GolangElasticAPI

About

The repository aims to provide interation of Golang with ElasticSearch DB. The thought used in building the APIs is based on providing relevant NewsBlocks (referred as cards in this project) for a user based on his/her geo-location.

This repository has four main folders and a few data generation scripts:

• Data Contains all the raw data we need to populate our elastic search indexes
• handlers Contains the API definition and corresponding request handlers
• models Contains the structural objects used in this project and DB connection details
• util Contains all the helper function that we need for functioning of our services

Installation

Install Golang and ElasticSearch

Usage

• Clone the repository
• Run the elastic server(Port: 9200)
• Create index users using the following structure:

{
      "settings": {
            "number_of_shards": 1,
            "number_of_replicas": 1,
            "analysis": {
                  "analyzer": {
                        "not_analyzed": {
                              "type": "custom",
                              "tokenizer": "standard"
                        }
                  }
            }
      },
      "mappings": {
            "properties": {
                  "user": {
                        "type": "text",
                        "analyzer": "not_analyzed"
                  },
                  "createdAt": {
                        "type": "date"
                  },
                  "state": {
                        "type": "text"
                  },
                  "latitude": {
                        "type": "float"      
                  },
                  "longitude": {
                        "type": "float"
                  },
                  "subDistrictCode": {
                        "type": "text",
                        "analyzer": "not_analyzed"
                  }
            }
      }
}

• Create index cards using the following structure:

{
      "settings": {
            "number_of_shards": 1,
            "number_of_replicas": 1,
            "analysis": {
                  "analyzer": {
                        "not_analyzed": {
                              "type": "custom",
                              "tokenizer": "standard"
                        }
                  }
            }
      },
      "mappings": {
            "properties": {
                  "cardId": {
                        "type": "text",
                        "analyzer": "not_analyzed"
                  },
                  "createdAt": {
                        "type": "date"
                  },
                  "title": {
                        "type": "text"
                  },
                  "state": {
                        "type": "text"
                  },
                  "latitude": {
                        "type": "float"      
                  },
                  "longitude": {
                        "type": "float"
                  },
                  "subDistrictCode": {
                        "type": "text",
                        "analyzer": "not_analyzed"
                  },
                  "createdBy": {
                        "type": "text",
                        "analyzer": "not_analyzed"
                  }
            }
      }
}

• Run the data creation scripts UserCreation.py and CardsCreation.py for data generation
• Go ahead and run the following command from the project root directory:

go run main.go

This will start the backend server at localhost:8080

To test the behavior out we have five APIs:

• Update user's geo-position
  • endPoint: /changePosition/{userID}/{latitude}/{longitude}
• Get all the users belonging to the particular sub-district (prefix query)
  • endPoint: /allUsersForSDCode/{subDistrictCode}
• Get all the cards for a particular user
  • endPoint: /allCardsForUser/{userID}/
• Get top relevant cards for the particular user based on geo-position
  • endPoint: /topCards/{latitude}/{longitude}
• Get nearest users for the particular user based on geo-position
  • endPoint: /topUsers/{latitude}/{longitude}

Go ahead and test it and let me know your thoughts over this.

Updates

Will be grooming this repository more with further developments. Cheers!