mauviette/notebooks/TRN_recsys.org

💈 Développez un moteur de recommandation de films

Preamble Emacs Setup noexport

(setq org-src-fontify-natively t)

(setq lsp-semantic-tokens-enable t) (setq lsp-enable-symbol-highlighting t)

(setq lsp-enable-file-watchers nil read-process-output-max (* 1024 1024) gc-cons-threshold 100000000 lsp-idle-delay 0.5 ;; lsp-eldoc-hook nil lsp-eldoc-enable-hover nil

;;pas de fil d'ariane lsp-headerline-breadcrumb-enable nil ;; pas de imenu voir menu-list lsp-enable-imenu nil ;; lentille lsp-lens-enable t

lsp-semantic-highlighting t lsp-modeline-code-actions-enable t )

(setq lsp-completion-provider :company lsp-completion-show-detail t lsp-completion-show-kind t)

(setq lsp-ui-doc-enable t lsp-ui-doc-show-with-mouse nil lsp-ui-doc-show-with-cursor t lsp-ui-doc-use-childframe t

lsp-ui-sideline-diagnostic-max-line-length 80

;; lsp-ui-imenu lsp-ui-imenu-enable nil ;; lsp-ui-peek lsp-ui-peek-enable t ;; lsp-ui-sideline lsp-ui-sideline-enable t lsp-ui-sideline-ignore-duplicate t lsp-ui-sideline-show-symbol t lsp-ui-sideline-show-hover t lsp-ui-sideline-show-diagnostics t lsp-ui-sideline-show-code-actions t )

(setq lsp-diagnostics-provider :none lsp-modeline-diagnostics-enable nil lsp-signature-auto-activate nil ;; you could manually request them via `lsp-signature-activate` lsp-signature-render-documentation nil)

Imports

%matplotlib inline %load_ext autoreload %autoreload 2

import sys import os import warnings warnings.filterwarnings("ignore") import pickle

import numpy as np import pandas as pd import matplotlib.pyplot as plt import seaborn as sns

from keras.preprocessing.image import load_img from keras.preprocessing.image import img_to_array from keras.applications.vgg16 import preprocess_input

from keras.applications.vgg16 import VGG16 from keras.models import Model

from sklearn.cluster import KMeans from sklearn.decomposition import PCA

from random import randint import pickle

:results:

:end:

Functions

def display_all(df): with pd.option_context("display.max_rows", 100, "display.max_columns", 100): display(df)

:results:

:end:

Org noexport

import IPython import tabulate

class OrgFormatter(IPython.core.formatters.BaseFormatter): format_type = IPython.core.formatters.Unicode('text/org') print_method = IPython.core.formatters.ObjectName('_repr_org_')

def pd_dataframe_to_org(df): return tabulate.tabulate(df, headers='keys', tablefmt='orgtbl', showindex='always')

ip = get_ipython() ip.display_formatter.formatters['text/org'] = OrgFormatter()

f = ip.display_formatter.formatters['text/org'] f.for_type_by_name('pandas.core.frame', 'DataFrame', pd_dataframe_to_org)

:results:

:end:

The Data Load Data

df = pd.read_csv('/Users/lss/Sites/simplon.ai/briefs/recsys/data/movie_metadata_prepared.csv')

:results: 0 - 06c3c7bc-1cc7-4984-a890-a11f3ea59b :end:

Glimpse at the data

display_all(df.describe(include='all').T)

:results:

	count	unique	top	freq	mean	std	min	25%	50%	75%	max
color	5024	2	Color	4815	nan	nan	nan	nan	nan	nan	nan
director_name	4939	2398	Steven Spielberg	26	nan	nan	nan	nan	nan	nan	nan
num_critic_for_reviews	4993	nan	nan	nan	140.194	121.602	1	50	110	195	813
duration	5028	nan	nan	nan	107.201	25.1974	7	93	103	118	511
director_facebook_likes	4939	nan	nan	nan	686.509	2813.33	0	7	49	194.5	23000
actor_3_facebook_likes	5020	nan	nan	nan	645.01	1665.04	0	133	371.5	636	23000
actor_2_name	5030	3032	Morgan Freeman	20	nan	nan	nan	nan	nan	nan	nan
actor_1_facebook_likes	5036	nan	nan	nan	6560.05	15020.8	0	614	988	11000	640000
gross	4159	nan	nan	nan	4.84684e+07	6.8453e+07	162	5.34099e+06	2.55175e+07	6.23094e+07	7.60506e+08
genres	5043	914	Drama	236	nan	nan	nan	nan	nan	nan	nan
actor_1_name	5036	2097	Robert De Niro	49	nan	nan	nan	nan	nan	nan	nan
movie_title	5043	4917	Ben-Hur	3	nan	nan	nan	nan	nan	nan	nan
num_voted_users	5043	nan	nan	nan	83668.2	138485	5	8593.5	34359	96309	1.68976e+06
cast_total_facebook_likes	5043	nan	nan	nan	9699.06	18163.8	0	1411	3090	13756.5	656730
actor_3_name	5020	3521	John Heard	8	nan	nan	nan	nan	nan	nan	nan
facenumber_in_poster	5030	nan	nan	nan	1.37117	2.01358	0	0	1	2	43
plot_keywords	4890	4760	based on novel	4	nan	nan	nan	nan	nan	nan	nan
movie_imdb_link	5043	4919	http://www.imdb.com/title/tt0232500/?ref_=fn_tt_tt_1	3	nan	nan	nan	nan	nan	nan	nan
num_user_for_reviews	5022	nan	nan	nan	272.771	377.983	1	65	156	326	5060
language	5031	47	English	4704	nan	nan	nan	nan	nan	nan	nan
country	5038	65	USA	3807	nan	nan	nan	nan	nan	nan	nan
content_rating	4740	18	R	2118	nan	nan	nan	nan	nan	nan	nan
budget	4551	nan	nan	nan	3.97526e+07	2.06115e+08	218	6e+06	2e+07	4.5e+07	1.22155e+10
title_year	4935	nan	nan	nan	2002.47	12.4746	1916	1999	2005	2011	2016
actor_2_facebook_likes	5030	nan	nan	nan	1651.75	4042.44	0	281	595	918	137000
imdb_score	5043	nan	nan	nan	6.44214	1.12512	1.6	5.8	6.6	7.2	9.5
aspect_ratio	4714	nan	nan	nan	2.2204	1.38511	1.18	1.85	2.35	2.35	16
movie_facebook_likes	5043	nan	nan	nan	7525.96	19320.4	0	0	166	3000	349000
end

numerical = df_raw.select_dtypes(include='number').columns categorical = df_raw.select_dtypes(exclude='number').columns

print(f"categorical columns are : {', '.join(str(x) for x in categorical)}") print(f"numerical columns are : {', '.join(str(x) for x in numerical)}")

categorical columns are : color, director_name, actor_2_name, genres, actor_1_name, movie_title, actor_3_name, plot_keywords, movie_imdb_link, language, country, content_rating

numerical columns are : num_critic_for_reviews, duration, director_facebook_likes, actor_3_facebook_likes, actor_1_facebook_likes, gross, num_voted_users, cast_total_facebook_likes, facenumber_in_poster, num_user_for_reviews, budget, title_year, actor_2_facebook_likes, imdb_score, aspect_ratio, movie_facebook_likes

path = "/Users/lss//Sites/simplon.ai/briefs/recsys/data/posters"

os.chdir(path)

posters = []

with os.scandir(path) as files: # loops through each file in the directory for file in files: if file.name.endswith('.jpg'): # adds only the image files to the flowers list posters.append(file.name)

:results:

:end:

img = load_img(posters[0], target_size=(224,224))

img = np.array(img)

print(img.shape)

:results:

:end:

reshaped_img = img.reshape(1,224,224,3) print(reshaped_img.shape)

:results:

:end:

Preprocessing

x = preprocess_input(reshaped_img)

:results:

:end:

The Model Architecture

Nous pouvons maintenant charger le modèle VGG et supprimer la couche de sortie manuellement. Cela signifie que la nouvelle couche finale est une couche entièrement connectée avec 4 096 nœuds de sortie. Ce vecteur de 4 096 nombres est le vecteur de caractéristiques que nous utiliserons pour regrouper les images.

model = VGG16()

model = Model(inputs=model.inputs, outputs=model.layers[-2].output)

:results:

:end:

features = model.predict(x, use_multiprocessing=True)

:results:

:end:

print(features.shape) features

:results:

  array([[0.4690976, 0.       , 0.       , ..., 0.       , 0.       ,
  3.0501497]], dtype=float32)

:end:

Bibliography References

bibliographystyle:unsrt bibliography:recsys.bib

Local Variables noexport