def get_views_features(views, advanced=False):
features = []
with mesure("building the vector space", indent=1):
# build the vector space
for view in views:
row = [
# lat, long
float(view['location'][0]), float(view['location'][1]),
int(view['duration']),
int(view['daytime']),
# split the day in 6 bits of 4 hours.
int(view['daytime']) / 4,
int(view['weekday'])
]
if advanced:
# get information coming from the browsing tree
pass # TODO Need a transformation in convert.py and re-import all data
features.append(row)
return features
def find_profiles_text(algo=None, training_set=None, user=None):
"""Find different user profiles using the TF/IDF metric (Term Frequency /
Inverse Document Frequency).
The stages of the pipeline are: 1. Vectorizer => 2. RandomizedPCA => 3. KMeans
The use of the randomized PCA is useful here to reduce the dimensionality of the
vectors space.
As we lack some data, the dimentionality reduction is made using an already
existing dataset, the 20 newsgroup dataset.
:parm algo: the algorithm to chose. Can be kmeans, meanshift or both (specified
by "all")
:param training_set: the training set to use for the word vectorisation.
The default setting is to use the 20 newsgroup dataset,
it is possible to use the documents by specifying "docs"
"""
# init some vars
if not algo:
algo = "all"
if not training_set:
training_set = "newsgroup"
print "Computing clusters using the TF-IDF scores,"\
" using %s algo and the %s training dataset" % (algo, training_set)
# we first train the pca with all the dataset to have a most representative
# model. Download the dataset and train the pca and the vector only if a
# pickled version is not available (i.e only during the first run).
wide_dataset = docs = None
vec_filename = os.path.join(OUTPUT_PATH, "pickle/vec-%s.pickle" % training_set)
pca_filename = os.path.join(OUTPUT_PATH, "pickle/pca-%s.pickle" % training_set)
pca2d_filename = os.path.join(OUTPUT_PATH, "pickle/pca2d-%s.pickle" % training_set)
with mesure(" loading vectors"):
if os.path.isfile(vec_filename):
vec = _load_obj(vec_filename)
else:
docs = _load_docs(docs, training_set)
vec = Vectorizer().fit(docs) # equivalent to CountVectorizer + TfIdf
_save_obj(vec, vec_filename)
with mesure(" loading PCA"):
if os.path.isfile(pca_filename):
pca = _load_obj(pca_filename)
else:
docs = _load_docs(docs, training_set)
print " reduce the dimentionality of the dataset to 100 components"
# whiten=True ensure that the variance of each dim of the data in the
# transformed space is scaled to 1.0
pca = RandomizedPCA(n_components=100, whiten=True).fit(vec.transform(docs))
_save_obj(pca, pca_filename)
# To visualize the data, we will project it on 2 dimensions. To do so, we
# will use a Principal Component Analysis (as we made in the first steps),
# but projecting on 2 dimensions.
with mesure(" loading PCA 2D"):
if os.path.isfile(pca2d_filename):
pca_2d = _load_obj(pca2d_filename)
else:
docs = _load_docs(docs, training_set)
print " reduce the dimensionality of the dataset to 2 components"
pca_2d = RandomizedPCA(n_components=2, whiten=True).fit(vec.transform(docs))
_save_obj(pca_2d, pca2d_filename)
# Now, go trough the whole resources for each users and try to find user
# profiles regarding TF-IDF
# as the process can take some time, there is a progressbar to keep the user
# updated about the status of the operation
for username in list(db.users.find().distinct('username')):
if user and user != username:
continue
# get all the resources for this user
urls = db.views.find({"user.username": username}).distinct("url")
if not urls:
continue # if we don't have any url for this user, go to the next one!
resources = list(db.resources.find({'url': {'$in': urls },
'blacklisted': False, 'processed': True}))
if not resources:
continue
print "processing %s (%s docs)" % (username, len(resources))
# get the docs content and names
docs = [res['content'] for res in resources]
urls = [res['url'] for res in resources]
# fit the contents to the new set of features the PCA determined
with mesure(" reduce dataset dimensions to 100"):
docs_transformed = pca.transform(vec.transform(docs))
# what we do have now is a matrix with 100 dimentions, which is not really
# useful for representation. Keeping this for later analysis is a good
# thing so let's save this model for comparing profiles against resources
# later
# TODO pickle the kmeans into mongodb ?
# project X onto 2D
with mesure(" reduce dataset dimensions to 2"):
docs_2d = pca_2d.transform(vec.transform(docs))
# run the clustering algorithm
if algo in ["kmeans", "all"]:
with mesure(" kmeans(5)"):
cluster = KMeans(k=5).fit(docs_transformed)
# get_words_from_clusters(cluster, 10, docs, vec)
# print "ngrams for km on %s" % username
# get_n_bigrams_from_clusters(cluster, docs, 5)
plot_2d(cluster, docs_2d, username, "kmeans", "Text-%s" % training_set)
plot_pie(cluster, username, "kmeans", "Text-%s" % training_set)
if algo in ["meanshift", "all"]:
with mesure(" meanshift"):
cluster = MeanShift().fit(docs_transformed)
# print "ngrams for ms on %s" % username
# get_n_bigrams_from_clusters(cluster, docs, 3)
plot_2d(cluster, docs_2d, username, "meanshift", "Text-%s" % training_set)
plot_pie(cluster, username, "meanshift", "Text-%s" % training_set)
if algo in ["affinity", "all"]:
with mesure(" affinity propagation"):
cluster = AffinityPropagation().fit(euclidean_distances(docs_transformed, docs_transformed))
plot_pie(cluster, username, "affinity", "Text-%s" % training_set)
def cluster_users(features=None):
"""Cluster the users, without using information about profiles.
Different features can be used to do so, at least text features and context
features.
"""
training_set="newsgroup"
docs = None
vec_filename = os.path.join(OUTPUT_PATH, "pickle/vec-%s.pickle" % training_set)
pca_filename = os.path.join(OUTPUT_PATH, "pickle/pca-%s.pickle" % training_set)
# get the training set, transform it to N dimensions
with mesure(" loading vectors"):
if os.path.isfile(vec_filename):
vec = _load_obj(vec_filename)
else:
docs = _load_docs(docs, training_set)
vec = Vectorizer().fit(docs) # equivalent to CountVectorizer + TfIdf
_save_obj(vec, vec_filename)
with mesure(" loading PCA"):
if os.path.isfile(pca_filename):
pca = _load_obj(pca_filename)
else:
docs = _load_docs(docs, training_set)
print " reduce the dimentionality of the dataset to 100 components"
# whiten=True ensure that the variance of each dim of the data in the
# transformed space is scaled to 1.0
pca = RandomizedPCA(n_components=100, whiten=True).fit(vec.transform(docs))
_save_obj(pca, pca_filename)
# for each user, get the contents related to him.
users_content = []
users_labels = []
for username in list(db.users.find().distinct('username')):
# get all the resources for this user
urls = db.views.find({"user.username": username}).distinct("url")
if not urls:
continue # if we don't have any url for this user, go to the next one!
resources = list(db.resources.find({'url': {'$in': urls },
'blacklisted': False, 'processed': True}))
if not resources:
continue
print "processing %s (%s docs)" % (username, len(resources))
# get the docs content and names
users_labels.append(username)
users_content.append(" ".join([res['content'] for res in resources]))
with mesure(" vectorise and reduce the dataset dimensions to 100"):
transformed_content = pca.transform(vec.transform(users_content))
# at the end, compute the similarity between users using different metrics
# kmeans 3 clusters
cluster = KMeans(3).fit(transformed_content)
plot_pie(cluster, "all", "kmeans", "text")
plot_2d(cluster, transformed_content, "all", "kmeans", "text")
user_list = [[users_labels[idx] for idx, _ in enumerate(cluster.labels_ == cluster_id) if _] for cluster_id in np.unique(cluster.labels_)]
# compute similarity scores
from ipdb import set_trace; set_trace()
def find_profiles_context(algo=None, user=None):
"""Find profiles based on:
* location of the views
* time of the day of the views
* time of the day
* day of the week
"""
if not algo:
algo = "all"
# get all users
for username in db.users.distinct("username"):
if user and user != username:
continue
urls = db.views.find({"user.username": username}).distinct("url")
resources = []
if not urls:
continue
print "processing %s (%s docs)" % (username, len(urls))
t0 = time.time()
progress = ProgressBar(
widgets=[" building the matrix for %s" % username,
Percentage(), Bar()])
for url in progress(urls):
# get the views related to this user and this url
views = db.views.find({"user.username": username, "url": url})
views = list(views)
indicators = ['average', 'mean', 'median', 'var', 'std']
row = [len(views), sum([int(v['duration']) for v in views])]
# TODO add location
daytimes = []
weekdays = []
for view in views:
daytimes.append(view['daytime'])
weekdays.append(view['weekday'])
for indicator in indicators:
row.append(getattr(np, indicator)(daytimes))
row.append(getattr(np, indicator)(weekdays))
resources.append(row)
resources = np.array(resources)
print "matrix generation took %s" % (time.time() - t0)
# project X on 2D
# print " project the dataset into 2d"
# pca_2d = RandomizedPCA(n_components=2, whiten=True).fit(resources)
# docs_2d = pca_2d.transform(resources)
# run the clustering algorithm
if algo in ["kmeans", "all"]:
with mesure(" kmeans(5)"):
cluster = KMeans(k=5).fit(resources)
plot_2d(cluster, resources, username, "kmeans", "Context")
plot_pie(cluster, username, "kmeans", "Context")
if algo in ["meanshift", "all"]:
with mesure(" meanshift"):
cluster = MeanShift().fit(resources)
plot_2d(cluster, resources, username, "meanshift", "Context")
plot_pie(cluster, username, "meanshift", "Context")
if algo in ["affinity", "all"]:
with mesure(" affinity propagation"):
cluster = AffinityPropagation().fit(euclidean_distances(resources, resources))
# plot_2d(cluster, resources, username, "affinity", "Context")
plot_pie(cluster, username, "affinity", "Context")
def get_profiles_similarity(usernames, N):
"""
Return a matrix of similarity between the users.
:usernames:
The list of usernames in the system
:N:
the number of profiles to find for each user
"""
# all the documents per profile will be stored in this variable
doc_profiles = []
# all the urls for each profiles will be put in this array
urls = []
# For each user, get his views
for username in usernames:
print "processing %s" % username
# don't use generators are we want to access it multiple times, so we
# actually need to store it in memory
views = list(db.views.find({"user.username": username, 'url': {
'$nin': list(db.resources.find({'blacklisted': True}).distinct('url'))
}}))
features = get_views_features(views)
# Run a clustering algorithm on the view
np_features = np.array(features)
#bandwidth = estimate_bandwidth(np_features, quantile=0.3)
#algo = MeanShift(bandwidth=bandwidth).fit(np_features)
# The distribution from the KMeans algorithm is better because we get
# more balanced clusters. MeanShift comes with a lot of clusters with
# less than 2 elements.
with mesure("clustering the context to find %s profiles" % N, indent=1):
algo = KMeans(N).fit(np_features)
# for each cluster, get the matching views
# this means iterating N times (where N is the number of cluster found)
for label in np.unique(algo.labels_):
profile_urls = []
for i, matches in enumerate(algo.labels_ == label):
view = views[i]
if matches and view['url'] not in profile_urls:
profile_urls.append(view['url'])
# save the urls of this profile for later use
urls.append(profile_urls)
resources = db.resources.find({
'url': {'$in': profile_urls}, # get the resources for those urls
'blacklisted': False, 'processed': True})
# Append the contents for this profile together
doc_profiles.append(" ".join([r['content'] for r in resources]))
# train the vectorizer on a big and sparse set of documents
# the vectorizer is loaded from disk to avoid recomputing it each time
with open(os.path.join(OUTPUT_PATH, "pickle", "vecnewsgroup.pickle")) as f:
vec = pickle.load(f)
# Same for the principal component analysis (PCA)
with open(os.path.join(OUTPUT_PATH, "pickle", "pca100-newsgroup.pickle")) as f:
pca = pickle.load(f)
# At this stage, all the documents are stored into memory, sometimes
# more than once for each resource. We want to vectorize them all and thus
# it can take some time.
with mesure("vectorizing %s profiles" % len(doc_profiles)):
vec_profiles = pca.transform(vec.transform(doc_profiles))
# Compute their similarity score
return euclidean_distances(vec_profiles, vec_profiles), urls
def find_profiles_text(algo=None, training_set=None, user=None):
"""Find different user profiles using the TF/IDF metric (Term Frequency /
Inverse Document Frequency).
The stages of the pipeline are: 1. Vectorizer => 2. RandomizedPCA => 3. KMeans
The use of the randomized PCA is useful here to reduce the dimensionality of the
vectors space.
As we lack some data, the dimentionality reduction is made using an already
existing dataset, the 20 newsgroup dataset.
:parm algo: the algorithm to chose. Can be kmeans, meanshift or both (specified
by "all")
:param training_set: the training set to use for the word vectorisation.
The default setting is to use the 20 newsgroup dataset,
it is possible to use the documents by specifying "docs"
"""
# init some vars
if not algo:
algo = "all"
if not training_set:
training_set = "newsgroup"
print "Computing clusters using the TF-IDF scores,"\
" using %s algo and the %s training dataset" % (algo, training_set)
# we first train the pca with all the dataset to have a most representative
# model. Download the dataset and train the pca and the vector only if a
# pickled version is not available (i.e only during the first run).
wide_dataset = docs = None
vec_filename = os.path.join(OUTPUT_PATH,
"pickle/vec-%s.pickle" % training_set)
pca_filename = os.path.join(OUTPUT_PATH,
"pickle/pca-%s.pickle" % training_set)
pca2d_filename = os.path.join(OUTPUT_PATH,
"pickle/pca2d-%s.pickle" % training_set)
with mesure(" loading vectors"):
if os.path.isfile(vec_filename):
vec = _load_obj(vec_filename)
else:
docs = _load_docs(docs, training_set)
vec = Vectorizer().fit(
docs) # equivalent to CountVectorizer + TfIdf
_save_obj(vec, vec_filename)
with mesure(" loading PCA"):
if os.path.isfile(pca_filename):
pca = _load_obj(pca_filename)
else:
docs = _load_docs(docs, training_set)
print " reduce the dimentionality of the dataset to 100 components"
# whiten=True ensure that the variance of each dim of the data in the
# transformed space is scaled to 1.0
pca = RandomizedPCA(n_components=100,
whiten=True).fit(vec.transform(docs))
_save_obj(pca, pca_filename)
# To visualize the data, we will project it on 2 dimensions. To do so, we
# will use a Principal Component Analysis (as we made in the first steps),
# but projecting on 2 dimensions.
with mesure(" loading PCA 2D"):
if os.path.isfile(pca2d_filename):
pca_2d = _load_obj(pca2d_filename)
else:
docs = _load_docs(docs, training_set)
print " reduce the dimensionality of the dataset to 2 components"
pca_2d = RandomizedPCA(n_components=2,
whiten=True).fit(vec.transform(docs))
_save_obj(pca_2d, pca2d_filename)
# Now, go trough the whole resources for each users and try to find user
# profiles regarding TF-IDF
# as the process can take some time, there is a progressbar to keep the user
# updated about the status of the operation
for username in list(db.users.find().distinct('username')):
if user and user != username:
continue
# get all the resources for this user
urls = db.views.find({"user.username": username}).distinct("url")
if not urls:
continue # if we don't have any url for this user, go to the next one!
resources = list(
db.resources.find({
'url': {
'$in': urls
},
'blacklisted': False,
'processed': True
}))
if not resources:
continue
print "processing %s (%s docs)" % (username, len(resources))
# get the docs content and names
docs = [res['content'] for res in resources]
urls = [res['url'] for res in resources]
# fit the contents to the new set of features the PCA determined
with mesure(" reduce dataset dimensions to 100"):
docs_transformed = pca.transform(vec.transform(docs))
# what we do have now is a matrix with 100 dimentions, which is not really
# useful for representation. Keeping this for later analysis is a good
# thing so let's save this model for comparing profiles against resources
# later
# TODO pickle the kmeans into mongodb ?
# project X onto 2D
with mesure(" reduce dataset dimensions to 2"):
docs_2d = pca_2d.transform(vec.transform(docs))
# run the clustering algorithm
if algo in ["kmeans", "all"]:
with mesure(" kmeans(5)"):
cluster = KMeans(k=5).fit(docs_transformed)
# get_words_from_clusters(cluster, 10, docs, vec)
# print "ngrams for km on %s" % username
# get_n_bigrams_from_clusters(cluster, docs, 5)
plot_2d(cluster, docs_2d, username, "kmeans",
"Text-%s" % training_set)
plot_pie(cluster, username, "kmeans", "Text-%s" % training_set)
if algo in ["meanshift", "all"]:
with mesure(" meanshift"):
cluster = MeanShift().fit(docs_transformed)
# print "ngrams for ms on %s" % username
# get_n_bigrams_from_clusters(cluster, docs, 3)
plot_2d(cluster, docs_2d, username, "meanshift",
"Text-%s" % training_set)
plot_pie(cluster, username, "meanshift", "Text-%s" % training_set)
if algo in ["affinity", "all"]:
with mesure(" affinity propagation"):
cluster = AffinityPropagation().fit(
euclidean_distances(docs_transformed, docs_transformed))
plot_pie(cluster, username, "affinity", "Text-%s" % training_set)
def cluster_users(features=None):
"""Cluster the users, without using information about profiles.
Different features can be used to do so, at least text features and context
features.
"""
training_set = "newsgroup"
docs = None
vec_filename = os.path.join(OUTPUT_PATH,
"pickle/vec-%s.pickle" % training_set)
pca_filename = os.path.join(OUTPUT_PATH,
"pickle/pca-%s.pickle" % training_set)
# get the training set, transform it to N dimensions
with mesure(" loading vectors"):
if os.path.isfile(vec_filename):
vec = _load_obj(vec_filename)
else:
docs = _load_docs(docs, training_set)
vec = Vectorizer().fit(
docs) # equivalent to CountVectorizer + TfIdf
_save_obj(vec, vec_filename)
with mesure(" loading PCA"):
if os.path.isfile(pca_filename):
pca = _load_obj(pca_filename)
else:
docs = _load_docs(docs, training_set)
print " reduce the dimentionality of the dataset to 100 components"
# whiten=True ensure that the variance of each dim of the data in the
# transformed space is scaled to 1.0
pca = RandomizedPCA(n_components=100,
whiten=True).fit(vec.transform(docs))
_save_obj(pca, pca_filename)
# for each user, get the contents related to him.
users_content = []
users_labels = []
for username in list(db.users.find().distinct('username')):
# get all the resources for this user
urls = db.views.find({"user.username": username}).distinct("url")
if not urls:
continue # if we don't have any url for this user, go to the next one!
resources = list(
db.resources.find({
'url': {
'$in': urls
},
'blacklisted': False,
'processed': True
}))
if not resources:
continue
print "processing %s (%s docs)" % (username, len(resources))
# get the docs content and names
users_labels.append(username)
users_content.append(" ".join([res['content'] for res in resources]))
with mesure(" vectorise and reduce the dataset dimensions to 100"):
transformed_content = pca.transform(vec.transform(users_content))
# at the end, compute the similarity between users using different metrics
# kmeans 3 clusters
cluster = KMeans(3).fit(transformed_content)
plot_pie(cluster, "all", "kmeans", "text")
plot_2d(cluster, transformed_content, "all", "kmeans", "text")
user_list = [[
users_labels[idx]
for idx, _ in enumerate(cluster.labels_ == cluster_id) if _
] for cluster_id in np.unique(cluster.labels_)]
# compute similarity scores
from ipdb import set_trace
set_trace()
def find_profiles_context(algo=None, user=None):
"""Find profiles based on:
* location of the views
* time of the day of the views
* time of the day
* day of the week
"""
if not algo:
algo = "all"
# get all users
for username in db.users.distinct("username"):
if user and user != username:
continue
urls = db.views.find({"user.username": username}).distinct("url")
resources = []
if not urls:
continue
print "processing %s (%s docs)" % (username, len(urls))
t0 = time.time()
progress = ProgressBar(widgets=[
" building the matrix for %s" % username,
Percentage(),
Bar()
])
for url in progress(urls):
# get the views related to this user and this url
views = db.views.find({"user.username": username, "url": url})
views = list(views)
indicators = ['average', 'mean', 'median', 'var', 'std']
row = [len(views), sum([int(v['duration']) for v in views])]
# TODO add location
daytimes = []
weekdays = []
for view in views:
daytimes.append(view['daytime'])
weekdays.append(view['weekday'])
for indicator in indicators:
row.append(getattr(np, indicator)(daytimes))
row.append(getattr(np, indicator)(weekdays))
resources.append(row)
resources = np.array(resources)
print "matrix generation took %s" % (time.time() - t0)
# project X on 2D
# print " project the dataset into 2d"
# pca_2d = RandomizedPCA(n_components=2, whiten=True).fit(resources)
# docs_2d = pca_2d.transform(resources)
# run the clustering algorithm
if algo in ["kmeans", "all"]:
with mesure(" kmeans(5)"):
cluster = KMeans(k=5).fit(resources)
plot_2d(cluster, resources, username, "kmeans", "Context")
plot_pie(cluster, username, "kmeans", "Context")
if algo in ["meanshift", "all"]:
with mesure(" meanshift"):
cluster = MeanShift().fit(resources)
plot_2d(cluster, resources, username, "meanshift", "Context")
plot_pie(cluster, username, "meanshift", "Context")
if algo in ["affinity", "all"]:
with mesure(" affinity propagation"):
cluster = AffinityPropagation().fit(
euclidean_distances(resources, resources))
# plot_2d(cluster, resources, username, "affinity", "Context")
plot_pie(cluster, username, "affinity", "Context")
