Made the entity-sorting keys configurable

Made the entity-sorting keys configurable, and made it possible to blacklist the "_id" property. This is required in order to write tests against external systems that generate nondeterministic "_id" values.

Improved validation

• IS-15992: Added a new validation check that checks that datatypes that has a share pipe includes the collect template
• IS-15992: Allow running validate offline (no dev node required)

Added a '-scheduler-dont-reset-pipes-or-delete-sink-datasets' commandline option

What's Changed

• IS-15927: Added a '-scheduler-dont-reset-pipes-or-delete-sink-datasets' commandline option by @knutj42 in #140. This command-line option should only be used in very special circumstances (like during manual testing and development).

Full Changelog: 2.8.0...2.8.1

Added support for the portal analytics API

Merge pull request #135 from sesam-community/IS-15613

IS-15613: support analytics api to register interactions

Support using .additional_parameters file during connector development

When doing connector development, additional parameters can now be added to the working directory of the connector. They will be uploaded as environment variables on doing sesam upload. See the README for more details.

Remove "permission" from webhook pipes during download/collapse

This release removes "permission" from the webhook pipes when running "sesam download".

Change priority for variables to replace in Sesampy

This release makes sure manifest parameters are replaced before datatype. In case the datatype is part of the manifest parameter, Sesampy had an unintended result in replacing variables. e.g. datatype is "contact" and the manifest parameter is "contacts", the collapsed config had the variable replaced by {{@ datatype @}}s. This update makes sure it is replaced with {{@ manifest_parameter @}}. The outcome will help merge templates.

TODO: It still can cause unintended results in case the manifest parameter is part of the datatype but changing the manifest is easier than datatype. Some validation may be required to check the potential issues.

Write secrets to all systems in a multi-system connector

This release writes the secrets to all systems if the connector uses more than one.

Update-schema fixes

• Don't recurse into objects or arrays when creating the CSV (only one level from the parent)
• Add an empty "description" property to null fields when bootstrapping

Support "object" type with no subschemas in update-schemas

update-schemas: If a property of type "object" is specified in a schema and it has no subschema (i.e. no properties of its own), it will still be included in the generated schema.csv file.