Stefan Urbanek
Stefan Urbanek
It would be nice if the highlighting distinguished between regular comments `//` and `/* ... */` and comments containing documentation: `///` and `/** ... */`. The former keeping the form...
Separate the `flask slicer serve` functionality from `cubes.server` into a separate module. Reasons: * allow other server type implementations, such as [protobuffer](https://developers.google.com/protocol-buffers/) or [thrift](https://thrift.apache.org/docs/) kind of server. * make the...
Add/modify documentation about how to create extensions and how to configure them. See #430. (replaces #417)
Potential mutability of model objects is sometimes causing unexpected behavior. All model objects should be immutable, including Cube. All setters should be removed. The problem is Cube creation and linking....
Add type annotations and remove all unnecessary type ambiguity. This is backward incompatible change and will make Cubes depend on Python ≧3.6. Lower versions of python will not be supported....
Add an object `SchemaInspector` that would inspect a database schema and would infer metadata objects: * fact tables based on prefix/suffix/pattern: `discover_fact_tables() → List[str]` * dimension tables based on prefix/suffix/pattern:...
Unittests are incomplete and broken. They are in very bad shape. This is a large task to be done. Tests to be reviewed: * [ ] Metadata * [ ]...
Housekeeping changes to the `Cube` object: * [ ] remove `aggregates_for_measure(name)` * [ ] remove `all_attributes()` * [ ] reconsider use of `attribute(name)` * [ ] try to replace the...
Separate all user-oriented metadata from `ModelObject` into `MetadataObjectDescription` or `MetadataObjectInfo`. Note: It is `Metadata*` not `Model*` because it will describe non-logical objects as well in the future. Reasons for the...
The `Dimension.from_metadata` method is overly complex and might also have type issues. To do: 1. [ ] analyse the function flow 2. [ ] split into smaller parts, especially to...