Module create 

File upload functionality for OpenAI-compatible APIs

This module provides functionality for uploading files to OpenAI-compatible APIs. Files can be uploaded for various purposes such as fine-tuning, batch processing, assistants, and more.

Overview

The create module contains:

• request: File upload request types and functionality
• response: File object response types
• Support for multipart/form-data file uploads
• Comprehensive error handling for file operations

Features

• File Upload: Upload files up to 512 MB in size
• Multiple Purposes: Support for various file purposes (assistants, batch, fine-tune, etc.)
• Expiration Policies: Configurable file expiration policies
• Multipart Uploads: Proper handling of multipart/form-data requests

Examples

Basic File Upload

This example demonstrates how to upload a file for batch processing:

use std::path::PathBuf;
use std::sync::LazyLock;
use openai_interface::files::create::request::{CreateFileRequest, FilePurpose};
use openai_interface::files::create::response::FileObject;
use openai_interface::rest::post::NoStream;

const MODELSCOPE_URL: &str = "https://dashscope.aliyuncs.com/compatible-mode/v1/files";
const MODELSCOPE_KEY: LazyLock<&str> =
    LazyLock::new(|| include_str!("../../../keys/modelstudio_domestic_key").trim());

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let file_path = PathBuf::from("src/files/create/file-test.txt");
    let create_file_request = CreateFileRequest {
        file: file_path,
        purpose: FilePurpose::Batch,
        ..Default::default()
    };

    let response: FileObject = create_file_request
        .get_response(MODELSCOPE_URL, MODELSCOPE_KEY.as_ref())
        .await?;

    println!("Uploaded file ID: {}", response.id);
    println!("File size: {} bytes", response.bytes);
    println!("Filename: {}", response.filename);
    Ok(())
}

File Upload with Custom Purpose

This example shows how to upload a file with a custom purpose:

use std::path::PathBuf;
use std::sync::LazyLock;
use openai_interface::files::create::request::{CreateFileRequest, FilePurpose};
use openai_interface::files::create::response::FileObject;
use openai_interface::rest::post::NoStream;

const MODELSCOPE_URL: &str = "https://dashscope.aliyuncs.com/compatible-mode/v1/files";
const MODELSCOPE_KEY: LazyLock<&str> =
    LazyLock::new(|| include_str!("../../../keys/modelstudio_domestic_key").trim());

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let file_path = PathBuf::from("src/files/create/file-test.txt");
    let create_file_request = CreateFileRequest {
        file: file_path,
        purpose: FilePurpose::Other("file-extract".to_string()),
        ..Default::default()
    };

    let response: FileObject = create_file_request
        .get_response(MODELSCOPE_URL, MODELSCOPE_KEY.as_ref())
        .await?;

    println!("Uploaded file with custom purpose: {}", response.id);
    Ok(())
}

File Size Limits

• Individual files can be up to 512 MB
• Total organization storage can be up to 1 TB
• Batch API files are limited to 200 MB
• Fine-tuning API only supports .jsonl files

Supported File Purposes

• assistants: Used in the Assistants API
• batch: Used in the Batch API (expires after 30 days by default)
• fine-tune: Used for fine-tuning
• vision: Images used for vision fine-tuning
• user_data: Flexible file type for any purpose
• evals: Used for eval data sets
• Custom purposes via the Other variant

Error Handling

The module provides comprehensive error handling for:

• File not found errors
• File read permissions
• Network connectivity issues
• API authentication failures
• Invalid file formats
• Size limit violations

Modules

request

response