Tag-Based Photo Search With Dropbox’s File Properties API
Dropbox has published a new tutorial that extends its earlier photo gallery sample app with tag-based search. The project combines the Microsoft Azure Face Recognition API with Dropbox’s File Properties API and JavaScript SDK to let users find pictures of specific people across a gallery.
The tutorial is available here, with full source code on GitHub.
How the Pieces Fit Together
The workflow relies on three main services:
- Microsoft Azure Face Recognition API — analyzes each image and identifies the people in it.
- Dropbox File Properties API — stores the resulting tags as metadata on the files, and handles search queries against those tags.
- Dropbox JavaScript SDK — glues the front end to the Dropbox platform, cutting down the amount of code needed for API calls and handling pagination of results when galleries grow large.
The File Properties API is one of Dropbox’s newer developer offerings. It allows third-party applications to attach property tags to files stored in Dropbox. In this tutorial, those properties correspond to the faces detected in each image, making it possible to search the entire gallery for a particular person.
Pagination and Scale
One of the more useful implementation details is the handling of large result sets. The tutorial demonstrates how to paginate through search results using the JavaScript SDK, which avoids performance problems when a gallery contains many files. This makes the tag-search feature practical beyond small demo collections.
