Document Query

Enables natural conversational interactions with documents stored in Smart Buckets. This endpoint allows users to ask questions, request summaries, and explore document content through an intuitive conversational interface. The system understands context and can handle complex queries about document contents.

The query system maintains conversation context throught the request_id, enabling follow-up questions and deep exploration of document content. It works across all supported file types and automatically handles multi-page documents, making complex file interaction as simple as having a conversation.

The system will:

• Maintain conversation history for context when using the same request_id
• Process questions against file content
• Generate contextual, relevant responses

Document query is supported for all file types, including PDFs, images, and audio files.

JavaScript

import Raindrop from '@liquidmetal-ai/lm-raindrop';

const client = new Raindrop({
  apiKey: process.env['RAINDROP_API_KEY'], // This is the default and can be omitted
});

async function main() {
  const documentQuery = await client.documentQuery.create({
    bucket: 'bucket',
    input: 'input',
    object_id: 'object_id',
    request_id: 'request_id',
  });

  console.log(documentQuery.answer);
}

main();

Python

import os
from lm_raindrop import Raindrop

client = Raindrop(
    api_key=os.environ.get("RAINDROP_API_KEY"),  # This is the default and can be omitted
)
document_query = client.document_query.create(
    bucket="bucket",
    input="input",
    object_id="object_id",
    request_id="request_id",
)
print(document_query.answer)

Go

package main

import (
  "context"
  "fmt"

  "github.com/LiquidMetal-AI/raindrop-go-sdk"
  "github.com/LiquidMetal-AI/raindrop-go-sdk/option"
)

func main() {
  client := raindrop.NewClient(
    option.WithAPIKey("My API Key"), // defaults to os.LookupEnv("RAINDROP_API_KEY")
  )
  documentQuery, err := client.DocumentQuery.New(context.TODO(), raindrop.DocumentQueryNewParams{
    Bucket: raindrop.F("bucket"),
    Input: raindrop.F("input"),
    ObjectID: raindrop.F("object_id"),
    RequestID: raindrop.F("request_id"),
  })
  if err != nil {
    panic(err.Error())
  }
  fmt.Printf("%+v\n", documentQuery.Answer)
}

Java

package com.raindrop.api.example;

import com.raindrop.api.client.RaindropClient;
import com.raindrop.api.client.okhttp.RaindropOkHttpClient;
import com.raindrop.api.models.documentquery.DocumentQueryCreateParams;
import com.raindrop.api.models.documentquery.DocumentQueryCreateResponse;

public final class Main {
    private Main() {}

    public static void main(String[] args) {
        // Configures using the `RAINDROP_API_KEY` environment variable
        RaindropClient client = RaindropOkHttpClient.fromEnv();

        DocumentQueryCreateParams params = DocumentQueryCreateParams.builder()
            .bucket("bucket")
            .input("input")
            .objectId("object_id")
            .requestId("request_id")
            .build();
        DocumentQueryCreateResponse documentQuery = client.documentQuery().create(params);
    }
}

CLI

# Query a document in a bucket
npx raindrop query document "What are the key points in this document?" \
  --bucket my-bucket \
  --object-id document.pdf