Add LLMSlim integration (document compressor) - #655

This adds integrations/llmslim.md for LLMSlim, an MIT-licensed Python library. Its LLMSlimDocumentCompressor is a Haystack component (2.x and 3) that sits between a retriever and the prompt builder. It keeps the query-relevant parts of the retrieved documents within a token budget. It runs locally: lexical ranking by default, or an optional 22M cross-encoder.

• PyPI: https://pypi.org/project/llmslim/ (the component ships in 0.8.1; install with pip install "llmslim[haystack]")
• Repo and guide: https://github.com/Thanatos9404/llmslim/blob/main/docs/integrations/HAYSTACK.md
• Type: Custom Component
• Tested with haystack-ai 2.0.0, 2.21.0 and 3.3.0. The usage example and its output in the page come from a real run on haystack-ai 3.3.0.

I couldn't find a compression or context-pruning component in the catalog yet, so this one may help people who need to fit retrieved documents into a prompt budget.

Disclosure: I'm the maintainer.