aspect-sdk 0.0.3

Python SDK for Aspect Media Engine API

Aspect Python SDK

This SDK delivers an intuitive interface for building applications with Aspect's video intelligence platform. Designed for Python developers, it provides comprehensive methods for video analysis, search, and management while eliminating complex integration overhead and reducing development time.

Installation

pip install aspect-sdk

Usage

Initialize the Aspect client

Sign up for your api key here and then head to the Api keys tab

from aspect_sdk import Aspect, AspectConfig

client = Aspect(AspectConfig(
    api_key='your-api-key',  # Required: Your API key 
))

Create an Index

Indexes are organizational containers for your assets. Think of them like folders or collections.

from aspect_sdk import IndexCreateRequest

index = client.indexes.create(IndexCreateRequest(
    name='My Video Collection',
    description='Collection of marketing videos',
    features=['embedding']  # Optional: specify which AI features to run by default on all assets when they're created in this index
))

print('Created index:', index.id)

Create an Asset

Assets are videos or images that get AI-indexed by the system. You can upload from a file path, URL, or file object.

from aspect_sdk import AssetCreateRequest

# Upload from file path
response = client.assets.create(AssetCreateRequest(
    index_id=index.id,
    name='video.mp4',
    asset_file='/path/to/video.mp4',
    # asset_url='https://example.com/video.mp4',  # can also optionally pass a url instead of file
    save_original=True,  # Whether to store the original file
    features=['transcription']  # Optional: specify which AI features to additionally run specifically for this asset (union with index default features)
))

asset_id = response.asset_id
task_id = response.task_id

print('Created asset from file:', asset_id)

# Wait on the asset to finish indexing its features
task = client.tasks.wait_for_done(task_id, WaitForDoneOptions(
    interval=5000,
    callback=lambda task: print(task.features)
))

if task.features.transcription.state == "failed":
    raise Exception("Transcription failed")

Search Assets

Search across your indexed video content using natural language queries.

# Basic search
search_results = client.search.run({
    'index_id': index.id,
    'query': "",
})
print("Search results", search_results)

# Note: The search API is currently being developed and will support
# parameters like filters, sorting, and pagination

API Structure

The SDK is organized into resource-based modules:

• client.indexes - Create and manage indexes (collections)
• client.assets - Upload and manage video assets
• client.search - Search across indexed content
• client.tasks - Monitor processing tasks
• client.users - User account management
• client.analyze - AI analysis operations

Asset Processing

When you create an asset, Aspect automatically runs tasks for the playground:

• Proxies - Optimized versions for streaming
• Previews - Thumbnail images and preview clips

You can choose to run these AI jobs whenever you want (either through index's default features, assets.create, or tasks.create):

• Transcription - Speech-to-text extraction
• Embedding - Semantic vector embeddings for search

You can monitor processing status through the task polling system or via webhooks.

Type Safety

This SDK is built with type safety in mind using Pydantic models and type hints:

from aspect_sdk import (
    IndexCreateRequest,
    AssetCreateResponse,
    AspectConfig,
    WaitForDoneOptions,
)

Error Handling

from aspect_sdk import ApiException

try:
    asset = client.assets.create(AssetCreateRequest(
        index_id='invalid-id',
        name='video.mp4',
        asset_file='/path/to/video.mp4',
        save_original=True
    ))
except ApiException as error:
    print('Failed to create asset:', str(error))

Advanced Usage

File Input Types

The SDK supports multiple file input types for asset creation:

# File path (string)
asset = client.assets.create(AssetCreateRequest(
    index_id=index_id,
    name='video.mp4',
    asset_file='/path/to/video.mp4',
    save_original=True
))

# Pathlib.Path object
from pathlib import Path
asset = client.assets.create(AssetCreateRequest(
    index_id=index_id,
    name='video.mp4',
    asset_file=Path('/path/to/video.mp4'),
    save_original=True
))

# Raw bytes
with open('/path/to/video.mp4', 'rb') as file:
    video_bytes = file.read()

asset = client.assets.create(AssetCreateRequest(
    index_id=index_id,
    name='video.mp4',
    asset_file=video_bytes,
    save_original=True
))

# File-like object
with open('/path/to/video.mp4', 'rb') as file:
    asset = client.assets.create(AssetCreateRequest(
        index_id=index_id,
        name='video.mp4',
        asset_file=file,
        save_original=True
    ))

License

MIT License