Utilize a phrase database provided below. The desired functionality is to:

1. Receive a sentence and provide a corresponding response mapping for it. The phrases will be of the format:

   catch phrase	keyword
   (0,2) block me (0,1)	BLOCK
   (0,1) kill myself	KILL
   (0,1) can you (0,1) me (0,1) joke (0,0)	JOKE
   Love me	LOVE

   (0,3) means it can allow 0 up to 3 words in the sentence.

   (0,2) are you doing (0, 1) should be able to catch sentences like “Hey what are you doing tonight?”, “What are you doing?” “Hi what are you doing after?”

2. Add phrases to the database (by reading from a csv file), delete, edit the mapping answer for it.

The project has been built on FastAPI. The following assumptions have been made;

1. The words in a regex will be valid regex words. i.e [A-Za-z0-9_]+
2. The words in the sentences may only be followed by 1 punctuation mark in the set (, . : ; ! ?). As such, hi, how are you doing today? is a valid sentence, but hi,, how are you doing today? is not for the input phrase: (0,2) are you doing (0,1).

On creation of a CatchPhrase object, the (minWords, maxWords) segments substituted by a regex and persisted to the Database. When you request to match a particular sentence, the sentence is split into its individual words, and, using the IN operator in SQL a subset of CatchPhrases objects are retrieved from the database. A full regex match is then performed against the records in these subset. The first record to return a full match will be returned as the mapping answer.

The following environment variable is available;

• DB_URL: set it to a valid connection string for a DB you want to connect to. If not provided, it will default to using a SQLLite database in the project directory.

Using Docker, you can run the project using the following command:

docker run -p 8000:8000 ghcr.io/ianminash/catcher:main

• Install the project dependencies from the requirements.txt file using pip or pipenv if using a virtual environment.
• Run the project using uvicorn:

  uvicorn main:app
  

  This will bring up the service on port 8000.