MovieMash is a website that generates a list of movies for your next movie marathon or when you are in a bingey mood! Input the amount of time you have and other options and get a list of awesome movies!
Install Python3. I would also recommend using Virtualenv to keep everything isolated for this project.
Go to project directory and enter virtualenv env.
Then, use source env/bin/activate to enter the environment.
Command is different for Windows users on CMD or PowerShell
Install the necessary packages: pip install -r requirements.txt
To run the server for development, use python main.py.
To run the server for production, use gunicorn --bind 0.0.0.0:5050 wsgi
Use whatever port and location you need
Visit the page in your browser and enjoy!
To seed the database, make sure you are in the root directory of the project have movies.tsv with UTF-8 encoding in that directory.
python seed.py
will seed the database in batches of 1000 movies. For each batch, the entire batch is scraped for synopses before all 1000 are inserted into elasticsearch at once.
To evaluate the use of time for the project, there is a script that uses somewhat random but vague queries, and identifies the difference between the time input and the filled up time.
python run_time_eval.py
We also have the following bash scripts for running the evaluations on queries to get NDCG-Partial:
./run_queries_eval.sh
./run_keywords_eval.sh
where the latter writes the results to the file keywords_results.tsv. To evaluate your own query, you can enter:
echo <query> | python evaluations.py
To see the NDCG-Partial scores.
If adding dependencies to the project, make sure to update the requirements.txt with the correct package name and version number. You can use pip freeze | grep <name> to get the correct format with version.
The scraper relies on connecting to an Elasticsearch instance. You can use Docker to create a single-node instance or install manually. I would also recommend getting Kibana as well to interact with and visualize Elasticsearch
If using Mac or Linux, try using
start-docker.shandstop-docker.shfor starting and stopping the containers for both Elasticsearch and Kibana. It will also handle install
For this project, we use an Elasticsearch instance on AWS
All static files, such as JS and CSS, go into the static/ directory.
All Jinja templates for the webpage go into the templates/ directory.
All other code for ranking and logic goes into the src/ directory.
Scripts for seeding and evaluating and the input files for them are all in the root directory for the project / and should be run from there.