Merge pull request #164 from davetownG/Issue-127-ChatSystem-Needs-README

Issue 127 chat system needs readme

Author: tebanieo

schema_design/SchemaExamples/ChatSystem/README.md

@@ -0,0 +1,80 @@
+# Chat System Data Modeling with Amazon DynamoDB
+
+## Overview
+
+This document outlines a use case using DynamoDB as a datastore for a chat system. For example, it is intended to be used within a team of several people in a game, or in a live video streaming situation where comments are added to the video. The flow of the system is as follows: create a room, join a room, speak in a room, leave a room, and delete a room.
+
+## Key Entities
+
+1. user
+2. room
+3. comment
+
+## Design Approach
+
+We employ a single table design coupled with a global secondary index (GSI). 
+The following key structures are used:
+
+  - Base table 
+    - For a user item:
+      - Partition key (PK)
+        - User:\<UserID\>
+      - Sort key (SK)
+        - \<timestamp\>
+    - For a room item:
+      - Partition key (PK)
+        - Room:\<RoomID\>
+      - Sort key (SK)
+        - "meta"
+
+    - Examples:  
+
+      | PK | SK | Sample Attributes |
+      | ----------- | ----------- | ----------- |
+      | User:UserA | 2023-04-01T14:00:00.001Z | RoomID, Comment, CreatedAt |
+      | Room:Art | `meta` | CreatedBy |
+
+  - GSI
+    - Partition key (RoomID)
+      - \<RoomID\>
+    - Sort key (CreatedAt)
+      - \<timestamp\>
+
+    - Example:  
+
+      | PK | SK | Sample Attributes |
+      | ----------- | ----------- | ----------- |
+      | Music | 2023-04-01T12:00:00.001Z | PK, SK, Comment |
+
+
+## Access Patterns
+
+The document covers 9 access patterns. For each access pattern, we provide:
+- Usage of Base table or GSI
+- Relevant DynamoDB operation (PutItem, DeleteItem, Query)
+- Partition and Sort key values
+- Other conditions or filters
+
+  | Access pattern | Base table/GSI | Operation | Partition key value | Sort key value | Other conditions/Filters |
+  | ----------- | ----------- | ----------- | ----------- | ----------- | ----------- |
+  | createChatRoom | Base table | PutItem | PK=\<RoomID\> | SK="meta" | if not exists |
+  | deleteChatRoom | Base table | DeleteItem | PK=\<RoomID\> | SK="meta" | CreatedBy=\<UserID\> |
+  | joinChatRoom | Base table | PutItem | PK=\<UserID\> | SK="Join:"\<roomID\>  | |
+  | leaveChatRoom | Base table | DeleteItem | PK=\<UserID\> | SK="Join:"\<roomID\>  | |
+  | addComments | Base table | PutItem | PK=\<UserID\> | SK=\<timestamp\> | |
+  | getAllComments | GSI | Query | PK=\<RoomID\> | | Limit 1 |
+  | getLatestComments | GSI | Query | PK=\<RoomID\> | | Limit 10 & ScanIndexForward = false |
+  | getFromLatestToSpecifiedPositionComments | GSI | Query | PK=\<RoomID\> | SK > \<FromPosition\> | |
+  | getFromPositionToPositionComments | GSI | Query | PK=\<RoomID\>  | SK between \<FromPosition\> and \<ToPosition\> | |
+
+  
+Please note: We add “Limit 1” for getAllComments since GSIs can have duplicate values. GSIs do not enforce uniqueness on key attribute values like the base table does.
+
+## Goals
+
+- Model relationships between users and posts efficiently
+- Ensure scalability using Amazon DynamoDB's single table design principles
+
+## Schema Design
+
+A comprehensive schema design is included, demonstrating how different entities and access patterns map to the DynamoDB table structure. [ChatSystemSchema.json](https://github.com/aws-samples/aws-dynamodb-examples/blob/master/schema_design/SchemaExamples/ChatSystem/ChatSystemSchema.json)