Enable context instance level persistence of graphs

[shashankpr]

The breakdown of tasks for this issue is as follows:

• [x] Implementation of multiple graph instance objects per meeting using hash map of contextId:instanceID: nx.Graph(graphID) (This will be in-memory)
• [x] Add utility function to write and read graph objects to pickle format. (Supported networkx persistence: https://networkx.github.io/documentation/latest/reference/readwrite/index.html)
• [x] Add utility / common package to handle S3 uploads and downloads of the pickled graph object. Graph object will be stored in this filename format - contextId:instanceId
• [x] Decide when to write graph to pickle and upload (Every n minutes or nats topic etc.)
• [x] Explore Amazon Neptune and Dynamo DB for persisting in DB
• [x] Request scoped persistence of graphs using a Repository pattern.

Reading and writing graphs — NetworkX 2.4rc1.dev20190613214358 documentation

[karthikmuralidharan]

Also, the persistence needs to be request scoped instead of dumping at the end of the meeting.

[shashankpr]

@karthikmuralidharan , I will implement the persistence of graph for every add_segments topic. It will read and write with every request scope of add_segments.