Product Requirement Document - LinDong123a/2017-2018-computing-thinking GitHub Wiki
Revision History:
| Version | Summary | Date | By |
|---|---|---|---|
| 0.1 | The start of the document to include all project requirements | 04/06/2017 | yangboz |
| 0.2 |
General Description
MutiChain based Dockerized Microservices include tensorflow powered AI and Python front-end user interface. 区块链承载的可编排，资源调度，弹性伸缩的容器化平台提供基于人工智能围棋相关的计算机思维一系列微服务.
Feature Perspective
This PRD is about getting spatial views production ready. Hence it can be considered a new feature, though from the coding perspective it's an improvement as the spatial views are already part of Couchbase since 2.0.
User Perspective
The user will use the spatial views the same way as the mapreduce views. Either through the Web UI or with and SDK.
Acceptable Constraints
The spatial views won't be as fast as the mapreduce views.
Target Release
0.0.1
Requirements
General Requirements
- Multidimensional index/queries
- Indexing geometries
UI Requirements
The UI is already there, but it could be improved to the level of the mapreduce views. Things the mapreduce views interface does and the spatial doesn't:
- Show syntax errors when a function is saved
- Display that there are no results when there were no results
Currently you need to create a mapreduce view before you can create a spatial view. This might be something to change.
Performance Requirements
Stability is the main focus for now. Performance comes later.
Behaviour on Topology changes
The expected behaviour on topology changes is exactly the same as for mapreduce views.
Failure mode behaviour
The expected failure mode is the same as for mapreduce views.
Possible Future Enhancements
- Geometry search
- More topological predicates for querying. I.e. not only query for everything intersecting a certain bounding box/geometry, but also be able to query for full containment or outside of some predicate.
- Support for values other than numbers
Testplan
Design Document
Workflow
Status
Implementation
GA Release
Original spatial views in 2.0. The multidimensional ones in 3.0. The production ready ones in Sherlock.