Automate checks of "code to label" mappings against data dictionary

[evansd]

There are various places in the TPP backend where we map internal codes to meaningful labels. For example:

ehrql/ehrql/backends/tpp.py

Lines 234 to 250 in dea0c83

	CASE Status
	WHEN 0 THEN 'Booked' COLLATE Latin1_General_CI_AS
	WHEN 1 THEN 'Arrived' COLLATE Latin1_General_CI_AS
	WHEN 2 THEN 'Did Not Attend' COLLATE Latin1_General_CI_AS
	WHEN 3 THEN 'In Progress' COLLATE Latin1_General_CI_AS
	WHEN 4 THEN 'Finished' COLLATE Latin1_General_CI_AS
	WHEN 5 THEN 'Requested' COLLATE Latin1_General_CI_AS
	WHEN 6 THEN 'Blocked' COLLATE Latin1_General_CI_AS
	WHEN 8 THEN 'Visit' COLLATE Latin1_General_CI_AS
	WHEN 9 THEN 'Waiting' COLLATE Latin1_General_CI_AS
	WHEN 10 THEN 'Cancelled by Patient' COLLATE Latin1_General_CI_AS
	WHEN 11 THEN 'Cancelled by Unit' COLLATE Latin1_General_CI_AS
	WHEN 12 THEN 'Cancelled by Other Service' COLLATE Latin1_General_CI_AS
	WHEN 14 THEN 'No Access Visit' COLLATE Latin1_General_CI_AS
	WHEN 15 THEN 'Cancelled Due To Death' COLLATE Latin1_General_CI_AS
	WHEN 16 THEN 'Patient Walked Out' COLLATE Latin1_General_CI_AS
	END AS status

Following #2003 we now have a copy of the data dictionary included in the test suite. We should come up with some scheme for either automatically generating, or at least automatically checking, these CASE expressions against the contents of the dictionary.