Datagrowth (postgres)

A version to move data from MySQL to Postgres. Introduces breaking changes.

Datascope (final draft)

This version adds a database migration and updates vital dependencies. 0.10.x are the last versions to support the draft API version. The 1.x versions will support a much stricter separation between projects.

Datagrowth (module)

A release that moves code shared between projects into a new module that will split off from the repo in a future release

• Adds logs under project directory
• Abandons Conda
• Growth commands can log to "datagrowth.command" to separate them from task processing output.
• Introduces a data dir that can be set on a machine basis. To share data between machines.
• Adds a QuerysetProcessor as a performative alternative for output processors
• Moves Resource, HttpResource and ShellResource into datagrowth
• Adds file deletion handler that can be connected to Resources
• Adds Collective.to_disk to easily use data in notebooks today
• Refactors ExtractProcessor to share code between HTML and XML
• Adds ShellResourceProcessor to execute shells on task servers
• Adds Tika as the first ShellResource
• Adds ibatch and datetime formatting as tools to datagrowth
• Moves all configuration to datagrowth and improves the flow of registering configuration defaults
• Migrates ImageDownload into separate models per app
• Adds TopicDetector and EntityDetector
• Fixes some problems with Wikipedia, but disables the feeds on production for now due to performance issues
• Migrates files to structure that scales well

Datagrowth

A version where stable code is moved into a directory that may grow into a separate package to be used inside other repos.

Stable

Provides a stable release for a number of upcoming demos

Future Fashion

This release includes:

• Adding _id properties to all content
• A way to send file data with HttpResource through FILE_DATA_KEYS attribute
• MicroServiceResource to use local machine learning model services
• A ShellResource to get data from shell commands
• A refactor of the ManifestResource
• Code to work with color similarity
• Renaming of future_fashion models to something more extendable
• Buggy versions of FileSorter, FileBalancer and SemanticDirectoryScan
• AWS as a release platform
• Basic clothing inventory
• Backend for the Paper Doll app

Discourse Scope

This release contains a working Discourse Scope version with reasonable performance.

Discourse Scope (pre-release 2)

Introduces commands to export and import data.
Fixes various problems with the pre-release

Discourse Scope (pre-release)

Most notable changes:

• drops support for 3.4

New features:

• completes the DiscourseSearchCommunity
• introduces the NumericFeaturesFrame
• adds the OnlineDiscourseRankProcessor
• adds a way to scope a Community/Dataset with a POST request on service
• adds setup_utrecht application
• introduces the UniformImagesCommunity
• introduces FileSorter, FileBalancer, SemanticDirectoryScan
• adds a FileSorter for ImageDownloads
• adds fill_from_directory for ImageGrid
• adds Wizenoze resources
• adds a split method to Collective/Collection
• introduces multiple ways to get data for resource request

Major changes and migrations:

• migrates many sources and communities to the correct apps
• changes all Manifestation uri's to have a hash that encodes all configuration (delete all Manifestations)
• splits Community/Dataset configuration into "growth" and "scope" configs
• revamps RankProcessor and ResourceProcessor
• renames topic_research to wiki_scope
• makes the cr parameter for GoogleImages optional
• better documentation
• better Locafora CSV
• parse_datetime_string now returns datetime(0) on error

Deprecations:

• deprecates Collective/Collection indexes
• deprecates public configs
• deprecates OrganismIterator
• deprecates generate_config

Bug fixes:

• prevents crashes when content type is unknown

Order endpoint

Adds a way to make orders