fix(codegen): strip empty self-closing HTML tags from documentation traits

[TrevorBurnham]

Fixes #7477

Problem

Several upstream AWS service models contain empty self-closing HTML tags like <p/> in their smithy.api#documentation trait values. These tags pass through code generation unmodified and render as literal text in:

• Generated README.md files (e.g. clients/client-lambda/README.md)
• JSDoc/TSDoc comments in generated TypeScript source (e.g. LambdaClient.ts)
• The published API reference at docs.aws.amazon.com

The Lambda client page is the one called out in the issue, but 30 service models are affected (RDS alone has 89 occurrences of <p/>).

Solution

Adds a new SanitizeDocumentation codegen integration that runs as a preprocessModel step — before any code generation. It walks every shape in the Smithy model, finds DocumentationTrait values containing empty self-closing HTML tags (<p/>, <br/>, <hr/>), and strips them.

Because this operates on the model before codegen, it fixes all downstream outputs at once: READMEs, JSDoc in .ts files, and anything else that consumes the documentation trait. It's also resilient to future model updates that reintroduce these tags.

Changes

• New file: codegen/smithy-aws-typescript-codegen/src/main/java/.../SanitizeDocumentation.java
  • Implements TypeScriptIntegration.preprocessModel
  • Uses a regex to match <p/>, <br/>, <hr/> (with optional whitespace)
  • Replaces affected shapes via ModelTransformer.replaceShapes
• Modified: META-INF/services/software.amazon.smithy.typescript.codegen.integration.TypeScriptIntegration
  • Registers the new integration

---

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.