This topic describes common terms that are used in Intelligent Media Management (IMM).
Metadata
IMM provides metadata management features that you can use to collect, analyze, and query metadata of files stored in services such as Object Storage Service (OSS) and Photo and Drive Service.
Metadata management supports the following file metadata:
Basic file information such as the file size, format, and creation time.
Information about the content of a document and information about files in a package.
Metadata information such as labels, figures, and text generated for media files such as images, videos, and audio based on AI-powered engines.
Project
A project is a fundamental unit for processing and managing files in IMM. You must create a project to use IMM features. You can configure quotas for datasets in a project.
Dataset
A dataset is a container for a collection of metadata stored in IMM. For example, you can develop an online storage application that uses OSS or Photo and Drive Service for data storage and submits the URLs of files in OSS or Photo and Drive Service to your IMM dataset for indexing. IMM automatically parses basic file information and performs AI-driven analysis on media files such as images, videos, and audio to extract labels, figures, and other information.
Workflow
A workflow template defines the steps of a metadata extraction task in IMM. When you create a project or dataset in IMM, you can specify a workflow template for the project or dataset.
For more information about the workflow templates that IMM provides, see Workflow templates.
Operator
An operator is a fundamental unit of workflow execution. The processing engine of IMM determines operators to use based on the specified workflow template to collect metadata and index the metadata into a dataset.
Request type
Synchronous request
A client sends a request and waits for a response, which is returned within hundreds of milliseconds. For example, face detection and content recognition in image analysis scenarios use synchronous requests.
Asynchronous request
IMM uses an asynchronous request mechanism for long-running tasks to prevent high client latency. For example, a client might time out and disconnect if it does not receive a response within a specific period, such as 10 seconds.
After a client sends a request, the system immediately returns an ID that you can use to query the status of the asynchronous operation. For example, a format conversion for a large file may take tens of seconds or even minutes to complete.
Task
A task is an executing entity that processes an asynchronous request. When you call an asynchronous API operation, the task ID is returned in the response. For example, when you send an asynchronous request to convert the format of a document, the executing entity that processes the request is a task.
You can query the status of tasks, delete tasks, and traverse tasks.