This is alpha quality software, use at your own risk.
elude
helps you download web pages quickly and anonymously for further processing e.g. scraping. The initial groundwork was inspired by Elite Proxy Finder project which attempted to find elite anonymous proxies. elude
builds upon this idea to provide a daemon that automatically finds these proxies and uses them to perform requests. It uses the asyncio
module extensively for handling concurrent connections.
For now, it provides the following:
- Grabs proxies from letushide and checkerproxy and performs regular test requests to make sure that they are still alive.
- Merges requests to the same URL so the target server only gets one request.
- Caches results in-memory to satisfy requests to the same URL (cache expiry is configurable).
- Prefetch feature where URLs can be queued to be requested and cached in the background and then retrieved on-demand later.
- Interface using pure Python (Python 3 only) and Redis.
The aim is to make this usable within Apache Spark jobs. The Spark driver program can first send prefetch requests to elude
daemon, and then the real Spark jobs can query the daemon in real-time as necessary. This enables Apache Spark to perform jobs directly on the downloaded data but it won't be limited to one request per worker thread/process.
- Documentation!
- Also fake user agent strings.
- Deposit fetched results to databases.
- Support for binary data and other mimetypes.