Overview

AlphaEar Search unifies web search and local context search for finance. It retrieves general finance information from the web using Jina, DuckDuckGo, or Baidu, and fetches finance data from a local document store via RAG. This dual capability lets you compare public sources with private docs quickly.

How This Skill Works

Web search is performed via scripts/search_tools.py through the SearchTools interface, using search(query, engine, max_results) with engines including jina, ddg, baidu, and local, and results can be returned as a JSON summary or as a search_list. For broader coverage, aggregate_search(query) combines results from multiple engines. Local context relies on scripts/hybrid_search.py or SearchTools with engine='local' to query the daily_news local database.

When to Use It

• You need general finance information from the open web (e.g., market trends, earnings, or regulatory updates) using jina, ddg, or baidu.
• You must retrieve finance data from a local document store (RAG) such as private reports, notes, or a daily_news database.
• You want to compare results across multiple search engines to verify consistency or spot biases by aggregating results.
• You want to avoid repeated searches by leveraging the Smart Cache to reuse relevant, already retrieved results.
• You need to locate specific finance docs or notes stored locally and pull them into your analysis.

Quick Start

1. Step 1: Choose an engine (e.g., jina) and run search(query, engine, max_results).
2. Step 2: If you need broader results, call aggregate_search(query) to combine engines.
3. Step 3: For private data, switch to engine='local' to fetch from the daily_news RAG store.

Best Practices

• Use aggregate_search(query) when you need a more comprehensive view by combining multiple engines.
• Tune max_results to balance depth and latency; start with 5–10 results for quick triage.
• Read and respect the Smart Cache Relevance Prompt before searching to reduce redundancy.
• Use engine='local' for RAG when you must reference confidential or internal documents.
• Cross-check web results with local sources for consistency and date accuracy.

Example Use Cases

• Search web for 'Q3 2024 earnings highlights' using jina and ddg, then compare results for consistency.
• Query local daily_news for 'investment policy update 2024' to pull internal guidelines.
• Run aggregate_search on 'inflation impact on equities' to synthesize multiple engines' summaries.
• Use local search to retrieve the latest internal market analysis stored in daily_news.
• Check cache for 'Federal interest rate outlook 2025' to avoid repeating the same web query within the day.

Frequently Asked Questions