Implement a Weaviate-powered semantic search system with schema design, data import, GraphQL queries, and multi-tenancy.
You are a search infrastructure engineer. Build a production Weaviate semantic search system. ## Configuration - **Weaviate Version**: [VERSION e.g. 1.27] - **Deployment**: [DEPLOY e.g. Docker / Weaviate Cloud / Kubernetes] - **Vectorizer**: [VECTORIZER e.g. text2vec-openai / text2vec-cohere / text2vec-transformers] - **Data Classes**: [CLASSES e.g. Article, Author, Category] - **Language**: [LANG e.g. Python / TypeScript / Go] ## Implementation ### 1. Schema Design (`schema/`) - Define classes for [CLASSES] with: - Properties with data types (text, int, number, boolean, date, geoCoordinates) - Cross-references between classes - Vectorizer configuration per class - Module configuration (generative, qna) - Multi-tenancy setup for SaaS use cases - Schema migration scripts for version updates ### 2. Data Import (`import/`) - Batch import with configurable batch_size and dynamic throttling - Error handling with dead letter queue - Progress tracking with ETA - Data validation before import - Cross-reference resolution during import - Incremental import (upsert semantics) ### 3. Search API (`search/`) - **Semantic search**: nearText, nearVector queries - **Hybrid search**: BM25 + vector with fusion algorithm (rankedFusion / relativeScoreFusion) - **Generative search**: RAG with inline generation - **Filters**: where filters with AND/OR/operand combinations - **Aggregation**: groupBy, count, sum, average over properties - **GraphQL query builder**: type-safe query construction in [LANG] ### 4. Multi-Tenancy (`tenancy/`) - Tenant CRUD operations - Tenant-specific data isolation - Bulk tenant operations - Tenant activity monitoring - Tenant data export/import ### 5. Monitoring (`monitoring/`) - Object count and vector dimension stats - Query latency tracking - Shard health monitoring - Backup and restore automation - Grafana dashboard templates Provide complete client library, Docker setup, and integration examples.
Free to copy and use. Compatible with Claude 4 Opus, GPT-5, Gemini 2.0 Flash, Llama 4 Maverick.
Deploy Weaviate via Docker or cloud. Define your schema classes and properties. Import data in batches, then test queries using the GraphQL explorer before building your search API.
Initial release
Sign in and download this prompt to leave a review.