Skip to main content

text2vec-jinaai

Overview​

The text2vec-jinaai module enables Weaviate to obtain vectors using JinaAI Embeddings.

Key notes:

  • As it uses a third-party API, you will need an API key.
  • Its usage may incur costs.
  • JinaAI requires a third-party API key. You can obtain one here.
  • When you enable the text2vec-jinaai model, you can use the nearText search operator.
  • The default model is jina-embeddings-v2-base-en.
Added in v1.22.3

Weaviate instance configuration​

Docker Compose file​

To use text2vec-jinaai, you must enable it in your Docker Compose file (docker-compose.yml). You can edit the Docker Compose file manually, or use the the Weaviate configuration tool to create a custom file.

Parameters​

ParameterRequiredPurpose
ENABLE_MODULESYesThe modules to enable. Include text2vec-jinaai to enable the module.
DEFAULT_VECTORIZER_MODULENoThe default vectorizer module. To make text2vec-jinaai the default for all classes, set it here.
JINAAI_APIKEYNoYour JinaAI API key. You can also provide the key at query time.

Example​

This Docker Compose file shows how to use JinaAI as the vectorizer.

  • It enables text2vec-jinaai.
  • It sets text2vec-jinaai as the default vectorizer.
  • It sets a JinaAI API key.
---
version: '3.4'
services:
  weaviate:
    image: cr.weaviate.io/semitechnologies/weaviate:1.24.10
    restart: on-failure:0
    ports:
     - 8080:8080
     - 50051:50051
    environment:
      QUERY_DEFAULTS_LIMIT: 20
      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'
      PERSISTENCE_DATA_PATH: "./data"
      ENABLE_MODULES: 'text2vec-jinaai'
      DEFAULT_VECTORIZER_MODULE: 'text2vec-jinaai'
      JINAAI_APIKEY: 'YOUR_JINAAI_API_KEY'  # Setting this parameter is optional, you can also provide the key at query time.
      CLUSTER_HOSTNAME: 'node1'
...

Collection configuration​

To configure how the module behaves in each collection, update the Weaviate schema.

API settings​

Parameters​

ParameterRequiredDefaultPurpose
modelNojina-embeddings-v2-base-enA model name, e.g. jina-embeddings-v2-small-en.

Example​

The following example configures the Document collection by setting the vectorizer to text2vec-jinaai and the model to jina-embeddings-v2-small-en:

{
  "classes": [
    {
      "class": "Document",
      "description": "A collection called document",
      "vectorizer": "text2vec-jinaai",
      "moduleConfig": {
        "text2vec-jinaai": {
          "model": "jina-embeddings-v2-small-en",
        }
      },
    }
  ]
}

Vectorization settings​

You can set vectorizer behavior using the moduleConfig section under each collection and property:

Collection level settings​

ParameterTypeDefaultPurpose
vectorizerstring-Sets the module for vectorization.
vectorizeClassNamebooleantrueWhether to include the class name during vectorization.

Property level settings​

ParameterTypeDefaultPurpose
skipbooleanfalseWhen true, does not include the property during vectorization.
vectorizePropertyNamebooleanfalseWhether to include the property name during vectorization.

Example​

{
  "classes": [
    {
      "class": "Document",
      "description": "A collection called document",
      "vectorizer": "text2vec-jinaai",
      "moduleConfig": {
        "text2vec-jinaai": {
          "model": "jina-embeddings-v2-small-en",
          "vectorizeClassName": false
        }
      },
      "properties": [
        {
          "name": "content",
          "dataType": ["text"],
          "description": "Content that will be vectorized",
          "moduleConfig": {
            "text2vec-jinaai": {
              "skip": false,
              "vectorizePropertyName": false
            }
          }
        }
      ]
    }
  ]
}

Query-time parameters​

API key​

You can supply the API key at query time by adding it to the HTTP header.

HTTP HeaderValuePurpose
X-Jinaai-Api-KeyYOUR-JINAAI-API-KEYJinaAI API key

Usage example​

This is an example of a nearText query that uses text2vec-jinaai.

import weaviate
from weaviate.classes.query import MetadataQuery, Move
import os

client = weaviate.connect_to_local(
    headers={
        "X-Jinaai-Api-Key": "YOUR_JINAAI_APIKEY",
    }
)

publications = client.collections.get("Publication")

response = publications.query.near_text(
    query="fashion",
    distance=0.6,
    move_to=Move(force=0.85, concepts="haute couture"),
    move_away=Move(force=0.45, concepts="finance"),
    return_metadata=MetadataQuery(distance=True),
    limit=2
)

for o in response.objects:
    print(o.properties)
    print(o.metadata)

client.close()

Additional information​

Available models​

The following models are available:

  • jina-embeddings-v2-base-en (Default)
  • jina-embeddings-v2-small-en

Questions and feedback​

If you have any questions or feedback, let us know in our user forum.