The data indexing feature of Alibaba Cloud Object Storage Service (OSS) enables you to search for multimodal data, such as images, videos, audio files, and documents, based on semantics. This allows for intelligent and accurate content searches.
Scenarios
A media company, Company B, created a bucket named mybucket in the China (Guangzhou) region to store a large amount of multimedia material. Because the file names did not contain descriptive information, searching for materials was inefficient. This inefficiency negatively impacted the productivity of ad creation and content planning.
Using OSS data indexing, Company B achieved the following objectives:
Multimodal search: Quickly find related multimedia files, such as images, videos, and audio files, by entering a keyword such as "Summer beach" and setting filter conditions. For example, you can search for images that are larger than 500 KB, wider than 100 px, and taller than 100 px.
Simple and easy to use: You can perform searches by calling the OSS Open API to quickly build an enterprise AI assistant.
Real-time updates: Search results are updated in near real-time as new files are uploaded, which prevents information lag.
Step 1: Enable data indexing
Log on to the OSS console.
Click Bucket List, and then click the name of the target bucket.
In the navigation pane on the left, choose .
On the Data Indexing page, if this is your first time using data indexing, follow the prompts to grant permissions to AliyunMetaQueryDefaultRole. This allows OSS to manage data in the bucket. After granting the permissions, click Enable Data Indexing.
Select AISearch and click Enable.
Step 2: Initiate a query request
Set the data search conditions based on the requirements of Company B.
In the navigation pane on the left, choose .
In the Search Content input box, enter "Summer beach".
Set File Size to greater than 500 KB.
Set Multimedia Type to Image.
Set Image Format to JPG/JPEG.
Set Image Width to greater than 100 px.
Set Image Height to greater than 100 px.
Click Query Now.
Step 3: Verify the results
The following figure shows the files that match the criteria.
Learn more
If your program requires significant customization, you can make REST API requests directly, but you must manually write code to calculate the signature. For more information, see Signature Version 4 and DoMetaQuery API.