#community-help

Error in Implementing Vector Search

TLDR Krish faced an error while implementing vector search. After recreating the collection as suggested by Kishore Nallan, the issue was resolved.

8mo

Join the chat

Apr 05, 2023 (8 months ago)

Photo of md5-7dff5651efcba5c9cdfac93918dc7eeb

Krish

06:02 AM

Hello! I am currently on 0.24 and trying to implement vector search. Below is my schema. But when I trying a search as mentioned below I am getting error: "{'results': [{'code': 400, 'error': 'Field embeddings does not have a vector query index.'}]}". Not sure where I am going wrong. Can you help please?
search_requests = {
'searches': [{
'collection': 'wh-records',
'q': '*',
'vector_query': 'embeddings:([1.4580078, -1.7529297, -0.16894531, 2.4335938,....], k:50)',
'per_page': 50,
'exclude_fields': 'embeddings'
}]
}
common_search_params = {}
result = typesenseClient.multi_search.perform(search_requests, common_search_params)
print(result)

schema = {
  "name": "wh-records",
  "fields": [
    {
      "name"  :  "id",
      "type"  :  "string"
    },
    {
      "name"  :  "title",
      "type"  :  "string",
      "facet" :  True
    },
    {
      "name"  :  "url",
      "type"  :  "string"
    },
    {
      "name"  :  "date",
      "type"  :  "string"
    },
    {
      "name"  :  "summary",
      "type"  :  "string",
      "facet" :  True
    },
    {
      "name"  :  "embeddings",
      "type"  :  "float[]",
      "num_dim" :  4096
    }
  ] 
}

Photo of md5-4e872368b2b2668460205b409e95c2ea

Kishore Nallan

06:09 AM

Query field name is wrong, should be "vector"

Krish

06:09 AM

Sorry, thats a copy paste error. it is embeddings

06:10

Krish

06:10 AM

 {
        "facet": false,
        "index": true,
        "infix": false,
        "locale": "",
        "name": "embeddings",
        "optional": false,
        "sort": false,
        "type": "float[]"
      },

06:10

Krish

06:10 AM

from the collection.

Kishore Nallan

06:10 AM

Try creating a standalone example where the error occurs and share it with me so I can reproduce

Krish

06:11 AM

standalone example as in?

06:11

Krish

06:11 AM

I did not see this issue in 0.24.0.rcn56

Kishore Nallan

06:12 AM

What are you using now?

Krish

06:12 AM

now I moved to 0.24.0

06:12

Krish

06:12 AM

let me try in curl and see

Kishore Nallan

06:13 AM

Try the example in the docs first: https://typesense.org/docs/0.24.0/api/vector-search.html#adding-a-vector-field

Krish

06:19 AM

This worked!

06:23

Krish

06:23 AM

So, what does this "field embeddings does not have a vector query index." mean?

Kishore Nallan

06:24 AM

I can't say unless you gave me a fully reproduce able example to run

06:25

Kishore Nallan

06:25 AM

But that error broadly means that somehow the field was not detected as vector field. This happens if you had omitted num_dim for example when creating collection.

Krish

06:26 AM

ok, let me create the collection again try

06:53

Krish

06:53 AM

dropping the collection and recreating it worked. May be there was something wrong in the previous one.

Typesense

Lightning-fast, open source search engine for everyone | Knowledge Base powered by Struct.AI

Indexed 3015 threads (79% resolved)

Join Our Community

Similar Threads

Integrating Semantic Search with Typesense

Krish wants to integrate a semantic search functionality with typesense but struggles with the limitations. Kishore Nallan provides resources, clarifications and workarounds to the raised issues.

11mo

Utilizing Vector Search and Word Embeddings for Comprehensive Search in Typesense

Bill sought clarification on using vector search with multiple word embeddings in Typesense and using them instead of OpenAI's embedding. Kishore Nallan and Jason informed him that their development version 0.25 supports open source embedding models. They also resolved Bill's concerns regarding search performance, language support, and limitations in the search parameters.

225

4mo

Hybrid Search Distance Threshold Issue

Anish has an issue with search results not respecting the vector distance threshold when using hybrid search. Jason explains additional fields cause `vector_distance` to only apply to vector search results and suggests opening a feature request on GitHub.

2mo

Issue with Embedding Error in Version 0.25.0.rc63

Bill reported a bug in version 0.25.0.rc63 regarding a problem with updating or emplacing a document and receiving an embedding error. This was resolved in version 0.25.0.rc65, but further discussion ensued regarding the function of 'index' in the update feature.

4mo

Understanding Vector Naming in Typesense

Koushik was confused by inconsistent vector naming in Typesense documents. Jason clarified that the vector can be named anything, as long as it has a 'num_dim' property.

4mo