async-web-search

Web Search

Async web search library supporting Google Custom Search, Wikipedia, arXiv, NewsAPI, GitHub, and PubMed APIs.

You can search across multiple sources and retrieve relevant, clean results in JSON format or as compiled text.

🌟 Features

• ⚡ Asynchronous Searching: Perform searches concurrently across multiple sources
• 🔗 Multi-Source Support: Query Google Custom Search, Wikipedia, arXiv, NewsAPI, GitHub, and PubMed
• 🧹 Content extraction and cleaning
• 🔧 Configurable Search Parameters: Adjust maximum results, preview length, and sources.

📋 Prerequisites

• 🐍 Python 3.8 or newer
• 🔑 API keys and configuration:
  • Google Search: Requires a Google API key and a Custom Search Engine (CSE) ID.
  • NewsAPI: Requires a free API key from newsapi.org.
  • arXiv: No API key required.
  • Wikipedia: No API key required.
  • GitHub: No API key required.
  • PubMed: No API key required.

Set environment variables:

export GOOGLE_API_KEY="your_google_api_key"
export CSE_ID="your_cse_id"
export NEWSAPI_KEY="your_newsapi_key"

📦 Installation

pip install async-web-search

🛠️ Usage

Example 1: Search across multiple sources

from web_search import WebSearch, WebSearchConfig

config = WebSearchConfig(sources=["google", "arxiv", "github", "newsapi", "pubmed"])
results = await WebSearch(config).search("quantum computing")

# results is a list of dicts with keys: url, title, preview, source
for result in results:
    print(f"Title: {result['title']}")
    print(f"URL: {result['url']}")
    print(f"Preview: {result['preview']}")
    print(f"Source: {result['source']}")
    print("---")

Example 1.1: Compiled search results as string

from web_search import WebSearch, WebSearchConfig

config = WebSearchConfig(sources=["google", "arxiv", "github"])
compiled_results = await WebSearch(config).compile_search("quantum computing")

print(compiled_results)  # Prints a formatted string with all results

Example 2: Google Search

from web_search import GoogleSearchConfig
from web_search.google import GoogleSearch

config = GoogleSearchConfig(
    api_key="your_google_api_key",
    cse_id="your_cse_id",
    max_results=5
)
results = await GoogleSearch(config)._search("quantum computing")

for result in results:
    print(result)

Example 3: Wikipedia Search

from web_search import BaseConfig
from web_search.wikipedia_ import WikipediaSearch

wiki_config = BaseConfig(max_results=5)
results = await WikipediaSearch(wiki_config)._search("deep learning")

for result in results:
    print(result)

Example 4: ArXiv Search

from web_search import BaseConfig
from web_search.arxiv import ArxivSearch

arxiv_config = BaseConfig(max_results=3)
results = await ArxivSearch(arxiv_config)._search("neural networks")

for result in results:
    print(result)

🌐 Production API Server

A FastAPI-based production server is available for teams that want to use async web search as a web service. The server is hosted at https://awebs.veedo.ai and can be run locally as well.

See server/README.md for detailed API documentation, endpoints, and deployment instructions.

📘 API Overview

🔧 Configuration

• BaseConfig: Shared configuration for all sources (e.g., max_results and timeout).
• GoogleSearchConfig: Google-specific settings (e.g., api_key, cse_id).
• WebSearchConfig: Configuration for the overall search process (e.g., sources to query).

📚 Classes

• WebSearch: Entry point for performing searches across multiple sources.
• GoogleSearch: Handles searches via Google Custom Search Engine API.
• WikipediaSearch: Searches Wikipedia and retrieves article previews.
• ArxivSearch: Queries arXiv for academic papers.

⚙️ Methods

• search(query: str): Main search method for WebSearch.
• _search(query: str): Source-specific search logic for GoogleSearch, WikipediaSearch, and ArxivSearch.

🤝 Contributing

We welcome contributions! To contribute:

• Fork the repository.
• Create a new branch (git checkout -b feature-name).
• Commit your changes (git commit -am "Add new feature").
• Push to the branch (git push origin feature-name).
• Open a pull request.

🧪 Running Tests

pytest -v

License

MIT