-
Notifications
You must be signed in to change notification settings - Fork 31
API Description
Note: This API description correspond to the latest version VDMS, in the develop branch.
Once the data usage and data collection/preprocessing requirements of an application are understood, it is time to connect the application to a VDMS server because VDMS offers the unique capabilities of storing metadata in a structured graph database as well as takes care of dealing with the actual data through one unified API, described on this page. We provide C++ and Python client libraries for linking with any application to enable VDMS communication.
While our metadata is stored in a graph database, we have tried to simplify the API to a level that hides some of the nitty gritties. It also opens up the possibility of parsing this to a relational backend, though that is not our goal in this project. We now describe the syntax and keywords for all the currently supported commands and indicate what is optional. We also give some examples and some explanations after each example to help understand and implement using our API. These examples try to use some of the schema elements from the example application schemas given earlier.
We define a set of JSON-based API calls that allow an application to interact with VDMS and exploit its strengths. The table below lists the calls provided, followed by their JSON specification. Our commands can be constructed using some basic building blocks and return results with certain meanings.
A user application will use the VDMS API by defining metadata conforming to the query protocol described below and passing along blobs. The client side VDMS library provides a simple query function that accepts a JSON string with commands and an array or vector of blobs. Internally, the library wraps the query string and blob in a protobuf (choice for now) and sends it to VDMS. It also receives a similarly wrapped response from VDMS and returns it to the client. The responses will require JSON parsing on the client side, starting with the metadata string that will indicate how to interpret blobs. We could substitute protobuf here with explicitly sending a string followed by sending binary data OR provide a link within the string block to indicate where to download the actual data from. For now, we will go with the image as a blob in the response message, and have the library in the client unpack the image and place it in a folder. In the future, we will consider adding a flag to indicate that VDMS will return urls and generate copies to download in the server.
Visual Data Management System - Intel Labs
FLINNG Library and Performance
Basic Building Blocks
Insert
- AddBlob
- AddBoundingBox
- AddConnection
- AddDescriptor
- AddDescriptorSet
- AddEntity
- AddImage
- AddVideo
- NeoAdd
Query
- ClassifyDescriptor
- FindBlob
- FindBoundingBox
- FindConnection
- FindDescriptor
- FindDescriptorSet
- FindEntity
- FindFrames
- FindImage
- FindVideo
- NeoFind
Update