Storage of the graph-based world representation used in RoboComp

My proposal aims to create a new component that allows to store the representation of the robot’s world over time, as well as query it. For this, it is necessary to store the graph structure used to represent the knowledge of the world that the robot possesses. In RoboComp, when robots need to perform complex tasks, its behaviour is based on missions. The mission begin from an initial state and through different transitions reaches a final state, that is, the robot completes the mission. The states are represented by a graph. My proposal is to use a JSON structure to store this graph. This allows its storage as a collection of documents in MongoDB. Consequently, each document could represent a state of the robot. Also, this slot is intended to study the viability of Neo4J for this purpose, as far as I am concerned, the best option would be to use the polyglot persistence, that consists in to use the database best suited for the type of data, in our case MongoDB, and in to use the database best suited for the type of queries, in our case Neo4j. Specifically I think that Neo4j Doc Manager could be our best choice.