Smart searching across millions of documents
Linear scalability and consistent response times
Our tests on corpora of millions of documents show a consistent response time. It can be used by means of a web application or a REST API.
As a web application, it offers a sophisticated but intuitive means of writing search terms.
Demo here.
Description of the technological basis
For storage and recovery, it uses the open-source Apache solr technology standard and KeyQ, our terminology extractor.
The web interface can be adapted to any kind of device (mobile phone, PC, tablet, etc.), using Bootstrap.
Business needs / application
Companies with large volumes of documents, in all kinds of formats (PDF, Word, etc.). Searching for information in their document repositories is a need shared by all kinds of companies.
Competitive advantages
Compared to the basic version, KeyQ (terminology extraction and online command searching, with the corpus stored in the memory), KeyQ-solr distributes the corpus documents to the hard disks of one or several machines using Apache solr technology. Its other features include:
– Filters: partial searches, upper/lower-case sensitivity, file metadata.
– Granularity selector for the information retrieved: page, paragraph or sentence.
– Detailed statistics of the corpus texts.
– Advanced visual analysis to identify thematic groupings.
– Handling of multiple corpora in different languages.
– Terminology management. Collaborative terminology evaluator (consensus among experts).
– Searches with logical operators.
Past performance references
We have handled corpora in fields such as:
- Biology: tens of thousands of scientific articles on Covid-19.
- Law: large documents with hundreds of pages, with instant access to the most relevant page/paragraph/sentence for a particular search.
- Energy and the environment: thousands of European Parliament documents in different formats.
KeyQ was developed in the Artificial Intelligence R&I Centre (AI.nnovation Space) (joint UPM-Accenture centre) between 2020 and 2021.
Protection
- Software registration
Stage of development
- Concept
- Research
- Lab prototype
- Industrial prototype
- Production
KeyQ-solr is what you were searching for. Now let us search for you.
