This is a simple chatbot that briefly converses with the user and finds restaurants based on the conversation from the yelp databse and sends a text message of the recommendations.
This was a part of Cloud Computing course at New York University, and was a way to get used to various AWS services
Following services were used in building this application out
- S3
- API Gateway
- Lambda
- Lex
- SQS
- SNS
- Elastic Search
- DynamoDB
- Yelp API for data sourcing
- S3 is used to host the frontend of the site
- API Gateway is used to connect the frontend with our chatbot and is defined by api.yaml
- Lambda hosts many functions, one of them is entry which acts as an interface between the user and Lex
- lex-validation is another Lambda function that works with Lex to validate input and finally send the gathered information to an SQS queue
- Every minute through CloudWatch events we trigger another Lambda, [recommendations](lambdas/recommendations.py].
- recommendations uses the data from the SQS entry to search for relevant results using Elastic Search and get detailed information about the restaurants from DynamoDB.
- Then it constructs and sends a text message filled with recommendations to the user using SNS
NOTE: Data is first scraped using the Yelp API, stored in DynamoDB for quick access and indexed in Elastic Search for quick search. Development code contains the programs used to accomplish this.