-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmain.py
319 lines (274 loc) · 13.2 KB
/
main.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
# -*- coding: utf-8 -*-
"""
Created on Sat Dec 1 12:07:56 2018
@author: bjtur
"""
import pandas as pd
import numpy as np
import gzip
import os
import urllib.request
import string
import re
import requests
from bs4 import BeautifulSoup
import math
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import linear_kernel
import matplotlib.pyplot as plt
if not os.path.isfile('group9-projFileInput.csv'):
if not os.path.isfile('rotten_tomatoes.csv'):
print('Scraping movie critic reviews from Rotten Tomoatoes...')
def getMovies(criticUrl):
r = requests.get(criticUrl)
criticHtml = r.text
bsyc2 = BeautifulSoup(criticHtml, "lxml")
temp = bsyc2.select('section[id="criticsReviewsChart_main"] > div > div > ul > li')[2].getText()
temp2 = [int(s) for s in temp.split() if s.isdigit()]
numPages = math.ceil(int(temp2[2])/50)
moveRatingsMap = {}
movieList = []
ratingList = []
for i in range(1,numPages + 1,1):
criticUrl = criticUrl + '?page=' + str(i)
r = requests.get(criticUrl)
criticHtml = r.text
bsyc2 = BeautifulSoup(criticHtml, "lxml")
temp2 = bsyc2.select('table[class="table table-striped"] > tr')
for a in temp2:
temp3 = a.select('td[class="center"]')[0].getText()
movieList.append(temp3)
for b in temp2:
temp4 = b.select('td > span["class"]')[0].get('class')[2]
ratingList.append(temp4)
#print(len(movieList))
#print(len(ratingList))
for index in range(len(movieList)):
moveRatingsMap[movieList[index]] = ratingList[index]
return moveRatingsMap
criticsNameURLMap = {}
baseUrl = 'https://www.rottentomatoes.com'
for alphabet in list(string.ascii_lowercase):
url = 'https://www.rottentomatoes.com/critics/authors?letter=' + alphabet
r = requests.get(url)
html = r.text
bsyc = BeautifulSoup(html, "lxml")
critics_table_list = bsyc.findAll('table',
{ "class" : "table table-striped borderless" } )
for alphabetCritics in critics_table_list:
alphabetCriticsList = alphabetCritics.findAll('a',{"href" : re.compile(r'/critic/')})
for alphabetCritic in alphabetCriticsList:
criticUrl = alphabetCritic.get('href')
criticUrl = baseUrl + criticUrl + '/movies'
criticsNameURLMap[alphabetCritic.contents[0]] = criticUrl
# i = 0
dataList = []
dataList.clear()
for k, v in criticsNameURLMap.items():
# print(k, v)
tempDict = getMovies(v)
for a1, b1 in tempDict.items():
# print(a1,b1)
a11 = a1[1:len(a1)-7]
a12 = a1[-5:-1]
a13 = 1 if b1=='fresh' else 0
dataList.append([k,a11,a12,a13])
# i += 1
# print(i)
# if (i==10):
# break
df = pd.DataFrame(dataList, columns = ['Critic', 'Movie', 'Year', 'Binary'])
pd.set_option('display.max_rows', len(df))
df.to_csv('rotten_tomatoes.csv')
pd.reset_option('display.max_rows')
print('Movie critic reviews successfully scraped!')
if not os.path.isfile('title.basics.tsv.gz'):
print('Downloading movie details from IMDB...')
url = r'https://datasets.imdbws.com/title.basics.tsv.gz'
urllib.request.urlretrieve(url,'title.basics.tsv.gz')
print('Movie details successfully downloaded!')
if not os.path.isfile('title.ratings.tsv.gz'):
print('Downloading movie ratings from IMDB...')
url = r'https://datasets.imdbws.com/title.ratings.tsv.gz'
urllib.request.urlretrieve(url,'title.ratings.tsv.gz')
print('Movie ratings successfuly downloaded!')
print('Merging datasets...')
with gzip.open('title.basics.tsv.gz', "rt", newline='', encoding='utf8') as fileIn:
imdbMovies = pd.read_csv(
fileIn,
delimiter='\t',
na_values = ['\\N'],
encoding='utf8',
quoting=3)
imdbMovies = imdbMovies[imdbMovies.titleType=='movie']
imdbMovies = imdbMovies.rename(columns={'primaryTitle':'Movie','startYear':'Year'})
imdbMovies = imdbMovies.drop(['titleType','originalTitle','isAdult','endYear','runtimeMinutes'],axis=1)
fileIn.close()
with gzip.open('title.ratings.tsv.gz', "rt", newline='', encoding='utf8') as fileIn:
imdbMovieRatings = pd.read_csv(
fileIn,
delimiter='\t',
na_values = ['\\N'],
encoding='utf8',
quoting=3)
fileIn.close()
rt = pd.read_csv('rotten_tomatoes.csv')
imdb = pd.merge(imdbMovies,imdbMovieRatings, on='tconst')
merge = pd.merge(imdb, rt, on=['Movie','Year'])
merge.to_csv('group9-projFileInput.csv',index=False)
print('Datasets successfully merged!\n')
ds = pd.read_csv('group9-projFileInput.csv')
dsDistinct = ds.drop_duplicates(['Movie','Year'], keep='first')
####user input
dsDistinct.reset_index(drop=True, inplace=True)
print("Welcome to Jasper Movie Recommendation Engine\n")
while(True):
name = input("Enter a movie name: ")
searchResult = dsDistinct[dsDistinct.Movie==name.title()]
if searchResult.empty==True:
print("####[ERROR]#### Your movie could not be found. Please try again.")
continue
if searchResult.Movie.count()==1:
break
searchResult = searchResult.sort_values(by='Year')
searchResult = searchResult.reset_index(drop=True)
while(True):
print("\nThe following movies were found with the same name:")
for index, row in searchResult.iterrows():
s = '{:2}. {} ({})'.format((index+1), row['Movie'], int(row['Year']))
print(s)
try:
selectedIndex = int(input("Enter an index from the list above to select your movie: "))
if selectedIndex>0 and selectedIndex<=searchResult.Movie.count():
print('Your selection: ' + str(searchResult.iloc[selectedIndex-1][1]) + " (" + str(int(searchResult.iloc[selectedIndex-1][2])) + ")")
searchResult = searchResult.iloc[selectedIndex-1]
break
else:
raise Exception
except:
print("\n####[ERROR]#### Invalid selection. Try again")
continue
break
print("Your movie was found!\n")
print("Running recommendation engine....\n")
#'data' is a df merge between imdb and rt data
#'searchResult' is a df containing one record from 'data' of the movie the user has entered
####cosine similarity analysis
dsSimilarity = ds.head(2500)
tf = TfidfVectorizer(analyzer='word', ngram_range=(1, 3), min_df=0, stop_words='english')
tfidf_matrix = tf.fit_transform(dsDistinct['genres'].values.astype('U'))
# x = v.fit_transform(df['Review'].values.astype('U'))
cosine_similarities = linear_kernel(tfidf_matrix, tfidf_matrix)
results = {}
for idx, row in dsDistinct.iterrows():
similar_indices = cosine_similarities[idx].argsort()[:-100:-1]
similar_items = [(cosine_similarities[idx][i], dsDistinct['tconst'][i]) for i in similar_indices]
# First item is the item itself, so remove it.
# Each dictionary entry is like: [(1,2), (3,4)], with each tuple being (score, item_id)
results[row['tconst']] = similar_items[1:]
def item(tconst):
return dsDistinct.loc[dsDistinct['tconst'] == tconst]['Movie'].tolist()[0].split(' - ')[0]
def recommend(item_id, num):
print("Recommending movies similar to " + item(item_id) + "...")
print("{:->130}".format(''))
recs = results[item_id][:num]
return pd.DataFrame(recs)
# Just plug in any item id here (1-500), and the number of recommendations you want (1-99)
# You can get a list of valid item IDs by evaluating the variable 'ds', or a few are listed below
# =============================================================================
myMovieId = searchResult.iloc[0][1]
recsdf = recommend(item_id=myMovieId, num=100)
recsdf.rename(columns={0:"Recommended Score", 1:"tconst"}, inplace=True)
aggregatedMovieCriticRating = ds.groupby('tconst')['Recommends'].mean()
movieCriticCounts = ds.groupby('tconst')['Recommends'].count()
aggregatedUserCriticRating = ds.groupby('tconst')['averageRating'].first()
userRatingCounts = ds.groupby('tconst')['numVotes'].first()
aggregatedMovieCriticRatingDf = aggregatedMovieCriticRating.to_frame()
movieCriticCountsDf = movieCriticCounts.to_frame()
aggregatedUserCriticRatingDf = aggregatedUserCriticRating.to_frame()
userRatingCounts = userRatingCounts.to_frame()
mergedRatings = pd.merge(aggregatedMovieCriticRatingDf,movieCriticCountsDf,how='left', on='tconst')
mergedRatings = pd.merge(mergedRatings, aggregatedUserCriticRatingDf,how='left', on='tconst')
mergedRatings = pd.merge(mergedRatings, userRatingCounts,how='left', on='tconst')
mergedRatings.rename(columns={'Recommends_x':'Average Critic Rating',
'Recommends_y':'Number of Critic Ratings',
'averageRating':'Average User Rating',
'numVotes':'Number of User Ratings'},
inplace=True)
mergedRatings['Average User Rating'] = mergedRatings['Average User Rating']/10
mergedRatings['Combined Rating'] = (0.7 * (mergedRatings['Average Critic Rating'] ) + 0.3 * ((mergedRatings['Average User Rating'])))
mergedRatings = mergedRatings.sort_values(by=['Combined Rating'], ascending=False)
combinedData = pd.merge(recsdf, mergedRatings, how="left", on="tconst")
combinedData["Final Rating"] = combinedData['Recommended Score'] * combinedData['Combined Rating']
combinedData = combinedData.sort_values(by=['Final Rating'], ascending=False)
top50Rec = combinedData.head(50)
MovieName = ds[['tconst','Movie']]
MovieName = MovieName.drop_duplicates(subset ="Movie", keep = 'first')
top50Rec_Final = pd.merge(top50Rec, MovieName,how = 'left', on = 'tconst' )
top50Rec_Final.shape
top50Rec_Final.drop_duplicates(subset='Movie', keep = 'first', inplace=True)
top50Rec_Final.sort_values(by = 'Final Rating', ascending= False)
top50Rec_Final = top50Rec_Final[top50Rec_Final.tconst != myMovieId]
top50Rec_Final = top50Rec_Final.dropna(subset=['Movie'])
top20Rec_Final = top50Rec_Final.head(20)
print("{:>30}{:>25}{:>25}{:>25}{:>25}".format('Movie','Average Critic Rating','Average User Rating','Similarity Score','Recommended Rating'))
print("{:->130}".format(''))
for i, row in enumerate(top20Rec_Final.values):
print("{:>30}{:>25.2}{:>25.2}{:>25.2}{:>25.2}".format(row[8],row[2],row[4],row[0],row[7]))
# Functions to generate plots
def Revenue(movieList):
results = []
for movie in movieList:
budget = None
grossing = None
movieCode = movie[0]
movieName = movie[1]
url = "https://www.imdb.com/title/"+movieCode + "/"
review_html = requests.get(url).text
soup = BeautifulSoup(review_html, 'lxml')
for h4 in soup.find_all('h4'):
if "Budget:" in h4:
budget = h4.next_sibling.strip()
budget = int(re.sub(r'[^0-9]','',budget))
if "Cumulative Worldwide Gross:" in h4:
grossing = h4.next_sibling.strip()
grossing = int(re.sub(r'[^0-9.]','',grossing))
if(budget != None and grossing != None):
results.append([movieCode, movieName, budget, grossing])
return results
# box office plot
movieList = np.array([top20Rec_Final['tconst'],top20Rec_Final['Movie'],top20Rec_Final['Average User Rating'],top20Rec_Final['Number of User Ratings']])
userRating = np.array(movieList[2]).astype(float)
totalRatings = np.array(movieList[3]).astype(int)
ratingLabel = np.array(movieList[1])
movieList = movieList.transpose()
results = Revenue(movieList)
arr = np.array(results)
arr = arr.transpose()
label = np.array(arr[1])
budget = np.array(arr[2]).astype(int)
revenue = np.array(arr[3]).astype(int)
ratio = revenue/budget
colors = np.random.rand(len(budget))
index = np.arange(len(ratio))
plt.figure(figsize=(15,10))
plt.barh(index, ratio)
plt.ylabel('Movie', fontsize=12)
plt.xlabel('Worlwide Gross/ Budget', fontsize=12)
plt.yticks(index, label, fontsize=10)
# plt.xticks(index, fontsize=10)
plt.title('Box Office Success', fontsize=20)
plt.show()
# user rating plot
x = ratingLabel
y = userRating
z = totalRatings/200
index = np.arange(len(ratingLabel))
colors = np.random.rand(len(ratingLabel))
plt.clf()
plt.figure(figsize=(15,10))
plt.xticks(index, fontsize=10, rotation=90)
plt.title('User Ratings', fontsize=20)
plt.ylabel('User Ratings', fontsize=10)
plt.scatter(x, y, z, alpha=0.5,c=colors)
plt.show()