Semantically Search PDF Business Reports

[seduerr91]

Description

Use Haystack’s Semantic Document Search with Hugging Face models in Gradio on any given pdf. Take a look:

Please find explanations on Medium.

Tagline

Helping non-technical users benefit from semantic document search in pdfs.

Author(s)

Authored by Seb: duerr.sebastian@gmail.com

Link to showcase

https://huggingface.co/spaces/seduerr/semantic_search

The us what task this project does

With deepsets.ai’s Haystack framework, we can answer questions about a a pdf's content with semantic search. Semantic search is powered by deep neural nets and allows you to search documents based on similar phrasing and context (word embeddings). This article by deepset.ai explains the specificities of semantic document search. In their article, the authors provide an intuitive illustration towards Semantic Document Search which is finding similar documents to queries: .

Some real life applications for such functionality is scanning and reading corporate business reports.

Haystack Nodes

• [ ] BM25Retriever
• [ ] TableTextRetriever
• [ ] EmbeddingRetriever
• [ ] DensePassageRetriever
• [X] TfidfRetriever
• [X] FARMReader
• [ ] TableReader
• [ ] RAGenerator
• [ ] TransformersSummarizer
• [ ] TransformersTranslator
• [ ] TransformersQueryClassifier
• [ ] SklearnQueryClassifier
• [ ] QuestionGenerator
• [ ] TransformersDocumentClassifier
• [ ] EntityExtractor
• [ ] RouteDocuments
• [ ] JoinDocuments
• [ ] JoinAnswers
• [ ] Docs2Answers
• [ ] Custom Node

Document Stores

• [ ] ElasticsearchDocumentsStore
• [ ] OpenSearchDocumsntStore
• [X] InMemoryDocumentStore
• [ ] WeaviateDocumentStore
• [ ] FAISSDocumentStore
• [ ] OpenDistroElasticsearchDocumentStore
• [ ] MilvusDocumentStore
• [ ] SQLDocumentStore
• [ ] PineconeDocumentStore

[github-actions[bot]]

Thank you for submitting a Haystack showcase. We've created a PR for review and will publish your showcase asap. Thank you for your contribution! :sparkles: