pigieligi's picture
Update utils.py
1f1c259
Raw History Blame Contribute Delete
5.5 kB
import pickle
import requests
import json
import numpy as np
def get_api_key():
with open('hopsworks_api_key.txt', 'r') as f:
key_text = f.read()
return key_text
def load_headers_dict():
with open('headers.pkl', 'rb') as f:
headers_loaded = pickle.load(f)
return headers_loaded
def get_movie_by_title(title):
headers = load_headers_dict()
url = f"https://api.themoviedb.org/3/search/movie?query={title}"
response = requests.get(url, headers=headers)
response_dict = json.loads(response.text)
return response_dict
def get_movie_details(movie_id):
headers = load_headers_dict()
url = f"https://api.themoviedb.org/3/movie/{movie_id}"
response = requests.get(url, headers=headers)
response_dict = json.loads(response.text)
return response_dict
def get_movies_list(lte_rd : str, gte_rd : str, pages : int = 1, first : int = 0):
headers = load_headers_dict()
all_results = []
# Check that the range of pages doesn't reach the max number of the request
url = f"https://api.themoviedb.org/3/discover/movie?page=1&primary_release_date.gte={gte_rd}&primary_release_date.lte={lte_rd}&include_adult=false&include_video=false&language=en-US&sort_by=revenue.desc"
tt_pages = requests.get(url, headers=headers).json()['total_pages']
if tt_pages > 500:
print('---------')
print('/!\ Number of pages greater than 500!! All films will not be accessible.')
print(f'Number of pages : {tt_pages}')
print('---------')
print()
if first + pages > tt_pages:
raise ValueError(f'The request does not have enough pages for this range. Number of page : {tt_pages}.')
# Fetch all pages
for i in range(1+first, first+pages+1):
print(f'Downloading page {i} ------')
url = f"https://api.themoviedb.org/3/discover/movie?include_adult=false&include_video=false&language=en-US&sort_by=revenue.desc&page={i}"
response = requests.get(url, headers=headers)
response_dict = json.loads(response.text)
results = response_dict['results']
all_results.extend(results)
return all_results
def get_keywords_related_films_average_score(movie_id):
headers = load_headers_dict()
url = f"https://api.themoviedb.org/3/movie/{movie_id}/similar?language=en-US&page=1&sort_by=popularity.desc"
response = requests.get(url, headers=headers)
avg_revenue = list()
for film in response.json()['results']:
movie_id = film['id']
movie_details = get_movie_details(movie_id)
if 'revenue' in movie_details.keys():
revenue = movie_details['revenue']
if revenue > 0:
avg_revenue.append(revenue)
if len(avg_revenue) > 0:
return sum(avg_revenue) / len(avg_revenue)
else:
return 0
def get_movie_cast(movie_id):
headers = load_headers_dict()
url = f"https://api.themoviedb.org/3/movie/{movie_id}/credits"
response = requests.get(url, headers=headers)
response_dict = json.loads(response.text)
return response_dict
def get_top3_genres_features(genres_list):
nb = 0
genres_id = list()
while len(genres_list) > 0 and nb < 3:
id = genres_list.pop(0)['id']
genres_id.append(id)
nb += 1
# Complete the 3-uplet with -1 corresponding to no genres
while len(genres_id) < 3 :
genres_id.append(-1)
return genres_id
def crew_popularity(movie_id):
headers = load_headers_dict()
url = f"https://api.themoviedb.org/3/movie/{movie_id}/credits?language=en-US"
crew_list = requests.get(url, headers=headers).json()['crew']
peoples = set()
for p in crew_list:
if p['job'] in ['Screenplay', 'Director', 'Story', 'Producer', 'Production Manager', 'Executive Producer', 'Casting']:
# To avoid counting many times the same man
peoples.add((p['popularity'], p['name']))
total_pop = 0
for pop, _ in peoples:
total_pop += pop
return total_pop
def get_latest():
headers = load_headers_dict()
url = "https://api.themoviedb.org/3/movie/latest?language=en-US"
movie_response_dict = requests.get(url, headers=headers)
return movie_response_dict
def extract_features(movie_id, latest=False):
movie_details = get_movie_details(movie_id)
# As we sort the request by his revenue, we can stop at the first occurence of a film with no revenues.
if not latest and movie_details['revenue'] == 0:
print('Last film with revenue saved for this request (no need to fetch more pages).')
return -1
# The budget of the film is a super relevant feature, if we do not know the budget, better to ditch out the film
if movie_details['budget'] > 0:
# Avg similar films revenues
movie_details['similar_revenues'] = get_keywords_related_films_average_score(movie_id)
# Crew popularity
movie_details['crew_popularity'] = crew_popularity(movie_id)
# Top10 movie_cast pop
cast = get_movie_cast(movie_id=movie_id)
popularities = [person['popularity'] for person in cast['cast']]
total_popularity = np.sum(np.sort(popularities)[-10:])
movie_details['top_cast_popularity'] = total_popularity
return movie_details
else:
print('Budget missing!')
return None
def load_model():
filename = 'finalized_modelk7.sav'
model = pickle.load(open(filename, 'rb'))
return model