Add InterSystems IRIS Document Store integration

[pedrohenrique-sc]

This PR adds InterSystems IRIS to the Haystack integrations

InterSystems IRIS is an enterprise-grade, high-performance multimodel data platform that features a native VECTOR SQL type and VECTOR_COSINE similarity function. It is widely used in healthcare, finance, and logistics for handling massive, mission-critical workloads.

This integration includes the following components for Haystack 2.x:

• IRISDocumentStore: A production-ready DocumentStore backed by InterSystems IRIS.
• IRISEmbeddingRetriever: Retrieves documents from the IRISDocumentStore using vector embedding similarity.
• IRISBm25Retriever: Retrieves documents from the IRISDocumentStore using Okapi BM25 keyword-based search.

Use case:
Enterprise data architectures often require keeping generative AI workloads close to the transactional or legacy data without duplicating it into standalone vector databases.

By adding this integration, developers can:

• Use InterSystems IRIS as a unified database for both relational metadata and vector embeddings.
• Build advanced Hybrid Search RAG pipelines by combining IRISEmbeddingRetriever (semantic search) and IRISBm25Retriever (lexical search) to improve retrieval accuracy.
• Deploy secure and scalable AI agents leveraging existing InterSystems infrastructure.

Documentation:
Full documentation and usage examples can be found here: InterSystems Iris DocumentStore

[kacperlukawski]

Since it's planned to be published on Haystack's website, I don't think we need a link to Haystack2, although our users would appreciate a link to InterSystems IRIS.

[kacperlukawski]

Suggested change

	In addition to the IRISDocumentStore, the library includes the following Haystack components which can be used in a pipeline:
	- IRISEmbeddingRetriever - A component used to query the vector store and find semantically related Documents. It uses VECTOR_COSINE natively in the database.
	- IRISBm25Retriever - A keyword-based retriever that implements Okapi BM25 over the stored documents.
	In addition to the `IRISDocumentStore`, the library includes the following Haystack components which can be used in a pipeline:
	- `IRISEmbeddingRetriever` - A component used to query the vector store and find semantically related Documents. It uses `VECTOR_COSINE` natively in the database.
	- `IRISBm25Retriever` - A keyword-based retriever that implements Okapi BM25 over the stored documents.

[kacperlukawski]

I would advice running it against all the supported python versions.

[kacperlukawski]

Suggested change

	Once installed, you can start using ``IRISDocumentStore`` as any other document stores that support embeddings.
	Once installed, you can start using `IRISDocumentStore` as any other document stores that support embeddings.

[kacperlukawski]

Suggested change

	Create a ```.env``` file using ```.env.example``` template and import the default config credentials for IntersystemsIris.
	Create a `.env` file using `.env.example` template and import the default config credentials for IntersystemsIris.

Wouldn't that create a block of code, instead of inline one?

[kacperlukawski]

Suggested change

	This integration introduces:
	- `IRISDocumentStore` A DocumentStore backed by InterSystems IRIS
	- `IRISEmbeddingRetriever`Retrieve documents from `IRISDocumentStore` by embedding similarity.
	- `IRISBm25Retriever ` Retrieve documents from `IRISDocumentStore` using Okapi BM25.
	This integration introduces:
	- `IRISDocumentStore` A DocumentStore backed by InterSystems IRIS
	- `IRISEmbeddingRetriever`Retrieve documents from `IRISDocumentStore` by embedding similarity.
	- `IRISBm25Retriever ` Retrieve documents from `IRISDocumentStore` using Okapi BM25.

[kacperlukawski]

Could you please describe the goal of this example? I would also advice splitting it into indexing and search.