search

practical-recommender-systems / moviegeek

A django website used in the book Practical Recommender Systems to illustrate how recommender algorithms can be implemented.
MIT License
901 stars 360 forks source link

Movie are not getting recommended #33

Closed saadsaw closed 4 years ago

saadsaw commented 4 years ago

I cloned the repo, installed PostgreSQL, configured settings.py to point at my PSQL installation, setup a TMDb API key and added the key with the appropriate format into a new .prs file.

I did "makemigrations" and "migrate --run-syncdb", both of them went through successfully, I even double checked through pgAdmin and all of the tables were created successfully.

Afterwards I ran "populate_moviegeek.py", the script threw out an error due to a size difference between one of the columns in the dataset and the DB, thus I resorted to use the 100K version of the MovieTweetings dataset, I copied the URL and pasted it as the default parameter for the download_movies function, and rerun the script, everything went through beautifully! All of the movies were added into the Database successfully.

Since I got the 100K version of the MovieTweetings Dataset, I also got the ratings for them, changed the URL inside "populate_ratings.py" and ran the script, everything worked here as well.

Last but not least I ran "populate_logs.py" and it worked as well.

Finally I ran the server using "python manage.py runserver" and I tried accessing the webpage, no issues here, however! No recommendations were being shown! The analytics page didn't work at all! The console showed diffferent kinds of errors, and I tried getting into the details page of a movie maybe I would get a "Movies similar to ..." recommendations, but didn't aswell.

All in all, I am desperately trying to get the script to work but I can't!

@kimfalk your help would be greatly appreciated!

My current python version: 3.7.7 Windows 10 Pro x64 All of the packages installed successfully using "pip install -r requirements.txt"

kimfalk commented 4 years ago

Hi Saad Sawash,

This site is described in the book Practical Recommender Systems, you should refer to that. But if you are looking to make the recommendations work have a look in the builder folder where all the algorithms for recommendations are.

The scripts to get the movietweetings should be working though so I will have a look at that.

saadsaw commented 4 years ago

Thanks for the quick resposne @kimfalk

I actually have bought your book, but still haven't started reading through it, I wanted to test out the website and see the recommendations in action before diving in.

Waiting for your update on the MovieTweetings issue.