Make PxWebApi work with Px.Utils

[likp]

The PX file reading portions should use Px.Utils for faster and less resource intensive data retrieval.

The goal is to have a new PX file parser implementation that is faster then the one we have today.

Minimum requirements

• Elimination should work
• Aggregations should work
• En need to set up testing to verify the different implementations.
• Setting to select which implementation to use

Nice to have

• Cell notes should work
• Break it out to its own assembly

[likp]

Current state
The PxUtils version of the PxModelBuilder, PxBuilder should have the same functionality as the current version, with the exception that it only supports loading groupings from a specific folder.

Anybody can try it out by setting the DataSource.Px.UsePxUtilsParser to true.

Known bugs

• ParseStringToList can not handle keyword values that span over multiple rows correctly. See CONTACTS for TAB005

Next steps
We still need to add some for testing that compare the new builder with the old one before we can include it into the upcoming releases of the API.

We should have to investigate the need to break it out to an assembly of its own and also make sure that we do not get any circular dependencies.