Clear schema cache between JSON files

[joelittlejohn]

When using JSON or YAML files (not schemas, but example data) the schemas will not have a proper file path in their ID, as there is no actual schema file. All schema IDs will be fragments only (relative references within the document). Once generation is complete for a file, these schemas should be thrown away and not reused.

Closes #1427

[unkish]

Would it make sense to take some of the changes from https://github.com/joelittlejohn/jsonschema2pojo/pull/1464 in particular:

• remove YamlRefIT as it is empty (dead code)
• rewrite SelftRefIT::nestedSelfRefsInStringContentWithoutParentFile such that it would act as an IT rather than unit test (schemaRule.generateAndCompile(...) vs. new SchemaMapper().generate(...))
• add || ruleFactory.getGenerationConfig().getSourceType() == SourceType.YAML to https://github.com/joelittlejohn/jsonschema2pojo/blob/65e1fdbcf4a4fd7298ea7ede0dc9c3e1651e9035/jsonschema2pojo-core/src/main/java/org/jsonschema2pojo/SchemaMapper.java#L127

Or separate PR would be better ?

[joelittlejohn]

• [x] remove YamlRefIT as it is empty (dead code)
• [ ] rewrite SelftRefIT::nestedSelfRefsInStringContentWithoutParentFile such that it would act as an IT rather than unit test (schemaRule.generateAndCompile(...) vs. new SchemaMapper().generate(...))
• [x] add || ruleFactory.getGenerationConfig().getSourceType() == SourceType.YAML

I've ticked off two of the above as part of this change.