newegg_scraper_chsbr 0.1.2 → 0.2.2
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/bin/newegg_scraper +1 -1
- data/lib/environment.rb +5 -1
- data/lib/newegg_scraper_chsbr/cli.rb +72 -14
- data/lib/newegg_scraper_chsbr/cpu.rb +5 -5
- data/lib/newegg_scraper_chsbr/data_grabber.rb +32 -64
- data/lib/newegg_scraper_chsbr/descriptors.rb +4 -1
- data/lib/newegg_scraper_chsbr/name.rb +1 -1
- data/lib/newegg_scraper_chsbr/scraper.rb +1 -1
- data/lib/newegg_scraper_chsbr/shipping.rb +6 -0
- data/lib/newegg_scraper_chsbr/user.rb +14 -5
- data/lib/newegg_scraper_chsbr/version.rb +1 -1
- metadata +11 -10
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: b6f2133128bd60c1e34d9d7ed98d65aa0bd5c305c804a4491738c22ea6b48931
|
4
|
+
data.tar.gz: f431eac85ce8616a3143537293e3b552a50056bb83b7afd84d6f0be525387b90
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 32761b7f137118cd0b9ff915b745078f28ed74402ded5cfee1163196a6412e5904e53a5c8be73a8c818e637fa98884f483c714c3c62297a4c9b0dbded09d5fbd
|
7
|
+
data.tar.gz: deefa949cc5df1ea06cd1bf75a6977277b76573ac2ad9c1b5a2fc6fb01dc921bb7ed4163c9a1f1df5518f19cbaeddbc5df6b6afc7a7e58f643205bc42ef5f7f9
|
data/bin/newegg_scraper
CHANGED
data/lib/environment.rb
CHANGED
@@ -1,6 +1,8 @@
|
|
1
1
|
require 'open-uri'
|
2
2
|
require 'nokogiri'
|
3
3
|
require "bundler/setup"
|
4
|
+
|
5
|
+
|
4
6
|
require_relative "./newegg_scraper_chsbr/version.rb"
|
5
7
|
require_relative './newegg_scraper_chsbr/data_grabber.rb'
|
6
8
|
require_relative './newegg_scraper_chsbr/scraper.rb'
|
@@ -8,4 +10,6 @@ require_relative './newegg_scraper_chsbr/user.rb'
|
|
8
10
|
require_relative './newegg_scraper_chsbr/name.rb'
|
9
11
|
require_relative './newegg_scraper_chsbr/price.rb'
|
10
12
|
require_relative './newegg_scraper_chsbr/cli.rb'
|
11
|
-
require_relative './newegg_scraper_chsbr/cpu.rb'
|
13
|
+
require_relative './newegg_scraper_chsbr/cpu.rb'
|
14
|
+
require_relative './newegg_scraper_chsbr/descriptors.rb'
|
15
|
+
require_relative './newegg_scraper_chsbr/shipping.rb'
|
@@ -4,28 +4,52 @@ class NeweggScraperChsbr::CLI
|
|
4
4
|
def initialize
|
5
5
|
NeweggScraperChsbr::DataGrabber.new
|
6
6
|
end
|
7
|
-
def
|
8
|
-
puts "
|
9
|
-
puts "Please enter y to
|
7
|
+
def call
|
8
|
+
puts "Were you interested in purchasing some CPU's?"
|
9
|
+
puts "Please enter y to do so."
|
10
10
|
@counter = gets.strip
|
11
|
-
|
11
|
+
corr_input = control_bool
|
12
|
+
while corr_input
|
13
|
+
puts "Were you interested in purchasing some CPU's?"
|
14
|
+
puts "Please enter y to do so.\nPlease enter n or exit to quit."
|
15
|
+
@counter = gets.strip
|
16
|
+
corr_input = control_bool
|
17
|
+
end
|
18
|
+
while @counter == 'y'
|
19
|
+
get_price_range
|
12
20
|
get_init_data_for_user
|
13
21
|
NeweggScraperChsbr::Cpu.display_cpu(@user)
|
14
22
|
@user.choose_cpu
|
15
23
|
NeweggScraperChsbr::Cpu.display_cpu_with_extras(@user)
|
16
|
-
puts "If you would like to
|
24
|
+
puts "If you would like to restart the program please enter : #{@counter}.\nTo exit type n or exit."
|
17
25
|
@counter = gets.strip
|
26
|
+
corr_input = control_bool
|
27
|
+
while corr_input
|
28
|
+
puts "Were you interested in purchasing some CPU's?"
|
29
|
+
puts "Please enter y to do so.\nPlease enter n or exit to quit."
|
30
|
+
@counter = gets.strip
|
31
|
+
corr_input = control_bool
|
32
|
+
end
|
33
|
+
|
18
34
|
end
|
19
35
|
puts "Thanks for using Chesbro's Scraper!"
|
20
36
|
end
|
21
37
|
def get_init_data_for_user
|
22
38
|
|
23
39
|
if @counter == 'y'
|
24
|
-
|
40
|
+
|
25
41
|
puts "Are you searching for Intel or AMD?"
|
26
42
|
puts "Please put i for Intel, or a for AMD."
|
27
43
|
puts "Please put x for either."
|
28
44
|
@cpu_make = gets.strip
|
45
|
+
not_found = cpu_make_bool
|
46
|
+
while not_found
|
47
|
+
puts "Are you searching for Intel or AMD?"
|
48
|
+
puts "Please put i for Intel, or a for AMD."
|
49
|
+
puts "Please put x for either."
|
50
|
+
@cpu_make = gets.strip
|
51
|
+
not_found = cpu_make_bool
|
52
|
+
end
|
29
53
|
puts "\n\nYou want to search for (an) #{NeweggScraperChsbr::Cpu.cpu_maker(@cpu_make)} CPU."
|
30
54
|
if @max_price.to_i != 0 && @min_price.to_i != 0
|
31
55
|
puts "You entered a maximum price of $#{@max_price}, and a minimum price of $#{@min_price}."
|
@@ -40,28 +64,62 @@ class NeweggScraperChsbr::CLI
|
|
40
64
|
end
|
41
65
|
end
|
42
66
|
def if_correct_info
|
43
|
-
puts "\
|
67
|
+
puts "\nIf you're satisfied with your selection, the computer will print\na list of CPUs that satisfy the criteria. Enter y to continue. Enter n to choose again."
|
44
68
|
@counter = gets.strip
|
45
69
|
if @counter == 'y'
|
46
70
|
@user = NeweggScraperChsbr::User.new(@cpu_make, @min_price, @max_price)
|
47
71
|
else
|
48
|
-
|
72
|
+
@counter = 'y'
|
73
|
+
call
|
49
74
|
end
|
50
75
|
end
|
51
76
|
def get_price_range
|
52
|
-
puts "
|
77
|
+
puts "Welcome. Would you like to set a budget? If no, set your maximum budget to 0"
|
78
|
+
puts "Please enter the integer value of your maximum price."
|
79
|
+
puts "$500.50, please enter as 501, or 500 to stay below budget."
|
80
|
+
@max_price = gets.strip
|
81
|
+
while @max_price.to_i < 0 || @max_price.to_i > 10000 || @max_price.include?("$") || @max_price.include?(".") || @max_price == nil
|
82
|
+
puts "Welcome. Would you like to set a budget? If no, set your maximum budget to 0"
|
53
83
|
puts "Please enter the integer value of your maximum price."
|
54
84
|
puts "$500.50, please enter as 501, or 500 to stay below budget."
|
55
85
|
@max_price = gets.strip
|
56
|
-
|
57
|
-
|
58
|
-
|
86
|
+
end
|
87
|
+
if @max_price.to_i != 0
|
88
|
+
puts "Please enter the minimum you would spend on a CPU."
|
89
|
+
puts "Please use the same format as above."
|
90
|
+
@min_price = gets.strip
|
91
|
+
|
92
|
+
while @min_price.to_i <= 0 || @min_price.to_i >= @max_price.to_i || @min_price.include?("$") || @min_price.include?(".") || @min_price == nil
|
59
93
|
puts "Please enter the minimum you would spend on a CPU."
|
60
94
|
puts "Please use the same format as above."
|
61
95
|
@min_price = gets.strip
|
62
96
|
end
|
63
|
-
|
97
|
+
end
|
64
98
|
end
|
99
|
+
def control_bool
|
100
|
+
if @counter.include?("y")
|
101
|
+
false
|
102
|
+
elsif @counter.include?("exit")
|
103
|
+
false
|
104
|
+
elsif @counter.include?("n")
|
105
|
+
false
|
106
|
+
else
|
107
|
+
true
|
108
|
+
end
|
109
|
+
end
|
110
|
+
def cpu_make_bool
|
111
|
+
if @cpu_make.include?("x")
|
112
|
+
@cpu_make = "x"
|
113
|
+
false
|
114
|
+
elsif @cpu_make.include?("a")
|
115
|
+
@cpu_make = "a"
|
116
|
+
false
|
117
|
+
elsif @cpu_make.include?("i")
|
118
|
+
@cpu_make = "i"
|
119
|
+
false
|
120
|
+
else
|
121
|
+
true
|
122
|
+
end
|
65
123
|
|
66
|
-
|
124
|
+
end
|
67
125
|
end
|
@@ -5,8 +5,8 @@ class NeweggScraperChsbr::Cpu
|
|
5
5
|
def initialize(name, price, shipping, desc)
|
6
6
|
@name = NeweggScraperChsbr::Name.new(name)
|
7
7
|
@price = NeweggScraperChsbr::Price.new(price)
|
8
|
-
@desc = desc
|
9
|
-
@shipping = shipping
|
8
|
+
@desc = NeweggScraperChsbr::Descriptors.new(desc)
|
9
|
+
@shipping = NeweggScraperChsbr::Shipping.new(shipping)
|
10
10
|
@@all << self
|
11
11
|
end
|
12
12
|
def self.all
|
@@ -25,7 +25,7 @@ class NeweggScraperChsbr::Cpu
|
|
25
25
|
counter = 1
|
26
26
|
@@all.each do | computer |
|
27
27
|
if user.max_price.to_i != 0 && user.min_price.to_i == 0
|
28
|
-
|
28
|
+
|
29
29
|
|
30
30
|
user.min_price = "1"
|
31
31
|
end
|
@@ -54,12 +54,12 @@ class NeweggScraperChsbr::Cpu
|
|
54
54
|
def self.display_cpu_with_extras(user)
|
55
55
|
|
56
56
|
user.chosen_cpu.each do | chosen_cpu |
|
57
|
-
puts "\n\n\n#{@@all[chosen_cpu - 1].name.name}\nHas a price of: #{@@all[chosen_cpu - 1].price.price}\nShipping: #{@@all[chosen_cpu - 1].shipping}\n\n Descriptive Points:\n"
|
57
|
+
puts "\n\n\n#{@@all[chosen_cpu - 1].name.name}\nHas a price of: #{@@all[chosen_cpu - 1].price.price}\nShipping: #{@@all[chosen_cpu - 1].shipping.price}\n\n Descriptive Points:\n"
|
58
58
|
@@all[chosen_cpu - 1].printDesc
|
59
59
|
end
|
60
60
|
end
|
61
61
|
def printDesc
|
62
|
-
@desc.each do | key, value |
|
62
|
+
@desc.bullets.each do | key, value |
|
63
63
|
puts " #{key + 1}: #{value}"
|
64
64
|
end
|
65
65
|
end
|
@@ -1,4 +1,3 @@
|
|
1
|
-
|
2
1
|
class NeweggScraperChsbr::DataGrabber
|
3
2
|
attr_reader :cpus
|
4
3
|
|
@@ -9,7 +8,7 @@ class NeweggScraperChsbr::DataGrabber
|
|
9
8
|
names = getNames
|
10
9
|
descHash = getCpuDesc
|
11
10
|
@cpus = makeCpus(prices, names, shipping_price, descHash)
|
12
|
-
end
|
11
|
+
end
|
13
12
|
|
14
13
|
def getCpuDesc
|
15
14
|
scraped = NeweggScraperChsbr::Scraper.new
|
@@ -18,7 +17,10 @@ class NeweggScraperChsbr::DataGrabber
|
|
18
17
|
css_next_link = scraped.xml_obj.css ".item-title"
|
19
18
|
css_next_link.each_with_index do | element, index |
|
20
19
|
if index != 0
|
21
|
-
|
20
|
+
|
21
|
+
if element.attributes["href"] != nil
|
22
|
+
pages << element.attributes["href"].text
|
23
|
+
end
|
22
24
|
end
|
23
25
|
end
|
24
26
|
description = {}
|
@@ -44,19 +46,19 @@ class NeweggScraperChsbr::DataGrabber
|
|
44
46
|
|
45
47
|
|
46
48
|
end
|
47
|
-
def split_price(html_element)
|
49
|
+
def split_price(html_element)
|
48
50
|
counter = 0
|
49
51
|
until counter == 100
|
50
|
-
if counter < 10
|
51
|
-
if html_element.text.include?(".0#{counter.to_s}")
|
52
|
-
|
53
|
-
price = html_element.text.split ".0#{counter.to_s}"
|
54
|
-
|
55
|
-
|
52
|
+
if counter < 10
|
53
|
+
if html_element.text.include?(".0#{counter.to_s}")
|
54
|
+
|
55
|
+
price = html_element.text.split ".0#{counter.to_s}"
|
56
|
+
|
57
|
+
|
56
58
|
price.delete_at 1
|
57
|
-
return price, ".0#{counter.to_s}"
|
59
|
+
return price, ".0#{counter.to_s}"
|
58
60
|
end
|
59
|
-
elsif counter >= 10
|
61
|
+
elsif counter >= 10
|
60
62
|
if html_element.text.include?(".#{counter.to_s}")
|
61
63
|
|
62
64
|
price = (html_element.text.split ".#{counter.to_s}")
|
@@ -67,59 +69,35 @@ class NeweggScraperChsbr::DataGrabber
|
|
67
69
|
counter += 1
|
68
70
|
end
|
69
71
|
end
|
70
|
-
def isCoolerOrMB?(name)
|
71
|
-
name.include?("Water") || name.include?("Air") || name.include?("Motherboard")
|
72
|
-
# any of the products include
|
73
|
-
# words that describe motherboards and/or
|
74
|
-
# cooling products.
|
72
|
+
def isCoolerOrMB?(name)
|
73
|
+
name.include?("Water") || name.include?("Air") || name.include?("Motherboard")
|
75
74
|
end
|
76
75
|
def getPrice
|
77
76
|
scraped_info = NeweggScraperChsbr::Scraper.new
|
78
|
-
css_price = scraped_info.xml_obj.css ".price-current"
|
79
|
-
# html that contains what I need to find the price.
|
77
|
+
css_price = scraped_info.xml_obj.css ".price-current"
|
80
78
|
|
81
|
-
prices = []
|
82
|
-
|
83
|
-
css_price.each do | piece | # Going 1 level deep into the html obj that contains the price text element.
|
84
|
-
|
85
|
-
temp_price = split_price piece # split_price will split the price, and return
|
86
|
-
# two variables in an array. The first element of
|
87
|
-
# the array will be the dollar amount.
|
88
|
-
# The second element will be the cent ammount because
|
89
|
-
# I deemed it easier to remove anything after the
|
90
|
-
# floating point, it was the only pattern that was
|
91
|
-
# consistent.
|
79
|
+
prices = []
|
92
80
|
|
81
|
+
css_price.each do | piece |
|
82
|
+
temp_price = split_price piece
|
93
83
|
if temp_price != nil
|
94
|
-
temp_price.flatten!
|
95
|
-
|
96
|
-
# to return the flattened array to temp_price.
|
97
|
-
|
98
|
-
prices << "#{temp_price[0]}#{temp_price[1]}" # Here prices is being shoveled the
|
99
|
-
# two elements, concatenated to create
|
100
|
-
# a string similar to "$00.00"
|
84
|
+
temp_price.flatten!
|
85
|
+
prices << "#{temp_price[0]}#{temp_price[1]}"
|
101
86
|
end
|
102
87
|
|
103
|
-
end
|
88
|
+
end
|
104
89
|
prices
|
105
90
|
end
|
106
91
|
def getNames
|
107
92
|
scraped = NeweggScraperChsbr::Scraper.new
|
108
|
-
names = []
|
109
|
-
# objects found on the page.
|
93
|
+
names = []
|
110
94
|
|
111
|
-
css_name = scraped.xml_obj.css ".item-title"
|
112
|
-
# object to be enumerated upon.
|
95
|
+
css_name = scraped.xml_obj.css ".item-title"
|
113
96
|
|
114
|
-
css_name.each_with_index do | name, index |
|
115
|
-
# to retrieve the strings.
|
97
|
+
css_name.each_with_index do | name, index |
|
116
98
|
|
117
|
-
index != 0 ? names << name.text : nil
|
118
|
-
|
119
|
-
# After I skip the first name
|
120
|
-
# I will shovel the text from the
|
121
|
-
# html object into the names array.
|
122
|
-
end #end names.each_with_index
|
99
|
+
index != 0 ? names << name.text : nil
|
100
|
+
end
|
123
101
|
names
|
124
102
|
end
|
125
103
|
def getShipping
|
@@ -136,23 +114,13 @@ class NeweggScraperChsbr::DataGrabber
|
|
136
114
|
|
137
115
|
|
138
116
|
cpus = []
|
139
|
-
prices.each_with_index do | price, index |
|
140
|
-
# each of the prices with their
|
141
|
-
# index so that the enumation can
|
142
|
-
# access the names array at the
|
143
|
-
# same time.
|
117
|
+
prices.each_with_index do | price, index |
|
144
118
|
|
145
|
-
if !isCoolerOrMB?(names[index])
|
146
|
-
# the name is a cooler
|
147
|
-
# or motherboard because
|
148
|
-
# there is a price associated
|
149
|
-
# with the coolers on the
|
150
|
-
# website.
|
151
|
-
|
119
|
+
if !isCoolerOrMB?(names[index])
|
152
120
|
cpus << NeweggScraperChsbr::Cpu.new(names[index], price, shipping[index], desc_hash[index])
|
153
121
|
|
154
|
-
end
|
155
|
-
end
|
122
|
+
end
|
123
|
+
end
|
156
124
|
cpus
|
157
125
|
end
|
158
126
|
end
|
@@ -4,7 +4,7 @@ class NeweggScraperChsbr::Scraper
|
|
4
4
|
|
5
5
|
URL = "https://www.newegg.com/p/pl?d=cpu"
|
6
6
|
def initialize(url = URL)
|
7
|
-
@xml_obj = Nokogiri::HTML(URI.open(url))
|
7
|
+
@xml_obj = Nokogiri::HTML(URI.open(url))
|
8
8
|
end
|
9
9
|
|
10
10
|
|
@@ -8,13 +8,22 @@ class NeweggScraperChsbr::User
|
|
8
8
|
@chosen_cpu = []
|
9
9
|
end
|
10
10
|
def choose_cpu
|
11
|
-
puts "Please enter the number of the CPU you'd like to
|
12
|
-
chosen = gets.strip
|
13
|
-
|
14
|
-
|
15
|
-
|
11
|
+
puts "Please enter the number of the CPU you'd like to see more about."
|
12
|
+
chosen = gets.strip
|
13
|
+
while !is_numeric?(chosen)
|
14
|
+
puts "Please enter the number of the CPU you'd like to see more about."
|
15
|
+
chosen = gets.strip
|
16
|
+
end
|
17
|
+
@chosen_cpu << chosen.to_i
|
18
|
+
puts "Would you like to see more about another CPU?\nEnter y to do so.\nEnter n to see details about the chosen CPU(s)"
|
19
|
+
input = gets.strip
|
20
|
+
|
21
|
+
if input == 'y'
|
16
22
|
choose_cpu
|
17
23
|
end
|
18
24
|
|
19
25
|
end
|
26
|
+
def is_numeric?(obj)
|
27
|
+
obj.match(/\A[^+-]?\d+?(\^.\d+)?\Z/) == nil ? false : true
|
28
|
+
end
|
20
29
|
end
|
metadata
CHANGED
@@ -1,43 +1,43 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: newegg_scraper_chsbr
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.
|
4
|
+
version: 0.2.2
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Lrd134
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2021-04-
|
11
|
+
date: 2021-04-27 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: nokogiri
|
15
15
|
requirement: !ruby/object:Gem::Requirement
|
16
16
|
requirements:
|
17
|
-
- - "
|
17
|
+
- - "~>"
|
18
18
|
- !ruby/object:Gem::Version
|
19
|
-
version:
|
19
|
+
version: 1.11.3
|
20
20
|
type: :runtime
|
21
21
|
prerelease: false
|
22
22
|
version_requirements: !ruby/object:Gem::Requirement
|
23
23
|
requirements:
|
24
|
-
- - "
|
24
|
+
- - "~>"
|
25
25
|
- !ruby/object:Gem::Version
|
26
|
-
version:
|
26
|
+
version: 1.11.3
|
27
27
|
- !ruby/object:Gem::Dependency
|
28
28
|
name: open-uri
|
29
29
|
requirement: !ruby/object:Gem::Requirement
|
30
30
|
requirements:
|
31
|
-
- - "
|
31
|
+
- - "~>"
|
32
32
|
- !ruby/object:Gem::Version
|
33
|
-
version: '0'
|
33
|
+
version: '0.1'
|
34
34
|
type: :runtime
|
35
35
|
prerelease: false
|
36
36
|
version_requirements: !ruby/object:Gem::Requirement
|
37
37
|
requirements:
|
38
|
-
- - "
|
38
|
+
- - "~>"
|
39
39
|
- !ruby/object:Gem::Version
|
40
|
-
version: '0'
|
40
|
+
version: '0.1'
|
41
41
|
description: To scrape CPU's off of Newegg and display them nicely.
|
42
42
|
email:
|
43
43
|
- larryc3200@gmail.com
|
@@ -55,6 +55,7 @@ files:
|
|
55
55
|
- lib/newegg_scraper_chsbr/name.rb
|
56
56
|
- lib/newegg_scraper_chsbr/price.rb
|
57
57
|
- lib/newegg_scraper_chsbr/scraper.rb
|
58
|
+
- lib/newegg_scraper_chsbr/shipping.rb
|
58
59
|
- lib/newegg_scraper_chsbr/user.rb
|
59
60
|
- lib/newegg_scraper_chsbr/version.rb
|
60
61
|
homepage: https://github.com/Lrd134/aa_phase_1_proj
|