feat: bedrock model reuse for OSS models

[mujtaba1747]

Description of changes:
Adds resource reuse support for open-source (OSS) models imported into Amazon Bedrock via ModelBuilder. When reuse_resources=True, the SDK now checks for existing imported models or in-progress import jobs before starting a new import, avoiding redundant work and reducing
import times.

Changes

• Reuse logic in bedrock_model_builder.py: Before importing an OSS model, the deploy flow now:
  a. Resolves a stable source identifier from the model package ARN or S3 artifact URI
  b. Searches for an already-imported model matching that source (via resource tags)
  c. Falls back to finding an in-progress import job and waiting for its completion
  d. Only starts a fresh import if no existing resource is found
• New model_reuse.py module: Provides find_existing_imported_model() and find_existing_model_import_job() — paginated Bedrock API lookups that match resources by a sagemaker.amazonaws.com/model-source tag
• Idempotent HF model path resolution: Guards against double-appending checkpoints/hf_merged/ when the config is already at the base S3 path
• Source tagging: Automatically tags imported models and import jobs with their source identifier for future reuse lookups

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