feat: add Cosmos DB NoSQL create-index sample (Java)

[diberry]

Summary

• add the new
  osql-create-index-java\ Maven sample
• use DefaultAzureCredential with Azure Cosmos DB and the Azure OpenAI client
• bulk upsert the shared pre-vectorized hotel dataset and run VectorDistance() queries for DiskANN and QuantizedFlat containers

Validation

• mvn test

[diberry]

Review Score: 15/16 [PASS]

Specialist	Dimension	Score	Notes
Sam	Correctness	2	Java SDK usage correct: CosmosClientBuilder, DefaultAzureCredential, executeBulkOperations(), OpenAI embeddings all properly implemented. pom.xml dependencies valid (azure-cosmos, azure-identity, azure-ai-openai v1.0.0-beta.16).
Scooter	Completeness	1.5	All files present and README is comprehensive. Article covers Prerequisites, Clone, Configure, Build/Run, and Structure. Minor: missing explicit cleanup section (less critical for data-plane only).
Beauregard	Conventions	2	Maven project layout is standard (src/main/java/com/azure/cosmos/createindex/). Package naming follows Azure conventions. Class names and records are appropriately named. Matches other language sample patterns.
Statler	Coverage	2	Excellent error handling: CosmosException caught with clear message, Config.validate() checks all required env vars, validateFieldName() regex prevents SQL injection, data file existence verified, VECTOR_ALGORITHM and container names validated.
Architect	Coherence	2	Code structure matches Python/TS pattern: Config handles loading/validation, App orchestrates workflow, DataPlane contains client factories and operations. Article aligns with code structure and explains each component.
Razor	Conciseness	1.5	No dead code or unused variables. SDK usage is efficient (contentResponseOnWriteEnabled=false, executeBulkOperations). Records for output types (IngestionSummary, QueryResult, QuerySummary) add clarity but could be more minimal.
Mirror	Consistency	2	Env vars match standard across samples (AZURE_COSMOSDB_, AZURE_OPENAI_, VECTOR_ALGORITHM, DATA_FILE_WITH_VECTORS). Output format consistent with other language samples. Correct containers (hotels_diskann, hotels_quantizedflat) and relative data path.
Lens	Clarity	2	README instructions are clear with step-by-step guidance and examples. Code comments minimal but sufficient. Error messages are actionable (e.g., 'Invalid embedding field name', missing env var details). Article well-structured with clear sections.
Total		15/16	PASS (≥13, no zeros) ✅ All critical criteria met: data-plane only, DefaultAzureCredential passed directly, field validation before SQL interpolation, VectorDistance() with field interpolation, executeBulkOperations() used, 'Azure OpenAI client' terminology, correct containers, shared data file, standard env vars, Tier 1 article structure.