|
--- |
|
title: Bolete |
|
emoji: π |
|
colorFrom: green |
|
colorTo: red |
|
sdk: streamlit |
|
sdk_version: 1.10.0 |
|
app_file: app.py |
|
pinned: false |
|
license: mit |
|
--- |
|
# bolete |
|
|
|
An information extraction and exploration app. Upload files with text. Bolete will then extract text, identify common keywords and entities, and create a simple search interface to explore the corpus. |
|
|
|
- Search of collection texts |
|
- how best to search collection? |
|
- Holmes? |
|
- Entities and frequencies |
|
- TODO run ner, return filters for most frequent ents in corpus |
|
|
|
- Fun |
|
- use with scispaCy rather than generic model |
|
en_core_sci_lg + en_core_sci_scibert |