and asset-to-system mapping
  - Enables direct relationships without relying on `customProperties.system_id`
  - Updated workspace JSON schema and Rust types

### Fixed

- **fix(import)**: ODCS `id` field now correctly preserved in import results (#35)
  - Previously the contract UUID was extracted but lost during `TableData` construction
  - Tables now retain their source UUIDs for proper system linkage

## [1.13.4] - 2026-01-08

### Added

- **feat(sql)**: Databricks SQL `USING` clause support for table format specifications
  - `USING DELTA` - Delta Lake tables
  - `USING PARQUET` - Parquet format tables
  - `USING CSV`, `USING JSON`, `USING ORC`, `USING AVRO` - Other formats
  - `USING ICEBERG` - Apache Iceberg tables
  - These clauses are now stripped during preprocessing, allowing table imports to succeed

### Fixed

- **fix(sql)**: Databricks SQL imports now correctly handle:
  - `CREATE TABLE ... USING DELTA` statements
  - `CREATE VIEW` statements (already supported, now tested)
  - `CREATE MATERIALIZED VIEW` statements (already supported, now tested)
  - Combined clauses like `USING DELTA TBLPROPERTIES (...)`

## [1.13.3] - 2026-01-08

### Fixed

- **fix(wasm)**: Rebuild WASM package to include v1.13.2 Decision Log and Knowledge Base bindings
  - NPM package was stuck at v1.13.1 and missing all DDL/KB WASM exports
  - Rebuilt with `wasm-pack build --target web --out-dir pkg --features wasm,openapi,odps-validation`
  - All 83 WASM functions now properly exported in the NPM package

## [1.13.2] - 2026-01-07

### Added

- **feat(wasm)**: Decision Log (DDL) WASM bindings for browser support
  - `parse_decision_yaml()` - Parse decision YAML files (.madr.yaml)
  - `parse_decision_index_yaml()` - Parse decisions index file (decisions.yaml)
  - `export_decision_to_yaml()` - Export decision to YAML format
  - `export_decision_index_to_yaml()` - Export decision index to YAML format
  - `export_decision_to_markdown()` - Export decision to MADR Markdown format
  - `create_decision()` - Create new decision with required fields
  - `create_decision_index()` - Create empty decision index
  - `add_decision_to_index()` - Add decision to index

- **feat(wasm)**: Knowledge Base (KB) WASM bindings for browser support
  - `parse_knowledge_yaml()` - Parse knowledge article YAML files (.kb.yaml)
  - `parse_knowledge_index_yaml()` - Parse knowledge index file (knowledge.yaml)
  - `export_knowledge_to_yaml()` - Export article to YAML format
  - `export_knowledge_index_to_yaml()` - Export knowledge index to YAML format
  - `export_knowledge_to_markdown()` - Export article to Markdown format
  - `create_knowledge_article()` - Create new knowledge article
  - `create_knowledge_index()` - Create empty knowledge index
  - `add_article_to_knowledge_index()` - Add article to index
  - `search_knowledge_articles()` - Search articles by title, summary, content, or tags

### Changed

- WASM package now exports 83 functions (up from 66 in v1.13.1)

## [1.13.1] - 2026-01-07

### Fixed

- **fix(validation)**: Move schema validation from CLI module to core validation module
  - Schema validation functions are now in `validation::schema` instead of `cli::validation`
  - Fixes WASM build failure when `cli` feature is not enabled but `schema-validation` is
  - Validation functions are now available to all SDK consumers (WASM, native, API)
  - Removed duplicate inline validation code from import/export modules
  - Removed redundant `cli/validation.rs` wrapper module

## [1.13.0] - 2026-01-07

### Added

- **feat(ddl)**: Architecture Decision Records (DDL) with MADR format support
  - `Decision` model with full MADR (Markdown Any Decision Records) field support
  - Decision lifecycle: draft, proposed, accepted, deprecated, superseded, rejected
  - Decision categories: architecture, technology, process, security, data, integration
  - `DecisionIndex` for tracking decision metadata and numbering
  - `DecisionImporter` for parsing decision YAML files
  - `DecisionExporter` for exporting decisions to Markdown format
  - Storage integration: `load_decisions()`, `save_decision()`, `export_decision_markdown()`
  - Database schema with `decisions` table and comprehensive indexes
  - CLI commands: `decision new`, `decision list`, `decision show`, `decision status`, `decision export`

- **feat(kb)**: Knowledge Base (KB) for domain documentation
  - `KnowledgeArticle` model for storing guides, references, tutorials, and runbooks
  - Article types: guide, reference, concept, tutorial, troubleshooting, runbook
  - Article lifecycle: draft, review, published, archived, deprecated
  - `KnowledgeIndex` for tracking article metadata and numbering
  - `KnowledgeImporter` for parsing knowledge YAML files
  - `KnowledgeExporter` for exporting articles to Markdown format
  - Storage integration: `load_knowledge()`, `save_knowledge()`, `export_knowledge_markdown()`
  - Database schema with `knowledge_articles` table and comprehensive indexes
  - CLI commands: `knowledge new`, `knowledge list`, `knowledge show`, `knowledge search`, `knowledge status`, `knowledge export`

- **feat(database)**: Extended database schema for DDL/KB (schema version 2)
  - `decisions` table with all MADR fields: number, title, status, category, date, deciders, context, drivers, options, decision, consequences, linked_assets, supersedes, superseded_by, compliance, rationale, additional_context
  - `knowledge_articles` table with all KB fields: number, title, article_type, status, summary, content, author, author_email, reviewers, linked_assets, linked_decisions, related_articles, skill_level, estimated_reading_time, review_frequency
  - Indexes for workspace, domain, status, category, type, author, and number queries
  - SQL modules: `decision_sql` and `knowledge_sql` with UPSERT, SELECT, DELETE, COUNT operations

- **feat(sync)**: Extended sync engine for DDL/KB
  - `sync_decisions()` and `sync_knowledge()` methods in DatabaseBackend trait
  - `export_decisions()` and `export_knowledge()` methods for database export
  - `sync_workspace_full()` now includes decisions and knowledge articles
  - `export_workspace_full()` returns complete workspace including DDL/KB content
  - Extended `SyncResult` with `decisions_synced` and `knowledge_synced` counts
  - Extended `SyncStatus` with `decision_count` and `knowledge_count`

### Documentation

- Updated README.md with Decision Records and Knowledge Base sections
- Updated CLI.md with `decision` and `knowledge` command documentation
- Updated LLM.txt with DDL/KB modules and architecture

## [1.12.0] - 2026-01-07

### Added

- **feat(workspace)**: Relationships now embedded in workspace.yaml
  - Added `relationships` field to Workspace struct
  - Added Relationship definition to workspace-schema.json
  - WASM bindings for relationship operations: `add_relationship_to_workspace()`, `remove_relationship_from_workspace()`, `get_workspace_relationships_for_source()`, `get_workspace_relationships_for_target()`
  - Relationships.yaml file is no longer required (merged into workspace schema)

- **feat(models)**: Added `color` field to Relationship model
  - Supports hex color codes (#RRGGBB, #RGB) or named colors
  - Used for relationship line colors in the UI
  - Pattern validation in workspace-schema.json

- **feat(database)**: Database schema updated with new relationship fields
  - Added `drawio_edge_id` column to relationships table for diagram integration
  - Added `color` column to relationships table for UI display
  - Updated DuckDB backend `sync_relationships` with new fields
  - Updated PostgreSQL backend `sync_relationships` with new fields

- **feat(schemas)**: Added workspace-schema.json
  - Complete JSON Schema for workspace.yaml validation
  - Includes Relationship definition with all fields

### Changed

- **refactor(storage)**: Flat file naming convention
  - Files now use format: `{workspace}_{domain}_{system}_{resource}.{type}.yaml`
  - Removed legacy domain-based directory structure
  - Updated README.md and ARCHITECTURE.md documentation

- **refactor(schemas)**: Removed domain-schema.json
  - Domain functionality now merged into workspace schema
  - Added dmnModels section to CADS schema
  - Added openApiSpecs section to CADS schema

### Documentation

- Updated LLM.txt with flat file structure and workspace relationships
- Updated README.md with new file naming convention
- Updated docs/ARCHITECTURE.md with flat structure documentation
- Updated schemas/README.md with workspace-schema.json

## [1.11.0] - 2026-01-07

### Added

- **feat(models)**: Complete ODCS v3.1.0 property field support in Column model
  - Added 25+ new fields: `id`, `businessName`, `physicalName`, `logicalTypeOptions`, `primaryKeyPosition`, `unique`, `partitioned`, `partitionKeyPosition`, `clustered`, `classification`, `criticalDataElement`, `encryptedName`, `transformSourceObjects`, `transformLogic`, `transformDescription`, `examples`, `defaultValue`, `authoritativeDefinitions`, `tags`, `customProperties`
  - Ensures zero data loss during import/export round-trips

- **feat(import)**: ColumnData struct now mirrors Column exactly
  - All ODCS v3.1.0 property fields preserved through WASM layer
  - Added `column_to_column_data()` helper for consistent Column→ColumnData conversion
  - Updated `column_data_to_column()` to preserve all fields during reconstruction

- **feat(models)**: Added new supporting types
  - `LogicalTypeOptions`: minLength, maxLength, pattern, format, minimum, maximum, precision, scale
  - `AuthoritativeDefinition`: type and URL for authoritative definition references
  - Both types exported from models module for external use

### Changed

- **refactor(import)**: All importers now use `..Default::default()` pattern for Column/ColumnData construction
  - Simplifies code and ensures new fields automatically get sensible defaults
  - Applied across: avro.rs, json_schema.rs, protobuf.rs, odcs.rs, odcl.rs, odcs_shared.rs, sql.rs

## [1.10.0] - 2026-01-06

### Added

- **feat(cli)**: Added `validate` command for standalone schema validation
  - Validates files against their respective schemas without importing
  - Supports formats: ODCS, ODCL, ODPS, CADS, OpenAPI, Protobuf, Avro, JSON Schema, SQL
  - Usage: `data-modelling-cli validate <format> [file]`
  - Accepts stdin with `-` as input

- **feat(import)**: Fixed protobuf JAR import to properly create parent columns for nested messages
  - Nested messages now create parent column with `OBJECT` or `ARRAY<OBJECT>` type
  - Repeated message fields use `.[]` notation for nested properties
  - ODCS export now correctly includes nested column structure

### Changed

- **refactor(validation)**: ODPS and OpenAPI validation now only require `schema-validation` feature
  - Removed separate `odps-validation` and `openapi` feature requirements for validation
  - Simplifies feature configuration for validation-only use cases

## [1.9.0] - 2025-01-06

### Added

- **feat(import)**: Added support for Google protobuf well-known wrapper types in JAR imports
  - `google.protobuf.StringValue` → `STRING`
  - `google.protobuf.Int32Value` → `INTEGER`
  - `google.protobuf.Int64Value` → `BIGINT`
  - `google.protobuf.DoubleValue` → `DOUBLE`
  - `google.protobuf.FloatValue` → `FLOAT`
  - `google.protobuf.BoolValue` → `BOOLEAN`
  - `google.protobuf.Timestamp` → `TIMESTAMP`
  - Wrapper types are correctly marked as nullable

- **feat(import)**: Added nested message extraction for protobuf JAR imports
  - Now extracts nested messages (e.g., `Bet.Cashout`, `Leg.EachWay`)
  - Supports deeply nested message hierarchies
  - Enums are now recognized and mapped to `STRING` type

- **feat(import)**: Consistent STRUCT flattening across all import types
  - All WASM import functions now return consistent flattened format
  - `STRUCT<...>` types flatten to `parent.field` notation
  - `ARRAY<STRUCT<...>>` types flatten to `parent.[].field` notation
  - Applied to: SQL, ODCS, ODCL, Avro, JSON Schema, Protobuf imports

### Fixed

- **fix(import)**: Fixed protobuf JAR parser incorrectly parsing multi-line comments as fields
  - Added proper tracking for `/* ... */` and `/** ... */` block comments
  - Added validation that field names must be valid identifiers
  - Added check that proto field lines must contain `=` for field number

- **fix(import)**: Fixed WASM SQL import not flattening STRUCT columns
  - WASM `import_from_sql` now returns nested columns with dot notation
  - Matches CLI behavior for consistent UI rendering

## [1.8.6] - 2025-01-06

### Fixed

- Fixed macOS code signing by ensuring MACOS_SIGNING_IDENTITY secret is properly configured

## [1.8.5] - 2025-01-06

### Added

- **feat(cli)**: Improved protobuf JAR import with dependency graph analysis
  - Made `INPUT` argument optional when `--jar` is provided
  - Added `--root-message` flag to specify root message for JAR imports
  - Implemented dependency graph analysis to auto-detect root message
  - Flattens all referenced messages into single unified schema with dot notation
  - Outputs single combined ODCS file instead of multiple files
  - Root message detection: prefers messages with no incoming refs and most outgoing refs
  - Pretty mode shows JAR analysis with message list and dependencies

- **feat(ci)**: Added optional macOS code signing and notarization to release workflow
  - Signing is skipped gracefully if secrets are not configured
  - Supports Developer ID Application certificates
  - Supports Apple notarization for Gatekeeper approval

## [1.8.4] - 2025-01-06

### Fixed

- Fixed workflow_call version detection using environment variables
- Fixed dev suffix only applying to release/* branches, not main

## [1.8.3] - 2025-01-06

### Changed

- Merged release-cli and release-wasm workflows into a single workflow
- Version is now passed from publish workflow to release workflow
- Removed blocking tag condition from create-release job

## [1.8.2] - 2025-01-05

### Added

- **feat(wasm)**: Workspace and DomainConfig WASM bindings
  - `create_workspace()` - Create a new workspace with owner
  - `parse_workspace_yaml()` - Parse workspace YAML to JSON
  - `export_workspace_to_yaml()` - Export workspace JSON to YAML
  - `add_domain_to_workspace()` - Add a domain reference to workspace
  - `remove_domain_from_workspace()` - Remove a domain reference from workspace
  - `create_domain_config()` - Create a new domain configuration
  - `parse_domain_config_yaml()` - Parse domain config YAML to JSON
  - `export_domain_config_to_yaml()` - Export domain config JSON to YAML
  - `get_domain_config_id()` - Extract domain ID from config
  - `update_domain_view_positions()` - Update view positions in domain config
  - `add_entity_to_domain_config()` - Add entity reference (system, table, product, asset, process, decision)
  - `remove_entity_from_domain_config()` - Remove entity reference from domain config

- **feat(loader)**: Workspace and DomainConfig loading/saving
  - `load_workspace()` / `save_workspace()` - Load/save workspace.yaml files
  - `load_domain_config()` / `save_domain_config()` - Load/save domain.yaml files
  - `load_domain_config_by_name()` - Load domain config by name within a workspace
  - `get_domain_id()` - Extract domain ID from domain.yaml
  - `load_all_domain_configs()` - Load all domain configs in a workspace

### Fixed

- **fix(loader)**: Domain ID extraction from domain.yaml
  - BPMN, DMN, and OpenAPI loaders now extract domain ID from domain.yaml instead of generating random UUIDs
  - Falls back to new UUID if domain.yaml is not found

## [1.8.1] - 2026-01-05

### Fixed

- **fix(wasm)**: Fixed WASM build compatibility issues
  - Upgraded jsonschema from 0.20 to 0.38.1 with `default-features = false` to avoid reqwest dependency in WASM builds
  - Updated validation error handling to use new jsonschema 0.38.1 API (single ValidationError instead of iterator)
  - Configured getrandom 0.3.4 with `wasm_js` feature for proper WASM support
  - Added `.cargo/config.toml` with `getrandom_backend="wasm_js"` configuration
  - WASM builds now work correctly with `wasm,openapi,odps-validation` features

## [1.8.0] - 2026-01-05

### Added

- **feat(odps)**: ODPS schema validation and CLI support
  - ODPS import/export with JSON Schema validation (requires `odps-validation` feature)
  - CLI commands: `import odps` and `export odps` for ODPS YAML files
  - Manual test script (`test-odps`) for round-trip testing with field preservation verification
  - WASM binding: `validateOdps()` function for JavaScript validation
  - Field preservation: All ODPS fields (required and optional) preserved during import/export round-trips
  - Empty array preservation: Empty optional arrays maintained in exported YAML
  - Feature flag support: Validation can be enabled/disabled via `odps-validation` feature flag
  - Enhanced ODPS import display: Shows tags, description, managementPorts, support, and team information

### Fixed

- **fix(cli)**: Fixed pre-commit hook issues
  - Fixed formatting and clippy warnings
  - Added feature declaration for `odps-validation` in Cargo.toml lints section
  - Fixed collapsed if statements and unnecessary unwrap patterns

## [1.7.1] - 2026-01-05

### Fixed

- **fix(ci)**: Fixed GitHub Actions workflow for Windows builds
  - Use PowerShell `Compress-Archive` instead of `zip` command for Windows artifacts
  - Explicitly set `shell: bash` for all bash-based steps to avoid PowerShell parsing errors
  - Added cross-platform checksum generation support

## [1.7.0] - 2026-01-05

### Added

- **feat(cli)**: Full CLI wrapper implementation (`data-modelling-cli`)
  - Comprehensive import/export commands for all supported formats (SQL, AVRO, JSON Schema, Protobuf, OpenAPI, ODCS)
  - Support for importing SQL files with various dialects (PostgreSQL, MySQL, SQLite, Generic, Databricks)
  - Support for `CREATE VIEW` and `CREATE MATERIALIZED VIEW` statements
  - Import from AVRO, JSON Schema, Protobuf, OpenAPI, and ODCS formats
  - Export to ODCS, AVRO, JSON Schema, Protobuf (proto2/proto3), and Protobuf descriptor formats
  - UUID override support for imported tables (`--uuid` flag)
  - Automatic ODCS file generation on import (can be disabled with `--no-odcs`)
  - External reference resolution for JSON Schema and OpenAPI (`$ref`, `import`)
  - Schema validation before import (optional `--validate` flag)
  - Protobuf descriptor export using `protoc` compiler
  - Comprehensive error messages with platform-specific installation guidance
  - GitHub Actions workflow for building and releasing CLI binaries (Linux, macOS, Windows)
  - Full test coverage for all import/export operations

- **feat(protobuf)**: Proto2 and Proto3 format support
  - `--protobuf-version` flag to select proto2 or proto3 (default: proto3)
  - Correct field labeling: proto2 uses `required`/`optional`, proto3 uses optional by default
  - Proper handling of `repeated` fields in both versions
  - Protobuf descriptor export supports both proto2 and proto3

- **feat(cli)**: OpenAPI support enabled by default
  - CLI builds now include OpenAPI feature by default in GitHub releases
  - CI/CD workflows updated to build with `openapi` feature enabled
  - Documentation updated with OpenAPI build instructions

### Changed

- **refactor(cli)**: Export operations now accept `.odcs.yaml` files directly as input
  - Removed requirement for `workspace.json` file
  - Simplified export workflow: `data-modelling-cli export <format> input.odcs.yaml output.<ext>`

### Fixed

- **fix(protobuf)**: Fixed invalid `optional repeated` syntax in proto3 exports
  - `repeated` fields no longer include `optional` keyword (proto3 compliant)
  - Proto2 correctly uses `required` for non-nullable fields

## [1.6.2] - 2026-01-04

### Added

- **feat(json-schema)**: Full JSON Schema validation support
  - Import: Extract all JSON Schema validation keywords (pattern, minimum, maximum, minLength, maxLength, enum, const, multipleOf, minItems, maxItems, uniqueItems, minProperties, maxProperties, allOf, anyOf, oneOf, not)
  - Export: Export validation conditions from Column quality rules and enum_values back to JSON Schema format
  - Added enum_values field to ColumnData structure to preserve enumeration values through import pipeline
  - Validation keywords stored as quality rules with source="json_schema" for proper round-trip preservation
  - Comprehensive integration test for validation conditions round-trip

## [1.6.1] - 2026-01-04

### Fixed

- **fix(sql)**: Fixed Databricks SQL parser issues with COMMENT clauses
  - Fixed multiline COMMENT clause parsing by converting newlines to spaces within quoted strings
  - Fixed escaped quotes in COMMENT clauses (`\'` converted to SQL standard `''`)
  - Added support for `MAP<...>` complex types (extracted and restored like STRUCT/ARRAY)
  - Added preprocessing to remove unsupported table-level `COMMENT` clauses
  - Added preprocessing to remove unsupported `CLUSTER BY AUTO` clauses
  - Improved SQL normalization to preserve quoted strings correctly

### Added

- **feat(sql)**: Added SQL parser test tool (`cargo run --example test_sql`)
  - Command-line tool for manually testing SQL parsing with different dialects
  - Supports reading from file, stdin, or command-line argument
  - Pretty-print option for detailed column information
  - Convenience script (`test_sql.sh`) for quick testing

## [1.6.0] - 2026-01-03

### Added

- **feat(sql)**: Databricks SQL dialect support
  - `DatabricksDialect` implementation for parsing Databricks-specific SQL syntax
  - Support for `IDENTIFIER()` function calls in table/view names with variable references and string concatenation
  - Support for variable references (`:variable_name`) in `STRUCT` and `ARRAY` type definitions
  - Support for variable references in column definitions (`column_name :variable TYPE`)
  - Support for variable references in `COMMENT` clauses
  - Support for `CREATE VIEW` and `CREATE MATERIALIZED VIEW` statements
  - Complex type extraction and restoration for `STRUCT<...>` and `ARRAY<...>` types
  - SQL normalization for handling multiline SQL statements
  - Comprehensive test coverage with unit and integration tests
  - Backward compatibility maintained with existing SQL dialects (PostgreSQL, MySQL, SQLite, Generic)

### Changed

- **docs**: Updated documentation to list all supported SQL dialects (PostgreSQL, MySQL, SQLite, Generic, Databricks)
- **docs**: Enhanced `SQLImporter` documentation with Databricks dialect examples and usage

## [1.5.0] - 2026-01-03

### Added

- **feat(bpmn)**: BPMN 2.0 model support
  - `BPMNImporter` for importing BPMN XML files with validation
  - `BPMNExporter` for exporting BPMN models in native XML format
  - `BPMNModel` struct for representing BPMN models
  - WASM bindings: `importBpmnModel()`, `exportBpmnModel()`
  - Storage integration: Models saved to `{domain_name}/{model_name}.bpmn.xml`
  - CADS asset references: CADS assets can reference BPMN models via `bpmn_models` field

- **feat(dmn)**: DMN 1.3 model support
  - `DMNImporter` for importing DMN XML files with validation
  - `DMNExporter` for exporting DMN models in native XML format
  - `DMNModel` struct for representing DMN models
  - WASM bindings: `importDmnModel()`, `exportDmnModel()`
  - Storage integration: Models saved to `{domain_name}/{model_name}.dmn.xml`
  - CADS asset references: CADS assets can reference DMN models via `dmn_models` field

- **feat(openapi)**: OpenAPI 3.1.1 specification support
  - `OpenAPIImporter` for importing OpenAPI YAML/JSON files with validation
  - `OpenAPIExporter` for exporting OpenAPI specs with YAML ↔ JSON conversion
  - `OpenAPIModel` struct for representing OpenAPI specifications
  - WASM bindings: `importOpenapiSpec()`, `exportOpenapiSpec()`
  - Storage integration: Specs saved to `{domain_name}/{api_name}.openapi.yaml` or `.openapi.json`
  - CADS asset references: CADS assets can reference OpenAPI specs via `openapi_specs` field

- **feat(converter)**: OpenAPI to ODCS converter
  - `OpenAPIToODCSConverter` for converting OpenAPI schema components to ODCS tables
  - Type mapping: OpenAPI types (string, integer, number, boolean) mapped to ODCS types
  - Constraint preservation: minLength, maxLength, pattern, minimum, maximum, enum converted to ODCS quality rules
  - Format support: date, date-time, email, uri, uuid, password formats handled
  - WASM bindings: `convertOpenapiToOdcs()`, `analyzeOpenapiConversion()`
  - Conversion reports with warnings and field mappings

- **feat(cads)**: Extended CADS asset references
  - `CADSDMNModel` struct for DMN model references
  - `CADSOpenAPISpec` struct for OpenAPI spec references
  - `CADSDMNFormat` and `CADSOpenAPIFormat` enums
  - CADS importer/exporter support for `dmn_models` and `openapi_specs` fields

- **feat(schemas)**: Schema reference files
  - BPMN 2.0 XSD schema placeholder in `schemas/bpmn-2.0.xsd`
  - DMN 1.3 XSD schema placeholder in `schemas/dmn-1.3.xsd`
  - OpenAPI 3.1.1 JSON Schema placeholder in `schemas/openapi-3.1.1.json`

### Changed

- **refactor(features)**: Feature flags for optional formats
  - `bpmn` feature flag for BPMN support (requires `quick-xml`)
  - `dmn` feature flag for DMN support (requires `quick-xml`)
  - `openapi` feature flag for OpenAPI support (uses existing `jsonschema`)

## [1.4.0] - 2026-01-03

### Added

- **feat(storage)**: Domain-based file organization
  - Domain directories with `domain.yaml` files
  - ODCS tables saved as `{name}.odcs.yaml` within domain directories
  - ODPS products saved as `{name}.odps.yaml` within domain directories
  - CADS assets saved as `{name}.cads.yaml` within domain directories
  - `ModelSaver::save_domain()` for saving complete domains with all associated files
  - `ModelLoader::load_domains()` for loading domains from domain directories
  - `ModelLoader::load_domains_from_list()` for loading specific domains
  - Backward compatibility maintained with legacy `tables/` directory structure

- **feat(schemas)**: Schema reference directory
  - `schemas/` directory with JSON Schema definitions for all supported formats
  - ODCS v3.1.0, ODCL v1.2.1, ODPS, and CADS schemas maintained for validation and reference
  - Comprehensive schema documentation in `schemas/README.md`

- **feat(docs)**: Architecture guide
  - Comprehensive Architecture Guide (`docs/ARCHITECTURE.md`) covering project decisions, use cases, and integration patterns
  - Updated README.md with domain-based file structure documentation
  - Updated LLM.txt with new architecture details

### Changed

- **refactor(storage)**: File organization now uses domain-based structure
  - Files organized by business domain instead of flat `tables/` directory
  - Each domain contains its definition and all associated ODCS/ODPS/CADS files
  - Legacy `tables/` directory still supported for backward compatibility

## [1.3.0] - 2026-01-03

### Added

- **feat(odcs)**: Complete ODCS v3.1.0 and ODCL v1.2.1 field preservation
  - Preserve `description`, `quality` arrays (with nested structures), and `$ref` references during import/export
  - Full schema coverage for ODCS v3.1.0 and ODCL v1.2.1 formats
  - Round-trip import/export preserves all fields

- **feat(tags)**: Enhanced tag support with Simple, Pair, and List formats
  - `Tag` enum supporting three formats: `Simple(String)`, `Pair(String, String)`, `List(String, Vec<String>)`
  - Auto-detection parsing with graceful degradation for malformed tags
  - Backward compatible with existing `Vec<String>` tags
  - Tag support in ODCS, ODCL, JSON Schema, AVRO, and Protobuf importers/exporters

- **feat(cads)**: Full CADS v1.0 support
  - CADS importer (`CADSImporter`) for AI/ML models, applications, pipelines
  - CADS exporter (`CADSExporter`) for serializing CADS assets
  - Support for all CADS asset kinds: AIModel, MLPipeline, Application, ETLPipeline, SourceSystem, DestinationSystem
  - Full metadata support: runtime, SLA, pricing, team, risk, compliance, validation profiles

- **feat(odps)**: Full ODPS (Open Data Product Standard) support
  - ODPS importer (`ODPSImporter`) for data products
  - ODPS exporter (`ODPSExporter`) for serializing data products
  - Contract ID validation against ODCS Tables
  - Support for input/output ports, support, team, and custom properties

- **feat(domain)**: Business Domain schema support
  - `Domain` struct for organizing systems, CADS nodes, and ODCS nodes
  - `System` struct representing physical infrastructure with DataFlow metadata
  - `CADSNode` and `ODCSNode` structs with shared reference support
  - ERD-style `SystemConnection` for system-to-system relationships
  - Crowsfeet notation `NodeConnection` for ODCS node relationships
  - Domain operations in `DataModel`: `add_domain()`, `add_system_to_domain()`, `add_cads_node_to_domain()`, `add_odcs_node_to_domain()`, etc.
  - Domain YAML import/export: `Domain::from_yaml()`, `Domain::to_yaml()`

- **feat(converter)**: Universal format converter
  - `convert_to_odcs()` function converts any format to ODCS v3.1.0
  - Auto-detection for SQL, ODCS, ODCL, JSON Schema, AVRO, Protobuf, CADS, ODPS, Domain formats
  - Format-specific conversion logic with informative error messages

- **feat(migration)**: DataFlow to Domain migration utility
  - `migrate_dataflow_to_domain()` converts DataFlow YAML to Domain schema
  - Preserves all metadata during migration
  - Maps DataFlow nodes to Systems and relationships to SystemConnections

- **feat(wasm)**: Comprehensive WASM bindings for new features
  - CADS import/export: `importFromCads()`, `exportToCads()`
  - ODPS import/export: `importFromOdps()`, `exportToOdps()`
  - Domain operations: `createDomain()`, `importFromDomain()`, `exportToDomain()`, `migrateDataflowToDomain()`
  - Domain management: `addSystemToDomain()`, `addCadsNodeToDomain()`, `addOdcsNodeToDomain()`
  - Tag operations: `parseTag()`, `serializeTag()`
  - Universal converter: `convertToOdcs()`

### Changed

- **refactor(dataflow)**: DataFlow format removed and replaced with Domain schema
  - DataFlow import/export modules removed
  - DataFlow WASM bindings removed
  - Migration utility provided for existing DataFlow files

- **refactor(tags)**: Enhanced tag support across all formats
  - `Table.tags` and `Relationship.tags` now use `Vec<Tag>` instead of `Vec<String>`
  - All importers parse enhanced tags using `Tag::from_str()`
  - All exporters serialize tags using `Tag::to_string()`

### Documentation

- Added comprehensive Schema Overview Guide (`docs/SCHEMA_OVERVIEW.md`)
- Updated README.md with Domain schema and new format support
- Updated LLM.txt with new modules and architecture details

## [1.2.0] - 2026-01-27

### Added

- **feat(metadata)**: Enhanced Data Flow node and relationship metadata support
  - Add `owner`, `sla`, `contact_details`, `infrastructure_type`, and `notes` fields to `Table` struct (for Data Flow nodes)
  - Add `owner`, `sla`, `contact_details`, `infrastructure_type` fields to `Relationship` struct (for Data Flow relationships)
  - Add `SlaProperty` struct with ODCS-inspired structure (property, value, unit, element, driver, description, scheduler, schedule)
  - Add `ContactDetails` struct with email, phone, name, role, and other fields
  - Add `InfrastructureType` enum with 70+ infrastructure types covering major cloud databases, container platforms, data warehouses, message queues, and BI/analytics tools
  - Add filter methods to `DataModel`: `filter_nodes_by_owner()`, `filter_relationships_by_owner()`, `filter_nodes_by_infrastructure_type()`, `filter_relationships_by_infrastructure_type()`, `filter_by_tags()`

### Changed

- **refactor(metadata)**: All new metadata fields are optional and maintain backward compatibility
- **refactor(filter)**: Updated `filter_by_tags()` to return tuple `(Vec<&Table>, Vec<&Relationship>)` for both nodes and relationships

## [1.1.0] - 2026-01-02

### Added

- **feat(wasm)**: Add comprehensive WASM bindings for validation, export, and model management
  - Add WASM bindings for input validation functions (`validate_table_name`, `validate_column_name`, `validate_uuid`, `validate_data_type`, `validate_description`)
  - Add WASM bindings for sanitization functions (`sanitize_sql_identifier`, `sanitize_description`)
  - Add WASM bindings for table validation (`detect_naming_conflicts`, `validate_pattern_exclusivity`)
  - Add WASM bindings for relationship validation (`check_circular_dependency`, `validate_no_self_reference`)
  - Add WASM binding for PNG export (`export_to_png`) - feature-gated behind `png-export`
  - Add async WASM bindings for model loading/saving (`load_model`, `save_model`) using browser storage backend
  - Add `Serialize`/`Deserialize` to validation result types for WASM interop
  - Add `Serialize`/`Deserialize` to `ModelLoadResult` and related types for WASM interop

### Changed

- **refactor(wasm)**: Extend WASM module with 14 new binding functions
  - All validation functions return JSON strings with structured results
  - Model loading/saving functions use `js_sys::Promise` for async operations
  - Error handling converts Rust errors to JavaScript-compatible errors

## [1.0.2] - 2025-12-31

### Fixed

- **fix(publish)**: Fix publish workflow and Cargo.toml metadata
  - Add required `description` field to Cargo.toml for crates.io publishing
  - Update publish workflow to use `cargo login` with environment variables instead of deprecated `--token` flag
  - Fix cargo publish warnings

## [1.0.1] - 2025-12-31

### Changed

- **refactor(ci)**: Update GitHub workflows
  - CI workflow now only runs on PR open/update events (not on main branch)
  - Removed publish job from CI workflow
  - Build/test only on Linux by default (macOS/Windows optional)
  - Publish workflow runs on manual trigger or merge to main
  - Publish workflow creates git tag from Cargo.toml version
  - Publish workflow verifies version matches CHANGELOG.md

- **fix(auth)**: Use derive Default for AuthMode enum
  - Replace manual Default impl with derive macro and #[default] attribute
  - Fixes clippy::derivable_impls warning

## [1.0.0] - 2024-12-19

### Added

#### Features

- **feat(models)**: Add comprehensive data model structures for tables, columns, and relationships
  - Add `Table` struct with support for database types, medallion layers, SCD patterns, and Data Vault classifications
  - Add `Column` struct with support for foreign keys, composite keys, constraints, and enum values
  - Add `Relationship` struct with cardinality, relationship types, foreign key details, and ETL job metadata
  - Add `DataModel` struct for managing collections of tables and relationships
  - Add `Position` struct for 2D canvas positioning
  - Add `CrossDomainConfig` for cross-domain table and relationship references

- **feat(import)**: Add comprehensive import functionality for multiple formats
  - Add `ODCSImporter` for importing ODCS v3.1.0 YAML format (legacy ODCL formats supported for import)
  - Add `SQLImporter` for parsing CREATE TABLE statements (PostgreSQL, MySQL, SQL Server dialects)
  - Add `JSONSchemaImporter` for importing JSON Schema definitions with nested object support
  - Add `AvroImporter` for importing AVRO schema files
  - Add `ProtobufImporter` for runtime parsing of proto3 syntax files
  - Add support for deeply nested structures and references in all importers
  - Add validation for imported table and column names

- **feat(export)**: Add comprehensive export functionality for multiple formats
  - Add `ODCSExporter` for exporting to ODCS v3.1.0 YAML format
  - Add `SQLExporter` for generating CREATE TABLE statements (PostgreSQL, MySQL, SQL Server dialects)
  - Add `JSONSchemaExporter` for exporting JSON Schema definitions
  - Add `AvroExporter` for exporting AVRO schema files
  - Add `ProtobufExporter` for exporting Protocol Buffer message definitions
  - Add `ODCLExporter` for legacy ODCL format support (deprecated)
  - Add `PNGExporter` for exporting ER diagrams as PNG images (feature-gated)

- **feat(storage)**: Add storage backend abstraction layer
  - Add `StorageBackend` trait for platform-agnostic file operations
  - Add `FileSystemStorageBackend` for native file system operations (feature-gated)
  - Add `BrowserStorageBackend` for WASM browser storage APIs (feature-gated)
  - Add `ApiStorageBackend` for HTTP API-based storage (feature-gated)
  - Add path traversal protection in all storage backends
  - Add domain slug validation for API backend

- **feat(model)**: Add model loading and saving functionality
  - Add `ModelLoader` for loading tables and relationships from storage backends
  - Add `ModelSaver` for saving tables and relationships to storage backends
  - Add `ApiModelLoader` for loading models from API endpoints
  - Add support for orphaned relationship detection
  - Add YAML file path tracking for tables

- **feat(validation)**: Add comprehensive validation functionality
  - Add `TableValidator` for detecting naming conflicts and pattern exclusivity violations
  - Add `RelationshipValidator` for detecting circular dependencies and self-references
  - Add input validation functions for table names, column names, data types, and UUIDs
  - Add SQL identifier sanitization for multiple database dialects
  - Add reserved word checking for SQL identifiers

- **feat(git)**: Add Git integration support
  - Add `GitService` for Git repository operations (feature-gated)
  - Add support for staging files, committing changes, and checking repository status
  - Add error handling for Git operations

- **feat(workspace)**: Add workspace management types
  - Add `WorkspaceInfo` for workspace metadata
  - Add `ProfileInfo` for user profile management
  - Add request/response types for workspace operations

- **feat(auth)**: Add authentication types
  - Add `AuthMode` enum for different authentication modes (Web, Local, Online)
  - Add `AuthState` for tracking authentication status
  - Add `GitHubEmail` for GitHub OAuth email selection
  - Add OAuth request/response types

- **feat(cross-domain)**: Add cross-domain table and relationship sharing
  - Add `CrossDomainTableRef` for referencing tables from other domains
  - Add `CrossDomainRelationshipRef` for referencing relationships from other domains
  - Add `CrossDomainConfig` for managing imported references

#### Documentation

- **docs**: Add comprehensive documentation coverage (~85%+)
  - Add module-level documentation for all public modules
  - Add function-level documentation with examples for all public functions
  - Add field-level documentation for all model structs
  - Add usage examples in documentation comments
  - Add security considerations documentation

#### Tests

- **test**: Add comprehensive test suite (~65+ tests)
  - Add unit tests for all validation functions
  - Add unit tests for model loading and saving
  - Add unit tests for Git service operations
  - Add unit tests for storage backend operations
  - Add unit tests for import/export functionality
  - Add unit tests for workspace and auth types
  - Add integration tests for ODCS import/export roundtrips
  - Add tests for deeply nested structures and references
  - Add tests for edge cases and error conditions

### Changed

- **refactor(import/odcs)**: Improve ODCS importer to fully parse table structure
  - Replace basic field parsing with full `ODCSImporter::parse_table()` usage
  - Implement complete STRUCT expansion with recursive helper function
  - Improve handling of complex union types in AVRO import

- **refactor(import/protobuf)**: Enhance Protobuf parser for complete proto3 support
  - Add support for nested messages with dot notation
  - Add support for repeated and optional fields
  - Add support for enum definitions
  - Add support for oneof and map types

- **refactor(export/odcs)**: Standardize on ODCS v3.1.0 format
  - Remove support for legacy ODCS formats in export (import still supported)
  - Ensure all exports conform to ODCS v3.1.0 schema

- **refactor(models)**: Improve model structure and methods
  - Change table ID generation from deterministic UUIDv5 to random UUIDv4
  - Add `get_unique_key()` method to `Table` for conflict detection
  - Add helper methods to `DataModel` for table and relationship lookup

- **refactor(validation)**: Improve validation error messages
  - Add detailed error messages with context
  - Add support for multiple validation errors per operation

### Fixed

- **fix(import/odcs)**: Fix comments to clarify ODCS single-table-per-file format
  - Update comments to reflect that ODCS only specifies one table per file
  - Fix STRUCT expansion to handle deeply nested structures correctly

- **fix(import/avro)**: Improve union type handling
  - Use first non-null type from union instead of defaulting to string

- **fix(model/loader)**: Fix table loading to use full ODCS parser
  - Replace basic field parsing with complete table structure parsing
  - Ensure all columns and metadata are properly loaded

- **fix(storage)**: Fix path traversal vulnerabilities
  - Add comprehensive path traversal protection in all storage backends
  - Add validation for domain slugs in API backend

- **fix(export)**: Fix clippy warnings and improve code quality
  - Replace `or_insert_with(Vec::new)` with `or_default()`
  - Fix redundant closure warnings
  - Fix unused enumerate index warnings

### Security

- **security(storage)**: Add path traversal protection
  - Block all path traversal attempts (../, ..\, etc.)
  - Validate domain slugs to prevent injection attacks
  - Sanitize SQL identifiers to prevent SQL injection

- **security(validation)**: Add input validation
  - Validate all user inputs before processing
  - Prevent SQL injection via malicious table/column names
  - Enforce maximum length limits on identifiers

### Performance

- **perf(validation)**: Optimize relationship validation
  - Use efficient graph algorithms for cycle detection
  - Cache validation results where appropriate

### Build

- **build**: Add feature flags for optional functionality
  - Add `api-backend` feature for HTTP API support
  - Add `native-fs` feature for native file system operations
  - Add `wasm` feature for WebAssembly support
  - Add `png-export` feature for PNG diagram export
  - Add `git` feature for Git integration
  - Add `databricks-dialect` feature for Databricks SQL support

### Removed

- **remove**: Remove legacy ODCS format support from export
  - ODCL exporter remains for import compatibility only
  - All exports now use ODCS v3.1.0 format

### Deprecated

- **deprecate(export/odcl)**: Mark ODCL exporter as deprecated
  - ODCL is End-of-Life standard
  - Use ODCS v3.1.0 format for new exports
  - ODCL import still supported for backward compatibility

---

## Version History

### [1.0.0] - 2024-12-19

Initial stable release with comprehensive feature set:

- Complete data modeling SDK with support for tables, columns, and relationships
- Multi-format import/export (ODCS, SQL, JSON Schema, AVRO, Protobuf)
- Storage backend abstraction for multiple platforms
- Comprehensive validation and security features
- Cross-domain sharing capabilities
- Git integration support
- Extensive test coverage (~65+ tests)
- Comprehensive documentation (~85%+ coverage)

---

## Types of Changes

- **Added** for new features
- **Changed** for changes in existing functionality
- **Deprecated** for soon-to-be removed features
- **Removed** for now removed features
- **Fixed** for any bug fixes
- **Security** for vulnerability fixes

[1.0.0]: https://github.com/your-org/data-modelling-sdk/releases/tag/v1.0.0