Redesign CQL to SQL generation

if field is 'id' //special case
  validate UUID
  right truncation (= range) is allowed for == and <>
  == <> < <= > >= are allowed if comparing to a valid UUID
if field in 'fullTextIndex' section //
  use ts_vector (right now mod-inv-storage it is used for title, languages, instanceTypeId)
else //this branch also includes handling of the JSON schema for the record
  if relation is <=, <,  >=, > then use < .. (same OP) + unaccent/lower function
  if relation is = then use ~ for a regex match + unaccent/lower function 1)
  if relation is == or <> then use LIKE / NOT LIKE  match +unaccent/lower function 1)

if relation is = (alias adj) or <> #= behaves the same as == for all index types but fulltext
  if field in 'fullTextIndex' section
    use ts_vector, if right truncation use :*
  else if field in 'ginIndex' section
    use LIKE + unaccent/lower, if truncation use "%"
  else if field in 'index' section
    use exact SQL match, no truncation allowed (it will result in linear scan)
  else fail
if relation == #behaves the same as = but does not work with fulltext index
  if field in 'ginIndex' section
    use LIKE + unaccent/lower, if truncation use "%"
  else if field in 'index' section
    use exact SQL match, no truncation allowed (results in linear scan)
  else fail
if relation < > <= >=
  if field in 'index'
    use SQL equivalents, no truncation allowed (results in linear scan)
  else fail

if relation is = (alias adj) or <> #= behaves the same as == for all index types but fulltext
  if field in 'fullTextIndex' section
    use ts_vector, if right truncation use :*
  else if field in 'ginIndex' section
    use LIKE + unaccent/lower, if truncation use "%"
  else if field in 'index' section
    use exact SQL match, no truncation allowed (it will result in linear scan)
  else fail
if relation == #behaves the same as = but does not work with fulltext index
  if field in 'ginIndex' section
    use LIKE + unaccent/lower, if truncation use "%"
  else if field in 'index' section
    use exact SQL match, no truncation allowed (results in linear scan)
  else fail
if relation < > <= >=
  if field in 'index'
    use SQL equivalents, no truncation allowed (results in linear scan)
  else fail

if relation is =
  if (modifier format number)
    use exact SQL match, warn if no btree index specified
  if (modifiers are ignoreAccents and ignoreCase) #which is the default value of those modifiers
    use ts_vector (fulltext), warn if no index specified
  else (none of the above)
     use LIKE + unaccent/lower, if truncation use "%", warn if no GIN index specified 
if relation is adj, all OR any
   use ts_vector (fulltext), warn if no index specified
if relation == OR <>
  if (modifers format string)
     use LIKE + unaccent/lower, if truncation use "%", warn if no GIN index specified 
  if (none of the above)
    use exact SQL match, warn if no btree index specified
if relation < > <= >=
  use exact SQL match, warn if no btree index specified

WHERE lower(f_unaccent(instance.jsonb->>'contributors')) ~ lower(f_unaccent('(^|[[:punct:]]|[[:space:]]|(?=[[:punct:]]|[[:space:]]))"name":([[:punct:]]|[[:space:]])+"water.*"($|[[:punct:]]|[[:space:]]|(?<=[[:punct:]]|[[:space:]]))'))