Haystack Document Store

Opensolr in Haystack pipelines

opensolr-haystack adds an Opensolr document store, a hybrid retriever and a file converter to Haystack. Your pipelines need no embedder components: documents and queries get their meaning vectors on Opensolr servers.

Install and first pipeline

pip install opensolr-haystack

from haystack import Document, Pipeline
from haystack_integrations.document_stores.opensolr import OpensolrDocumentStore
from haystack_integrations.components.retrievers.opensolr import OpensolrHybridRetriever

# credentials come from OPENSOLR_EMAIL and OPENSOLR_API_KEY
store = OpensolrDocumentStore(index="my_index", create_if_missing=True)
store.write_documents([Document(content="Hybrid search ranks words and meaning together")])

pipe = Pipeline()
pipe.add_component("retriever", OpensolrHybridRetriever(document_store=store))
print(pipe.run({"retriever": {"query": "keyword and meaning search"}}))

What you get

• OpensolrDocumentStore: the full document store protocol, with standard filters.
• OpensolrHybridRetriever: words and meaning ranked together, with alpha for the balance, or meaning only.
• OpensolrConverter: files and folders turned into documents, a PDF page by page, scans and pictures through OCR.
• search_by_image(): search with a photo.
• ai_answer(): a grounded answer written from the best documents, with no generator component and no LLM key.
• A keyword-only mode that uses none of your AI allowance.

Links

• Haystack product page
• PyPI and GitHub, with the full guide in the README.
• No account yet? Use the public demo account.