Add arguments to pass document parsing and chunking options to google_discovery_engine_data_store

[etekin-amfam]

Google's VertexAI Search & Conversation Data Stores have further options to refine how documents are ingested that can significantly improve performance. However, these options need to be passed at creation, and are currently not supported by terraform. As such, we're having to revert to deploying these data stores manually instead of terraform., making things harder to manage. Adding these options to the terraform module so we can deploy data stores via terraform would make maintenance significantly easier and less manual.
For the arguments I'm describing, see https://cloud.google.com/generative-ai-app-builder/docs/parse-chunk-documents#parse-chunk-rag, mainly, a documentProcessingConfig field needs to be specifiable, as shown in https://cloud.google.com/generative-ai-app-builder/docs/parse-chunk-documents#example and https://cloud.google.com/generative-ai-app-builder/docs/parse-chunk-documents#turn-on-chunking

b/330174625

[dariowho]

Hi @rileykarson, I can take this one

[dariowho]

@etekin-amfam the PR for this issue does not include Layout Parsing and Chunking configuration, as those options were not available in either v1 or v1beta when this was implemented.

Those parameters are now available, I opened a separate issue to include them: #18390