Homec4science
Diffusion invenio-infoscience (master) Browse

R3600/invenio/legacy/bibrank/downloads_indexer.py
dd44fdec50bbmaster

Last Change
Raw File

downloads_indexer.py
View Options

# -*- coding: utf-8 -*-
##
## This file is part of Invenio.
## Copyright (C) 2005, 2006, 2007, 2008, 2010, 2011 CERN.
##
## Invenio is free software; you can redistribute it and/or
## modify it under the terms of the GNU General Public License as
## published by the Free Software Foundation; either version 2 of the
## License, or (at your option) any later version.
##
## Invenio is distributed in the hope that it will be useful, but
## WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
## General Public License for more details.
##
## You should have received a copy of the GNU General Public License
## along with Invenio; if not, write to the Free Software Foundation, Inc.,
## 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA.
__revision__ = "$Id$"
import os
import time
import calendar
import string
from six import iteritems
from invenio.legacy.dbquery import run_sql
def append_to_file(path, content):
"""print result in a file"""
if os.path.exists(path):
file_dest = open(path,"a")
file_dest.write("Hit on %s reads:" % time.strftime("%Y-%m-%d %H:%M:%S", time.localtime()))
file_dest.write(content)
file_dest.write("\n")
file_dest.close()
return content
def get_download_weight_filtering_user(dic, keys):
""" update the dictionnary.Without duplicates.Count maximum one hit per user per hour"""
for k in keys:
weight = 0
user_ips = run_sql("select count(distinct client_host) from rnkDOWNLOADS where id_bibrec=%s group by id_bibdoc" % k)
for ip in user_ips:
weight = weight + ip[0]
dic[k] = weight
return dic
def get_download_weight_total(dic, keys):
""" update the dictionnary.Count all the hit"""
for k in keys:
values = run_sql("select count(*) from rnkDOWNLOADS where id_bibrec=%s %s" % (k,";"))
dic[k] = values[0][0]
return dic
def uniq(alist):
"""Remove duplicate element in alist
Fastest without order preserving"""
set = {}
map(set.__setitem__, alist, [])
return set.keys()
def database_tuples_to_single_list(tuples):
"""convert a tuple extracted from the database into a list"""
return [elem[0] for elem in tuples]
def new_downloads_to_index (last_updated):
"""id_bibrec of documents downloaded since the last run of bibrank """
id_bibrec_list = database_tuples_to_single_list(run_sql("select id_bibrec from rnkDOWNLOADS where download_time >=\"%s\"" % last_updated))
res = uniq(id_bibrec_list)
return res
def filter_downloads_per_hour_with_docid (keys, last_updated):
"""filter all the duplicate downloads per user for each hour intervall"""
for k in keys:
id_bibdocs = run_sql("select distinct id_bibdoc from rnkDOWNLOADS where id_bibrec=%s" % k)
for bibdoc in id_bibdocs:
values = run_sql("""select DATE_FORMAT(download_time,"%%Y-%%m-%%d %%H"), client_host from rnkDOWNLOADS where id_bibrec=%s and id_bibdoc=%s and download_time >=\"%s\";""" % (k, bibdoc[0], last_updated))
for val in values:
date_res = val[0]
date1 = "%s:00:00" % (date_res)
date2 = compute_next_hour(date_res)
duplicates = (run_sql("select count(*) from rnkDOWNLOADS where id_bibrec=%s and id_bibdoc=%s and download_time>='%s' and download_time<'%s' and client_host=%s;" % (k, bibdoc[0], date1, date2, val[1]))[0][0])-1
run_sql("delete from rnkDOWNLOADS where id_bibrec=%s and id_bibdoc=%s and download_time>='%s' and download_time<'%s' and client_host=%s limit %s;" % (k, bibdoc[0], date1, date2, val[1], duplicates))
def filter_downloads_per_hour (keys, last_updated):
"""filter all the duplicate downloads per user for each hour intervall"""
for k in keys:
values = run_sql("""select DATE_FORMAT(download_time,"%%Y-%%m-%%d %%H"), client_host from rnkDOWNLOADS where id_bibrec=%s and download_time >=\"%s\";""" % (k, last_updated))
for val in values:
date_res = val[0]
date1 = "%s:00:00" % (date_res)
date2 = compute_next_hour(date_res)
duplicates = (run_sql("select count(*) from rnkDOWNLOADS where id_bibrec=%s and download_time>='%s' and download_time<'%s' and client_host=%s;" % (k, date1, date2, val[1]))[0][0])-1
run_sql("delete from rnkDOWNLOADS where id_bibrec=%s and download_time>='%s' and download_time<'%s' and client_host=%s limit %s;" % (k, date1, date2, val[1], duplicates))
def compute_next_hour(date_res):
"""treat the change of the year, of (special)month etc.. and return the date in database format"""
next_date = ""
date_res, date_hour = string.split(date_res, " ")
date_hour = string.atoi(date_hour)
if date_hour == 23:
date_year, date_month, date_day = string.split(date_res, "-")
date_year = string.atoi(date_year)
date_month = string.atoi(date_month)
date_day = string.atoi(date_day)
if date_month == 12 and date_day == 31:
next_date = "%s-%s-%s 00:00:00" % (date_year + 1, 01, 01)
elif calendar.monthrange(date_year, date_month)[1] == date_day:
next_date = "%s-%s-%s 00:00:00" % (date_year, date_month + 1, 01)
else :
next_date = "%s-%s-%s 00:00:00" % (date_year, date_month, date_day + 1)
else :
next_hour = date_hour + 1
next_date = "%s %s:00:00" % (date_res, next_hour)
return next_date
def get_file_similarity_by_times_downloaded(dic, id_bibrec_list):
"""For each id_bibrec, get the client_host and see which other id_bibrec these users have also downloaded.
Return update dictionnary of this form
{id_bibrec:[(id_bibrec1,score),(id_bibrec2,score)],id_bibrec:[(),()]...}
Take long time so let's see bibrank_downloads_similarity which compute in fly the similarity for
a particular recid."""
dic_result = {}
if id_bibrec_list != []:
tuple_string_id_bibrec_list = str(tuple(id_bibrec_list))
if len(id_bibrec_list) == 1:
tuple_string_id_bibrec_list = tuple_string_id_bibrec_list.replace(',','')
#first compute the download similarity between the new documents
#which have been downloadwd since the last run of bibrank
dic_news = {}
res = run_sql("select id_bibrec,client_host from rnkDOWNLOADS where id_bibrec in %s;" % tuple_string_id_bibrec_list)
for res_elem in res:
id_bibrec_key = res_elem[0]
client_host_value = str(res_elem[1])
if id_bibrec_key in dic_news.keys():
tmp_list = dic_news[id_bibrec_key]
if client_host_value not in dic_news[id_bibrec_key]:
tmp_list.append(client_host_value)
dic_news[id_bibrec_key] = tmp_list
else :
list_client_host_value = []
list_client_host_value.append(client_host_value)
dic_news[id_bibrec_key] = list_client_host_value
#compute occurence of client_host
for j in dic_news.keys():
list_tuple = []
tuple_client_host = str(tuple(dic_news[j]))
if len(tuple(dic_news[j])) == 1:
tuple_client_host = tuple_client_host.replace(',','')
res2 = run_sql("select id_bibrec,count(*) from rnkDOWNLOADS where client_host in %s and id_bibrec in %s and id_bibrec != %s group by id_bibrec;" % (tuple_client_host, tuple_string_id_bibrec_list, j)) #0.0023 par requete
list_tuple.append(list(res2))
dic_result[j] = list_tuple[0]
#merge new values with old dictionnary
return merge_with_old_dictionnary(dic, dic_result)
def merge_with_old_dictionnary(old_dic, new_dic):
"""For each key id_bibrec in new_dic add the old values contained in old_dic
Return not ordered merged dictionnary"""
union_dic = {}
for (key, value) in iteritems(new_dic):
if key in old_dic.keys():
old_dic_value_dic = dict(old_dic[key])
tuple_list = []
old_dic_value_dic_keys = old_dic_value_dic.keys()
for val in value:
if val[0] in old_dic_value_dic_keys:
tuple_list.append((val[0], val[1]+ old_dic_value_dic[val[0]]))
del old_dic_value_dic[val[0]]
else :
tuple_list.append((val[0], val[1]))
old_dic_value_dic_items = old_dic_value_dic.items()
if old_dic_value_dic_items != []:
tuple_list.extend(old_dic_value_dic_items)
union_dic[key] = tuple_list
else :
union_dic[key] = value
for (key, value) in iteritems(old_dic):
if key not in union_dic.keys():
union_dic[key] = value
return union_dic
c4science · Help