CHANGELOG.md

Change Log

All notable changes to this project will be documented in this file. See the change log guidelines for information on how to structure the file.

Unreleased

3.3.1

Fixed

• Opening projects that have MS Access database resources
• IndexOutOfBoundsException when calling Groovy helper functions
• Do not add STARTINDEX parameter to non-paginated WFS GetFeature requests
• Loading resources in headless mode from URL when remote server responds with a redirect
• Loading a project in headless mode no longer fails in cases where code lists cannot be imported
• Loading INSPIRE schemas from local resources when online version is not available
• New projects could be saved only as a project archive if the last project loaded was an archive

Added

• Support for saving changes directly to hale connect
• Support for partitioning GML output to multiple files
• Support for table type MATERIALIZED VIEW when importing a PostgreSQL database schema
• Check for remote changes when sharing project to hale connect
• Support for Double columns for the CSV reader
• GML reader parameters ignoreMappingRelevant and suppressParsingGeometry
• Property constraint CodeListAssociation
• Type constraint MappingRelevantIfFeatureType

3.3.0

Added

• Integration with the online collaboration platform hale connect: log in to hale connect, import shared transformation projects and upload projects.
• Spatial Index for instances with geometries that can be queried via the new Groovy geometry helper "spatialIndexQuery"
• Spatial Join transformation function: join instances based on the spatial relation of their geometry properties
• Groovy geometry helper function "boundaryCovers" that can be used to determine if the boundary of a the geometry covers all points of a line
• Use arbitrary SQL queries as a source schema and data source
• Import/Export hale schema definitions as JSON
• DMG image for macOS installation

Changed

• Application title is now "hale studio"
• "Load project from templates..." has been removed in favour of hale connect integration
• Cached schema definition is now used always if loading source fails

Fixed

• Fixed content assistance in RegEx Analysis function
• Fixed resource copying in hale Project Archive writer
• Fixed links on About screen
• Fixed updating a cell when the source or target types are changed to a parent of the original type

3.2.0

Added

• Added support for several Arc-based GML geometry types to be interpolated when read: Arc, ArcString, Circle, CircleByCenterPoint. The interpolation is based on one of two algorithms that can be selected on import.
• Project Validator that validates exported instances based on validator configuration (e.g. rules or schemas) imported into the project
• New transformation function Assign collected values allows the assignment of all values collected by a Groovy transformation function. The new function automatically converts collected values to references if the target property takes references.
• Better usage of available space in Alignment and Mapping views
• Content assistance for project variables in transformation function wizards such as Formatted String, Regex Analysis and Assign
• Request pagination for WFS requests. Users can now choose to activate request pagination for WFS sources.
• IO Provider extensions can now have a configurationContentType to describe the content type of configuration files for this provider
• Total number of imported instances is now shown in progress dialog (if known)
• The InstanceResolver interface has been extended to allow resolving multiple references at once. Implementations can use this to optimize resolving of multiple references

Changed

• The HTML documentation that can be generated for an alignment is now much more performant for large mappings due to lazy loading and rendering
• The Merge function now uses an iterative approach for merging instances which allows for processing more data in a Merge
• Allow using SimpleDateFormat and UUID classes in groovy scripts by default
• When loading data from CSV files the data is now streamed (similar to XML data sources) and not loaded at once into memory
• When a CSV files has more columns than defined in the schema, this is now a warning, not an error

Fixed

• Deselecting in a type selector could lead to an exception
• Removed CRS selection dialog and UI dependency from MS SQL plugin
• Schema selection configuration for JDBC driver is optional
• Fixed wrong tooltip in Mapping view
• Fixed error when loading hale schema definitions in respect to schema elements w/ primitive bindings

3.1.0

Added

• Support reading from and writing to MS SQL databases
• Instance counts are now calculated for condition and index contexts as well
• You can now hide optional properties in the schema explorer
• SKOS format code lists can now be loaded
• Validation based on a Schematron file can now be performed on an encoded XML/GML file
• To ensure topological consistency in respect to interpolated geometries, other geometries may optionally also be moved to the interpolation grid
• GML Encoding: It is now possible to specify a number format for geometry ordinates, e.g. if you want a fixed precision after the decimal point
• During validation in hale also check property values against an assigned code list
• CLI commands can now be marked experimental
• Added extension point for custom validators for the hale pre-encoding validation

Changed

• GML encoding: Automatic conversion of polygon geometries to line geometries when there are no possibilities to encode a surface has been changed to produce a LineString for a Polygon and a MultiLineString for a MultiPolygon
• The contents of XML Alignment files now are sorted where possible, to have a reproducable encoding for the same mapping and a nice diff when used in version control
• The contents of Hale Schema Definition files now are sorted where possible, to have a reproducable encoding for the same schema and a nice diff when used in version control
• When reading GML geometries composite 2D geometries (e.g. CompositeSurface, Surface with multiple patches, CompositeCurve, etc.) are now by default combined to a single geometry if possible

Removed

• File based databases can no longer be loaded via From database - instead use From file

Fixed

• Using a previously as Hale Schema Definition file exported database schema as source schema when loading data from the database now properly supports loading the geometries
• Using the value unpopulated for GML nilReason attributes does not conform to the GML specification. The proposal to use this value has been changed to other:unpopulated to conform with the specification. Also, when encoding GML, unpopulated will be replaced by other:unpopulated where encountered in nilReason attributes to support mappings created in previous versions
• Fixed error in instance partitioning for WFS-T upload when encountering unresolvable references
• Fixed instance partitioning for WFS-T upload producing too many small parts
• Resolving local resources via bundles can no longer yield streams to directories/packages
• Groovy geometry helper functions don't fail for null input
• Exporting project archives for new projects now works as expected
• On export to GeoServer AppSchema via the REST endpoint the provided URL now may end with a slash
• Mapping cells in the mapping view are now ensured to be updated when edited or deleted
• When exporting data to XML files include root element schema in schema location attribute

Deprecated

• Deprecated old style HTML mapping documentation, instead use the new HTML+SVG mapping documentation
• Deprecated INSPIRE 3.0 specific mapping functions (INSPIRE Identifier and Geographical Name), instead map to sub-properties individually

3.0.0

Added

• Users can now configure custom maps based on a Tile URL pattern or a Web Map Service
• Generic command line interface with commands that can be added via an extension point, added commands for documentation generation
• Transformation command line interface: Data included as source for the transformation using the now can be filtered, also arguments can now be passed to the command line interface using an arguments file
• MS Access database reader for .mdb files based on UCanAccess
• Three new priority levels for mapping cells
• Collector helper class for easily collecting information or building indexes in Groovy functions, also added a helper function that will create a default collector in a transformation context
• Internal instance validation now includes validation of local XLink references. Local references that cannot be resolved result in a warning
• hale checks the version of a project that is loaded and displays a warning when the project was created with a newer version of hale than the one currently used
• Support for project variables has been added, that allow customizing the behavior of an alignment. The variables can be defined and stored in the project, and overridden for transformation execution
• SHA-256 Groovy helper function
• Several Groovy helper functions for dealing with and creating geometries
• Users can now define custom transformation functions (using Groovy) that are stored in the alignment
• For XML Schema it is now possible to also map types w/o global element definition (option on schema import)
• Interior Point transformation function that determines a point that us guaranteed to lie inside a geometry
• Support for 1-dimensional arrays as multi-valued properties for JDBC schema and data sources (only tested with PostgreSQL)
• XmlSchemaReader can be configured with specific content for anyType elements (to be able to map them properly; not configurable via the UI)
• WFS-T writers now can be configured for services protected with HTTP Basic Authentication
• New decimal parameter added to identify the float value in CSV file with specified separator

Changed

• New default base map is Stamen Terrain
• The WFS 2.0 Feature Collection writer can now be configured to skip counting features (allowing direct streaming instead of temporary storing all features)
• Mapping cells are now sorted in a specific order when they are saved, to have easier understandable diffs when using .halex projects in version control
• Writing XmlDateTime values when encoding XML now uses a fixed default timezone (UTC) instead of the system time zone
• Type transformations are now executed in order according to their configured priority (this means that for instance a type transformation w/ high priority is guaranteed to be completely executed before a type transformation with normal priority)
• Several improvements to the HTML+SVG mapping documentation
• The winding order of GML geometries is now fixed when encoding a GML file, also the behavior regarding winding order is now configurable
• Groovy restrictions: Access to the QName class is now allowed by default
• Process URI and URL values as Strings when comparing keys in a Join
• Mapping cell explanations can now be available in multiple languages, by default English and German are supported
• Groovy script input variables are no longer converted to Strings to allow handling instances and other kinds of objects
• An encoding can now be specified when loading a schema from a Shapefile
• hale now requires Java 8
• hale is now based on Eclipse Mars (from Eclipse Luna)
• hale libraries can now be used on a non-OSGi environment (this implies a lot of internal changes regarding service provision and discovery)
• Groovy type transformations now may return multiple results with multiple _target closures
• Proxy credentials where the user name includes a backslash (\) are interpreted as NTLM credentials with the user name specifying both domain and user
• XML/GML writers will write nilReason attributes only if the element is actually nil

Removed

• The OpenStreetMap based map provided by MapQuest has been removed as the service is no longer available (see also #64)

Fixed

• Errors when converting the alignment model on saving will no longer result in empty files overriding your previously stored alignment
• Time information for dates no longer is lost when stored in the internal database
• Concatenating a Ring geometry to a LinearRing will no longer result in duplicate coordinates
• Concurrent access to the same source instances could result in exceptions and invalid objects being transformed (see #96)
• Drastically reduced threads created by FinalizableReferenceQueues for internal database handle cleanup
• PostGIS: classify columns with type geometry as geometry columns, even if there is no corresponding entry in the geometry_columns table
• PostGIS: assume lon/lat axis order instead of lat/lon for geographic coordinate reference systems
• Correctly use cursors in JDBC connections (loading tables in batches, not completely)
• Returning muliple instances from Groovy functions now always uses the correct bindings

2.9.4 - 2015-11-01

Changes so far have been documented in the hale help.