class RecSys$()$:

def $\_\_$init$\_\_($self$,$data$)$:

self.data$=$data

self.allusers $=$ list$($self$.$data.users$[\text{'}$uID$\text{'}])$

self.allmovies $=$ list$($self$.$data.movies$[\text{'}$mID$\text{'}])$

self.genres $=$ list$($self$.$data.movies.columns.drop$([\text{'}$mID$\text{'},$ 'title', 'year'$]))$

self.mid$2$idx $=$ dict$($zip$($self$.$data.movies.mID,list$($range$($len$($self$.$data.movies$)))))$

self.uid$2$idx $=$ dict$($zip$($self$.$data.users.uID,list$($range$($len$($self$.$data.users$)))))$

self.Mr$=$self.rating$\_$matrix$()$

self.Mm$=$None

self.sim$=$np$.$zeros$(($len$($self$.$allmovies$),$len$($self$.$allmovies$)))$

def rating$\_$matrix$($self$)$:

$"""$

Convert the rating matrix to numpy array of shape $($#allusers,#allmovies$)$

$"""$

ind$\_$movie $=[$self$.$mid$2$idx$[$x$]$ for x in self.data.train.mID$]$

ind$\_$user $=[$self$.$uid$2$idx$[$x$]$ for x in self.data.train.uID$]$

rating$\_$train $=$ list$($self$.$data.train.rating$)$

return np$.$array$($coo$\_$matrix$(($rating$\_$train, $($ind$\_$user, ind$\_$movie$)),$ shape$=($len$($self$.$allusers$),$ len$($self$.$allmovies$))).$toarray$())$

def predict$\_$everything$\_$to$\_3($self$)$:

$"""$

Predict everything to $3$ for the test data

$"""$

# Generate an array with $3$s against all entries in test dataset

# your code here

return np$.$full$($len$($self$.$data.test$),3)$

def predict$\_$to$\_$user$\_$average$($self$)$:

$"""$

Predict to average rating for the user.

Returns numpy array of shape $($#users,$)$

$"""$

# Generate an array as follows:

# $1.$ Calculate all avg user rating as sum of ratings of user across all movies$/$number of movies whose rating $>0$

# $2.$ Return the average rating of users in test data

# your code here

sum$\_$ratings $=$ np$.$sum$($self$.$Mr$,$ axis$=1)$

total$\_$ratings $=$ np$.$sum$(($self$.$Mr $>0).$astype$($int$),$ axis$=1)$

avg$\_$ratings $=$ sum$\_$ratings $/$ total$\_$ratings

# avg$\_$ratings $=$ np$.$nanmean$($self$.$Mr$,$ axis$=1)$

user$\_$indices $=[$self$.$uid$2$idx$[$x$]$ for x in self.data.test.uID$]$

return avg$\_$ratings$[$user$\_$indices$]$

# pass

def predict$\_$from$\_$sim$($self$,$uid,mid$)$:

$"""$

Predict a user rating on a movie given userID and movieID

$"""$

# Predict user rating as follows:

# $1.$ Get entry of user id in rating matrix

# $2.$ Get entry of movie id in sim matrix

# $3.$ Employ $1$ and $2$ to predict user rating of the movie

# your code here

user$\_$idx $=$ self.uid$2$idx$[$uid$]$

movie$\_$idx $=$ self.mid$2$idx$[$mid$]$

valid$\_$ratings $=($self$.$Mr$[$user$\_$idx$]!=0)$

# print$($self$.$Mr$.$shape, self.sim.shape$)$

sim$\_$ratings $=($self$.$Mr$[$user$\_$idx$])*($self$.$sim$[$:$,$ movie$\_$idx$])$

valid$\_$sim $=($self$.$sim$[$:$,$ movie$\_$idx$])*($valid$\_$ratings$)$

sim$\_$sum $=$ np$.$sum$($valid$\_$sim$)$

return np$.$sum$($sim$\_$ratings$)/($sim$\_$sum if sim$\_$sum $!=0$ else np$.$nan$)$

# movie$\_$rating $=$ self.Mr$[$user$\_$idx$]$

# sim$\_$rating $=$ self.sim$[$movie$\_$idx$]$

#print$($movie$\_$rating$[$movie$\_$rating $>0],$ sim$\_$rating$[$sim$\_$rating $>0])$

#print$($sim$\_$ratings$[$sim$\_$ratings $>0])$

# sim$\_$sum $=$ np$.$sum$($self$.$sim$[$movie$\_$idx, :$])$

# predicted$\_$rating $=$ np$.$sum$($sim$\_$ratings$)/($sim$\_$sum if sim$\_$sum $!=0$ else $1)$

# return np$.$sum$($sim$\_$ratings$)/($sim$\_$sum if sim$\_$sum $!=0$ else $0)$

# return predicted$\_$rating

# def predict$\_$from$\_$sim$($self$,$ uid, mid$)$:

# Predict a user rating on a movie given userID and movieID

# user$\_$index $=$ self.uid$2$idx$[$uid$]$

# movie$\_$index $=$ self.mid$2$idx$[$mid$]$

# user$\_$rating $=$ self.Mr$[$user$\_$index, :$]$

# sim$\_$scores $=$ self.sim$[$movie$\_$index, :$]$

# return np$.$dot$($user$\_$rating, sim$\_$scores$)/$ np$.$sum$($sim$\_$scores$[$user$\_$rating $>0])$

# pass

def predict$($self$)$:

$"""$

Predict ratings in the test data. Returns predicted rating in a numpy array of size $($# of rows in testdata,$)$

$"""$

# your code here

predictions $=[]$

for i in range$($len$($self$.$data.test$))$:

uid $=$ self.data.test.iloc$[$i$][\text{'}$uID$\text{'}]$

mid $=$ self.data.test.iloc$[$i$][\text{'}$mID$\text{'}]$

predictions.append$($self$.$predict$\_$from$\_$sim$($uid$,$ mid$))$

return np$.$array$($predictions$)$

# pass

def rmse$($self$,$yp$)$:

yp$[$np$.$isnan$($yp$)]=3$ #In case there is nan values in prediction, it will impute to $3.$

yt$=$np$.$array$($self$.$data.test.rating$)$

return np$.$sqrt$((($yt$-$yp$)**2).$mean$())$

class ContentBased$($RecSys$)$:

def $\_\_$init$\_\_($self$,$data$)$:

super$().\_\_$init$\_\_($data$)$

self.data$=$data

self.Mm $=$ self.calc$\_$movie$\_$feature$\_$matrix$()$

def calc$\_$movie$\_$feature$\_$matrix$($self$)$:

$"""$

Create movie feature matrix in a numpy array of shape $($#allmovies, #genres$)$

$"""$

# your code here

movie$\_$features $=$ np$.$zeros$(($len$($self$.$allmovies$),$ le