Movie lens with mongodb and Django

---

MongoDB document database was chosen to maximize the availability at the expense of consistency, as the data do not need to have a strong consistency and can have a more flexible schema.

ratings, tags, and genome tags were embedded in movie as the information could be retrieves in a single query when needed. There was no need to access any of them as an object on its own and joins/lookups were avoided. In addition, there are just a few hundreds of embedded documents, for this reason embedding was preferred. However, if more and more tags, genome tags, and ratings are to be included, then the database schema must be changed from embedding to referencing.

Simplified database schema

{

"_id": "ObjectId('ObjectId')",
"movieId": "id",
"title": "title",
"genres": ["genre1", "genre2", ...],
"year": "year",
"imdbId": "imdbId",
"tmdbId": "tmdbId",
"ratings": [
        {"userId": "id", "rating": "rating", "timestamp": timestamp}, ...
    ],
"tags": [
        {"userId": "id", "tag": "tag", "timestamp": timestamp}, ...
    ],
"genome_tags": [
        {"tagId": "id", "tag": "tag", "relevance": relevance}, ...
    ]

}

---

Install, configure, and run the application

---

Create environment with the dependencies

conda env create -f environment.yml

Activate environment

conda activate Movie_Lens

Create directory to save database

mkdir -p mongodb/data/db

Run mongoDB

mongod --noauth --dbpath mongodb/data/db

Run python Populate script

usage: Populate_db.py [-h] folder

Populate mongo Database with data from movie lens.

positional arguments: folder A folder with csv files.

optional arguments: -h, --help show this help message and exit

Run server

python Movies_Project/manage.py runserver

---