Fix consecutive-acronym handling in generated examples

[pdv]

Summary

Fixes broken imports/references in generated examples when an OpenAPI tag contains consecutive acronyms (e.g. Bits AI SRE).

The library class (BitsAiSreApi) generates correctly because the per-language pipeline goes raw-tag → snake → camel. But the example template re-camel_cased the spaceless form BitsAISRE (set in conftest from the BDD .feature file) which produced BitsAisreApi.

Changes:

• templates/example.j2: use raw_tag|camel_case|upperfirst instead of name|camel_case|upperfirst (5 sites, 6 occurrences).
• .generator/conftest.py: stash the raw tag in context["api_instance"] so the template can use it.

Behavior is unchanged for any tag without consecutive acronyms.

Test plan

• Renamed Bits AI → Bits AI SRE in .generator/schemas/v2/openapi.yaml and ran ./generate.sh.

Links

• Spec PR (originating change): https://github.com/DataDog/datadog-api-spec/pull/5522
• Sibling: Python — Fix consecutive-acronym handling in generated examples and apis/__init__.py datadog-api-client-python#3454
• Sibling: Rust — Fix consecutive-acronym handling in generated examples datadog-api-client-rust#1530