alicloud-ai-search-opensearch

What This Skill Does

The alicloud-ai-search-opensearch skill enables the OpenClaw agent to interface directly with Alibaba Cloud OpenSearch Vector Search Edition. By utilizing the alibabacloud-ha3engine Python SDK, the agent can programmatically push documents, perform complex HA (High Availability) queries, and execute SQL-style searches against your indexed data. This skill is engineered to streamline RAG (Retrieval-Augmented Generation) pipelines, allowing the agent to perform vector-based semantic retrieval or structured SQL filtering seamlessly within its workspace.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/cinience/alicloud-ai-search-opensearch

Ensure you have a configured Python virtual environment, as the SDK requires specific dependencies. After installation, define the mandatory environment variables in your configuration: OPENSEARCH_ENDPOINT, OPENSEARCH_INSTANCE_ID, OPENSEARCH_USERNAME, OPENSEARCH_PASSWORD, and OPENSEARCH_DATASOURCE. These credentials enable the agent to authenticate and interact with your specific OpenSearch instance.

Use Cases

• RAG Pipeline Orchestration: Automatically index new documentation or internal knowledge base entries so the agent can provide context-aware responses.
• Vector Retrieval: Perform high-speed semantic searches to find relevant information based on embeddings for complex queries.
• Structured Data Analysis: Execute SQL queries on OpenSearch indices to extract insights from structured document fields, useful for log analysis or metadata reporting.
• Content Management: Handle real-time updates by pushing additions or deletions to your searchable indices as the project evolves.

Example Prompts

1. "Index the current file content into the OpenSearch vector store using the 'general' datasource."
2. "Search the knowledge base for documents matching 'API authentication' and return the top 5 results."
3. "Run a SQL query to select all document IDs from the 'tech-docs' index where the category is 'security'."

Tips & Limitations

• Large Queries: If your search criteria exceed 30KB, please switch to the RESTful search API, as the SDK is optimized for standard query execution.
• Primary Key Management: Always specify your unique primary key field (defaulting to 'id') when pushing documents to avoid indexing conflicts.
• Performance: HA queries are significantly faster for vector-keyword hybrid retrieval; prefer these over SQL for latency-sensitive applications.
• Security: Never hardcode credentials in scripts; always utilize the environment variables defined in the prerequisites to maintain a secure posture.