Simple data validation + error messages

[ivansg44]

Right now, there is no data validation. Any errors when uploading data due to formatting issues are caught, and shown to the user in raw form.

We do not know exactly which data standards we want to set yet, but we should have some simple data validation, and deliver error messages for some invalidaties in plain english:

• Config file:
  • Check if values in config file match format
    • Type checks e.g., int, strings, lists
    • Look into JSON validation tools
  • Check for reference errors in config file
    • Is a referenced attribute actually in the tabular dataset?
• Tabular dataset:
  • Check if values in tabular dataset match format in config file
    • e.g., config file specifies attribute by name, but the values in the attribute column of tabular dataset should be int, string, etc.
• Matrix:
  • Check if all headers in tabular also in matrix
  • Ensure numeric matrix