sdl60660 / letterboxd_recommendations Goto Github PK

This should make the setup/run a lot more straightforward

the recommendation list is showing movies that I've already seen

It's arguably easier to mentally filter results from poster alone, for example, ignoring movies you've already seen. I suggest a grid or gallery view in 5 cols x 10 rows since the search returns 50 results.

Can't seem to scrape my profile's movie on any device or network. Seems to be logging "redis_get_user_data_job_status" as failed.

My tastes have changed significantly over the years, and it would be nice to have the option for your engine to recommend based on only movies I have rated in the last few years, for example.

Hello, I've tried the website today and the genres blacklist section doesn't work, for example I asked for no romance and animation movies, and my recommandations were full of romance movies and some animations too

The date thing does work fine though

Hi Sam, as per our mail, I think the usefulness of the tool could be increased a lot by having the results be between a decade of the users choosing, as well as excluding all films of a certain 'meta' genre, those being documentary, and animation. I know movielens.org suffers from this same issue, and will often end up recommending you things like 'stand up' comedy specials or documentaries, even though your preferences are clearly for fiction films.

I don't think its necessary to have every genre listed, but I do think documentary and animation should definitely be their own thing when generating a list of results. As for the decade, it makes it a lot easier when doing any serious kind of film research and wanting to get a good vibe for what the highest rated films looked or sounded like for that specific point in time.

Would it be possible to redistribute snapshots of the dataset so that there wouldn't need to be duplicated scraping? e.g. via automated torrents or something

The download I just did from your engine included the row: https://letterboxd.com/film/film:590400/ which when uploaded to Letterboxd, had no match. I went back to the list, still open on your site, and it was identified as [Movie poster for Das Boot] [Das Boot (1987)] https://letterboxd.com/film/film:590400/) 8.36 – so presumably the TV series was briefly in Letterboxd, before being deleted by them.

As the additional identifying data was in your cache anyway, would you consider an option to include it in the download? It's not uncommon for obscure but interesting films to disappear from Letterboxd, whether because they're ‘video’, or have never had a theatrical release or festival screening.

• Add material UI accordion to Result component with more details retrieved from tmdb
  (Data is already in the mongo database and being returned to the client, just not being surfaced right now)

Since around 2 weeks ago?

• Grab watchlist while scraping for ratings and return to the client
• Add checkbox for filtering out any items already on watchlist and maybe a badge marker on movies in recommendation list