Autocomplete Web Service

Flask based web service that suggests words starting with a given prefix. The words are obtained from a pre-configured corpus text file. This was implemented in the context of a python challenge.

DISCLAIMER: This web service can be unsuitable for big corpus, once it loads them entirely in memory.

Requirements

This web service has the following dependencies:

• Flask
• pygtries

Install them in your environment before running the server. You can use the requirements file at the root of the project:

pip install -r requirements.txt

Configuration

The web service is configured using the following environment variables:

• CORPUS_PATH: The path for the corpus text file. It should be a list of words, one by line.

Set the configurations variable before running the server: export CORPUS_PATH=<corpus_file_path>

Execution

To run the server:

1. Set the flask application module, which is implemented in the flaskr directory: export FLASK_APP=flaskr.

2. Run the flask application: flask run.

API

The web service provides the following end-point.

GET /suggestions

Query Parameters:

• q - prefix for the suggestions be be obtained;
• limit - number of results to be obtained;
• offset - position of the first result to be obtained from all suggestions.

Response

The response is a application/json with the following keys:

• total - total number of suggestions for the given prefix;
• limit - number of results being presented;
• offset - position of the first result from all suggestions;
• next - URL for the next page of results.

Example

Request:

GET /suggestions?q=F&limit=10&offset=10

Response:

{
    limit : 10,
    next: "suggestions?limit=10&offset=20",
    offset: 10,
    results: [
        "FaceApp",
        "FaceLock for apps",
        "FaceLOOK for Facebook",
        "Facetune",
        "FaceSwap Face Swap Live",
        "Fast Racing 3D",
        "Fast Reboot",
        "Fast Download Manager",
        "Fast Followers",
        "Fast Facebook Video Downloader",
    ],
    total: 393
}

Tests

All unit tests can be run using pytest from the project's root directory.

pytest