https://github.com/ksylvest/voyageai
A ruby client for VoyageAI
https://github.com/ksylvest/voyageai
ruby voyageai
Last synced: 8 months ago
JSON representation
A ruby client for VoyageAI
- Host: GitHub
- URL: https://github.com/ksylvest/voyageai
- Owner: ksylvest
- License: mit
- Created: 2024-10-21T21:15:23.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-09-29T17:56:58.000Z (9 months ago)
- Last Synced: 2025-10-10T00:57:32.928Z (8 months ago)
- Topics: ruby, voyageai
- Language: Ruby
- Homepage: https://voyageai.ksylvest.com
- Size: 43 KB
- Stars: 7
- Watchers: 1
- Forks: 4
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE.txt
Awesome Lists containing this project
README
# VoyageAI
[](https://github.com/ksylvest/voyageai/blob/main/LICENSE)
[](https://rubygems.org/gems/voyageai)
[](https://github.com/ksylvest/voyageai)
[](https://voyageai.ksylvest.com)
[](https://circleci.com/gh/ksylvest/voyageai)
`voyageai` is a ruby client for [VoyageAI](https://www.voyageai.com)
## Installation
```bash
gem install voyageai
```
## Usage
### Embedding
#### Generating Single Embedding
```ruby
require 'voyageai'
input = 'A quick brown fox jumps over the lazy dog.'
voyageai = VoyageAI::Client.new(api_key: 'pa-...') # or configure ENV['VOYAGEAI_API_KEY']
embed = voyageai.embed(input)
embed.model # "..."
embed.usage # "#"
embed.embedding # [0.0, ...]
```
#### Generating Multiple Embeddings
```ruby
require 'voyageai'
input = [
'John is a musician.',
'Paul is a plumber.',
'George is a teacher.',
'Ringo is a doctor.',
]
voyageai = VoyageAI::Client.new(api_key: 'pa-...') # or configure ENV['VOYAGEAI_API_KEY']
embed = voyageai.embed(input)
embed.model # "..."
embed.usage # "#"
embed.embeddings # [[0.0, ...], ...]
```
### Reranking
```ruby
require 'voyageai'
query = 'Who is the best person to call for a toilet?'
documents = [
'John is a musician.',
'Paul is a plumber.',
'George is a teacher.',
'Ringo is a doctor.',
]
voyageai = VoyageAI::Client.new(api_key: 'pa-...') # or configure ENV['VOYAGEAI_API_KEY']
rerank = voyageai.rerank(query:, documents:, top_k: 3)
rerank.model # "..."
rerank.usage # "#"
rerank.results # [#]
```
## Configuration
```ruby
require 'voyageai'
VoyageAI.configure do |config|
config.api_key = 'pa-...' # defaults to ENV['VOYAGEAI_API_KEY']
config.host = 'https://api.voyageai.com'
config.version = 'v1'
config.timeout = 15 # seconds
config.logger = Logger.new(STDOUT)
end
```
## Examples
```ruby
require "voyageai"
Entry = Data.define(:document, :embedding)
VOYAGEAI = VoyageAI::Client.new
DOCUMENTS = [
"John is a musician.",
"Paul is a plumber.",
"George is a teacher.",
"Ringo is a doctor.",
"Lisa is a lawyer.",
"Stuart is a painter.",
"Brian is a writer.",
"Jane is a chef.",
"Bill is a nurse.",
"Susan is a carpenter.",
].freeze
embeddings = VOYAGEAI.embed(DOCUMENTS, input_type: "document").embeddings
ENTRIES = DOCUMENTS.zip(embeddings).map do |document, embedding|
Entry.new(document:, embedding:)
end
# @param src [Array]
# @param dst [Array]
#
# @return [Float]
def euclidean_distance(src, dst)
Math.sqrt(src.zip(dst).map { |a, b| (a - b)**2 }.reduce(:+))
end
# @param query [String]
# @param limit [Integer]
#
# @return [Array]
def nearest_documents(query:, limit: 4)
embedding = VOYAGEAI.embed(query, input_type: "query").embedding
ENTRIES
.sort_by { |entry| euclidean_distance(entry.embedding, embedding) }
.first(limit)
.map(&:document)
end
# @param query [String]
def search(query:)
documents = nearest_documents(query:)
results = VOYAGEAI.rerank(query:, documents:, top_k: 2).results
puts "query=#{query.inspect}"
results.each do |reranking|
document = documents[reranking.index]
puts("document=#{document.inspect} relevance_score=#{reranking.relevance_score}")
end
end
search(query: "What do George and Ringo do?")
search(query: "Who works in the medical field?")
```
```
query="What do George and Ringo do?"
document="Ringo is a doctor." relevance_score=0.67968755
document="George is a teacher." relevance_score=0.58593755
```
```
query="Who works in the medical field?"
document="Bill is a nurse." relevance_score=0.55078125
document="Ringo is a doctor." relevance_score=0.50390625
```