Weaviate

• https://github.com/semi-technologies/weaviate

see

• https://www.semi.technology/news/fosdem-2020.html
• https://www.semi.technology/documentation/weaviate/current/about/philosophy.html#about-the-contextionary
• Word2vec
• BERT
• GloVe

Fill with dictionary "e.g. english".

• https://stackoverflow.com/questions/27652335/semantic-search-engines/60361878#60361878
• https://www.semi.technology/news/fosdem-2020.html
• https://en.wikipedia.org/wiki/Bob_van_Luijt
• Weaviate playground
• https://www.semi.technology/documentation/weaviate/current/about/philosophy.html#about-the-contextionary

• https://nlp.stanford.edu/pubs/glove.pdf
• http://wiki.bitplan.com/index.php/Weaviate
• https://hackernoon.com/how-weaviates-graphql-api-was-designed-t93932tl
• https://medium.com/semi-technologies/finding-answers-in-complex-standardizations-documents-using-weaviates-semantic-search-modules-4dcdb83527fb

Versions

• https://medium.com/semi-technologies/weaviate-version-1-2-x-now-supports-transformer-models-4a12d858cce3

Installation

see https://weaviate.io/developers/weaviate/installation

Docker

Prerequisites

sudo apt-get install curl docker docker-compose

Installweaviate

git clone https://github.com/weaviate/weaviate.git
Cloning into 'weaviate'...
remote: Enumerating objects: 124394, done.
remote: Counting objects: 100% (14150/14150), done.
remote: Compressing objects: 100% (1626/1626), done.
remote: Total 124394 (delta 12845), reused 13458 (delta 12510), pack-reused 110244
Receiving objects: 100% (124394/124394), 945.83 MiB | 33.55 MiB/s, done.
Resolving deltas: 100% (86422/86422), done.
cd weaviate
docker build --target weaviate -t weaviate .
docker compose up

docker-compose.yml

version: '3.4'
services:
  weaviate:
    image: weaviate
    ports:
      - 8080:8080
    environment:
      CONTEXTIONARY_URL: contextionary:9999
      QUERY_DEFAULTS_LIMIT: 25
      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'
      PERSISTENCE_DATA_PATH: './data'
      ENABLE_MODULES: 'text2vec-contextionary'
      DEFAULT_VECTORIZER_MODULE: 'text2vec-contextionary'
      AUTOSCHEMA_ENABLED: 'false'
  contextionary:
    environment:
      OCCURRENCE_WEIGHT_LINEAR_FACTOR: 0.75
      EXTENSIONS_STORAGE_MODE: weaviate
      EXTENSIONS_STORAGE_ORIGIN: http://localhost:8080
      NEIGHBOR_OCCURRENCE_IGNORE_PERCENTILE: 5
      ENABLE_COMPOUND_SPLITTING: 'false'
    image: semitechnologies/contextionary:en0.16.0-v1.2.1

Installation Issue

• https://github.com/semi-technologies/weaviate/issues/1215

Python Library

• https://www.semi.technology/documentation/weaviate/current/client-libs/python.html

sample code

Integration tests

• https://github.com/semi-technologies/weaviate-python-client/blob/master/integration/client_functions.py