Vector Search
#13994
Replies: 2 comments 6 replies
-
Vector search in MongoDB is just an aggregation stage, so you can use it with Mongoose's |
Beta Was this translation helpful? Give feedback.
1 reply
-
Here is a script for Vector Search Implementation const mongoose = require('mongoose');
const vectorSchema = new mongoose.Schema({
plot: String,
genres: [String],
year: Number,
myVector: [Number]
});
const Embedded = mongoose.model('vectortest', vectorSchema, 'vectortest');
async function run() {
await mongoose.connect('');
// await mongoose.connection.dropDatabase();
await Embedded.create({
plot: 'Test plot A',
genres: ['Action', 'Adventure', 'Family'],
year: 1993,
myVector: [0.75, 0.1]
});
const agg = [
{
'$vectorSearch': {
'index': 'vector_index',
'path': 'myVector',
'queryVector': [0.75, 0.1], // Adjust the query vector as needed
'numCandidates': 6,
'limit': 1
}
},
{
'$project': {
'_id': 0,
'plot': 1,
'genres': 1,
'year': 1,
'score': {
'$meta': 'vectorSearchScore'
}
}
}
];
const res = await Embedded.aggregate(agg);
console.log('Search result:', res);
console.log('Done');
}
run(); |
Beta Was this translation helpful? Give feedback.
5 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
When will we get Vector Search implementation?
Beta Was this translation helpful? Give feedback.
All reactions