Force minimum of one header row; Add to docs

[brynpickering]

Fixes #573

Based on the decision in #573, I have forced the need for a header row by setting header=[0] if a user does not provide columns data in their data source config.

It can lead to a difficult to understand failure. When a user forgets a header row, data ends up as the column / index level names. This gets picked up later when Calliope tries to match index/column levels to what has been defined in the config. That seems to be the only place we can catch it, but it leads to an error like this:

calliope.exceptions.ModelError: (data_sources, ds_name) | Trying to set names for index but names in the file do no match names provided | in file: ['bar1', 'bar2'] | defined: ['test_row1', 'test_row2'].

Maybe just adding to that error with something like "Make sure you have a header row in your CSV file"?

Summary of changes in this pull request

• Force header=[0] as a minimum
• Added to docs with a bit about how to structure the CSV
• Updated tests

Reviewer checklist

• Test(s) added to cover contribution
• Documentation updated
• Changelog updated
• Coverage maintained or improved

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.96%. Comparing base (872978d) to head (65badad).
Report is 13 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #596      +/-   ##
==========================================
+ Coverage   95.85%   95.96%   +0.10%     
==========================================
  Files          24       24              
  Lines        3619     3666      +47     
  Branches      788      735      -53     
==========================================
+ Hits         3469     3518      +49     
+ Misses         86       84       -2     
  Partials       64       64              

Files	Coverage Δ
src/calliope/preprocess/data_sources.py	100.00% <100.00%> (ø)