pcgs 1.0.0
Sign up to get free protection for your applications and to get access to all the features.
- data/lib/pcgs.rb +339 -0
- metadata +46 -0
data/lib/pcgs.rb
ADDED
@@ -0,0 +1,339 @@
|
|
1
|
+
require "hpricot"
|
2
|
+
require "open-uri"
|
3
|
+
|
4
|
+
class PCGS
|
5
|
+
class Wallet
|
6
|
+
attr_accessor :coins
|
7
|
+
|
8
|
+
attr_accessor :tables
|
9
|
+
|
10
|
+
def initialize(type=nil)
|
11
|
+
doc = PCGS.scrape("http://www.pcgs.com/prices/")
|
12
|
+
cols = doc.search("//div[@class='twocolumn']").first.search("//div[@class='col']")
|
13
|
+
boxes = cols.map{|col|col.search("//div").find_all{|div|div["class"].include?("box")}}.flatten
|
14
|
+
as = boxes.map{|box|box.search("//ul").first.search("//a")}.flatten
|
15
|
+
urls = as.map{|a|"http://www.pcgs.com"+a["href"]}
|
16
|
+
if not type.nil?
|
17
|
+
urls = as.find_all{|a|a.inner_html==type}.map{|a|"http://www.pcgs.com"+a["href"]}
|
18
|
+
end
|
19
|
+
|
20
|
+
@tables = []
|
21
|
+
|
22
|
+
urls.each do |url|
|
23
|
+
# p "Getting prices for #{url.split('title=')[1].gsub('+',' ')}..."
|
24
|
+
doc = PCGS.scrape(url)
|
25
|
+
ti = PCGS.table_info(doc)
|
26
|
+
ti.each do |t|
|
27
|
+
# p " Making table for #{t[:tab_title]}, #{t[:coin_grade_type]}."
|
28
|
+
@tables << PCGS.objectify_table_info(t)
|
29
|
+
end
|
30
|
+
end
|
31
|
+
|
32
|
+
@coins = []
|
33
|
+
|
34
|
+
self.coinify_tables
|
35
|
+
|
36
|
+
self
|
37
|
+
end
|
38
|
+
|
39
|
+
def coinify_tables
|
40
|
+
self.tables.each do |table|
|
41
|
+
self.coinify_table(table)
|
42
|
+
end
|
43
|
+
end
|
44
|
+
|
45
|
+
def coinify_table(table)
|
46
|
+
# p "Adding #{table.coin_type} coins to wallet..."
|
47
|
+
table.rows.each do |row|
|
48
|
+
if row.elements.first.is_a?(Integer)
|
49
|
+
if row.elements[1] != "Type"
|
50
|
+
pcgs_no = row.elements[0].to_i
|
51
|
+
description = row.elements[1]
|
52
|
+
design = row.elements[2]
|
53
|
+
grade_type = table.coin_grade_type
|
54
|
+
subtype = row.coin_subtype
|
55
|
+
type = table.coin_type
|
56
|
+
y = description.to_s.scan(/\d\d\d\d/)
|
57
|
+
if y.empty?
|
58
|
+
year = ""
|
59
|
+
elsif y.size == 1
|
60
|
+
year = y[0]
|
61
|
+
elsif y.size == 2
|
62
|
+
if description.gsub(" ","").include?(y[0]+"-"+y[1])
|
63
|
+
year = y[0]+"-"+y[1]
|
64
|
+
else
|
65
|
+
year = y[0]
|
66
|
+
end
|
67
|
+
end
|
68
|
+
mint_mark = description.to_s.scan(/\d\d\d\d\-[A-Z]/)[0][-1] rescue ""
|
69
|
+
row.elements[3..-1].each do |e|
|
70
|
+
i = row.elements.index(e)
|
71
|
+
grade = row.header_row.elements[i]
|
72
|
+
price = row.elements[i]
|
73
|
+
coin = self.add_coin(pcgs_no, description, design, grade, price, grade_type, subtype, type, year, mint_mark)
|
74
|
+
coin = coin.set_name
|
75
|
+
# p "Adding a #{coin.name} to the wallet."
|
76
|
+
end
|
77
|
+
end
|
78
|
+
end
|
79
|
+
end
|
80
|
+
end
|
81
|
+
|
82
|
+
def add_coin(pcgs_no, description, design, grade, price, grade_type, subtype, type, year, mint_mark)
|
83
|
+
coin = PCGS::Coin.new(pcgs_no, description, design, grade, price, grade_type, subtype, type, year, mint_mark)
|
84
|
+
self.coins << coin
|
85
|
+
coin
|
86
|
+
end
|
87
|
+
|
88
|
+
def coins_like(str)
|
89
|
+
wallet = self
|
90
|
+
types = wallet.coins.find_all{|c|c.type_like?(str)}
|
91
|
+
names = wallet.coins.find_all{|c|c.name_like?(str)}
|
92
|
+
(types+names).flatten.uniq
|
93
|
+
end
|
94
|
+
end
|
95
|
+
|
96
|
+
class Table
|
97
|
+
attr_accessor :coin_type
|
98
|
+
attr_accessor :tab
|
99
|
+
attr_accessor :coin_grade_type
|
100
|
+
attr_accessor :rows
|
101
|
+
attr_accessor :doc
|
102
|
+
|
103
|
+
def initialize(coin_type, tab, coin_grade_type, doc)
|
104
|
+
@coin_type = coin_type
|
105
|
+
@tab = tab
|
106
|
+
@coin_grade_type = coin_grade_type
|
107
|
+
@rows = []
|
108
|
+
@doc = doc
|
109
|
+
end
|
110
|
+
|
111
|
+
def add_row(coin_subtype, elements, header_row)
|
112
|
+
row = PCGS::Row.new(coin_subtype, elements, header_row)
|
113
|
+
self.rows << row
|
114
|
+
row
|
115
|
+
end
|
116
|
+
|
117
|
+
def add_rows(rows)
|
118
|
+
if not rows.empty?
|
119
|
+
table = self
|
120
|
+
length = rows.first.size
|
121
|
+
subtype = table.coin_type
|
122
|
+
header_row = nil
|
123
|
+
hh = "<img src=\"images/expand.gif\" align=\"left\" class=\"expandcollapseimages\" border=\"0\" />"
|
124
|
+
rows.each do |row|
|
125
|
+
if row.size == length
|
126
|
+
if row.first.to_s.include?(hh)
|
127
|
+
row[0] = row[0].gsub(hh,"").to_i
|
128
|
+
end
|
129
|
+
if row.first.is_a?(Integer)
|
130
|
+
table.add_row(subtype, row, header_row)
|
131
|
+
elsif row.first[0..3] == "PCGS"
|
132
|
+
header_row = table.add_row("Header", row, nil)
|
133
|
+
end
|
134
|
+
elsif row.size == 1
|
135
|
+
if row.first.include?("Price Changes") && rows[1].first.include?("Collectors Corner")
|
136
|
+
subtype = row.first.split(",")[0]
|
137
|
+
end
|
138
|
+
end
|
139
|
+
end
|
140
|
+
end
|
141
|
+
end
|
142
|
+
end
|
143
|
+
|
144
|
+
class Row
|
145
|
+
attr_accessor :coin_subtype
|
146
|
+
attr_accessor :elements
|
147
|
+
attr_accessor :header_row
|
148
|
+
|
149
|
+
def initialize(coin_subtype, elements, header_row)
|
150
|
+
@coin_subtype = coin_subtype
|
151
|
+
@elements = elements
|
152
|
+
@header_row = header_row
|
153
|
+
end
|
154
|
+
end
|
155
|
+
|
156
|
+
class Coin
|
157
|
+
attr_accessor :pcgs_no
|
158
|
+
attr_accessor :description
|
159
|
+
attr_accessor :design
|
160
|
+
attr_accessor :grade
|
161
|
+
attr_accessor :price
|
162
|
+
attr_accessor :grade_type
|
163
|
+
attr_accessor :subtype
|
164
|
+
attr_accessor :type
|
165
|
+
|
166
|
+
attr_accessor :year
|
167
|
+
attr_accessor :mint_mark
|
168
|
+
attr_accessor :name
|
169
|
+
|
170
|
+
NAMES = {"halfcent" => ["halfcent", "half cent", "hapenny", "halfpenny", "half penny", "1/2cent", "1/2 cent", "1/2penny", "1/2 penny"], "cent" => ["cent", "penny", "1cent", "1 cent", "1penny", "1 penny", "onepenny", "one penny", "draped bust cent", "lincoln"], "twocent" => ["two cent", "two cents", "twocent", "twocents", "2cent", "2cents", "2 cent", "2 cents", "2penny", "2pennies", "2 penny", "2 pennies"], "threecent" => ["threecent", "threecents", "three", "three cent", "three cents", "3cent", "3cents", "3 cent", "3 cents"], "nickel" => ["nickel", "fivecent", "fivecents", "five cents", "5cent", "5cents"], "dime" => ["dime", "tencent", "tencents", "ten cents", "10cent", "10cents"], "quarter" => ["quarter", "twentyfivecent", "twentyfivecents", "twentyfive cents", "25cent", "25cents"], "halfdollar" => ["halfdollar", "half dollar"], "dollar" => ["dollar", "onedollar", "1dollar", "one dollar", "1 dollar"], "commemorative" => ["commemorative"], "silver" => ["silver", "silver coins", "silver eagles"], "gold" => ["gold", "gold coins"], "gold dollar" => ["gold dollar", "golddollar"], "silver dollar" => ["silver dollar", "silverdollar"], "silver commemorative" => ["silver commemorative", "silvercommemorative"], "gold commemorative" => ["gold commemorative", "goldcommemorative"], "templeton reid" => ["templeton", "templetonreid", "templeton reid"]}
|
171
|
+
|
172
|
+
def initialize(pcgs_no, description, design, grade, price, grade_type, subtype, type, year, mint_mark)
|
173
|
+
@pcgs_no = pcgs_no
|
174
|
+
@description = description
|
175
|
+
@design = design
|
176
|
+
@grade = grade
|
177
|
+
@price = price
|
178
|
+
@grade_type = grade_type
|
179
|
+
@subtype = subtype
|
180
|
+
@type = type
|
181
|
+
@year = year
|
182
|
+
@mint_mark = mint_mark
|
183
|
+
end
|
184
|
+
|
185
|
+
def type_like?(type)
|
186
|
+
type.include?(type)
|
187
|
+
end
|
188
|
+
|
189
|
+
def name_like?(name)
|
190
|
+
name.include?(name)
|
191
|
+
end
|
192
|
+
|
193
|
+
def set_name
|
194
|
+
coin = self
|
195
|
+
possible_names = []
|
196
|
+
PCGS::Coin::NAMES.values.each_index do |i|
|
197
|
+
v = PCGS::Coin::NAMES.values[i]
|
198
|
+
score = 0
|
199
|
+
v.each do |name|
|
200
|
+
n = coin.subtype.downcase.split(" ") & name.split(" ")
|
201
|
+
score += n.size
|
202
|
+
end
|
203
|
+
if score > 0
|
204
|
+
possible_names << [PCGS::Coin::NAMES.keys[i], score]
|
205
|
+
end
|
206
|
+
end
|
207
|
+
coin.name = possible_names.sort_by{|e|e[1]}.reverse.first.first
|
208
|
+
coin
|
209
|
+
end
|
210
|
+
end
|
211
|
+
|
212
|
+
def self.scrape(url)
|
213
|
+
open(URI(url)){|f|Hpricot(f)}
|
214
|
+
end
|
215
|
+
|
216
|
+
def self.table_from_url(url, n)
|
217
|
+
doc = PCGS.scrape(url)
|
218
|
+
table_from_doc(doc, n)
|
219
|
+
end
|
220
|
+
|
221
|
+
def self.table_from_doc(doc, n)
|
222
|
+
doc.search("//div[@id='blue-table']").search("//table")[n]
|
223
|
+
end
|
224
|
+
|
225
|
+
def self.table_info(doc)
|
226
|
+
coin_type = doc.to_s.scan(/<a href=\"\/prices\">Home<\/a>[^<]+</).first.gsub("<a href=\"/prices\">Home</a> > ","").split("\r")[0]
|
227
|
+
tabs = doc.search("//div[@id='blue-table']").search("//ul")
|
228
|
+
tab_titles = tabs.map{|t|get_tabs(t)}
|
229
|
+
tab_links = tabs.map{|tab|tab.search("//a").map{|a|a["href"]}}
|
230
|
+
t = []
|
231
|
+
(0..(tab_links.size-1)).each do |i|
|
232
|
+
tab_links[i].each_index do |j|
|
233
|
+
tab_title = tab_titles[i][j]
|
234
|
+
link = tab_links[i][j]
|
235
|
+
table = table_from_url(link, i)
|
236
|
+
coin_grade_type = ["MS","PR"][i]#doc.to_s.scan(/Copper Type Coins, \S\S/).map{|t|t.gsub("Copper Type Coins, ","")}[i]
|
237
|
+
t << {:coin_type => coin_type, :coin_grade_type => coin_grade_type, :tab_title => tab_title, :link => link, :table => table}
|
238
|
+
end
|
239
|
+
end
|
240
|
+
t
|
241
|
+
end
|
242
|
+
|
243
|
+
def self.get_trs(table)
|
244
|
+
table.search("//tr")
|
245
|
+
end
|
246
|
+
|
247
|
+
def self.get_tds(tr)
|
248
|
+
tr.search("//td")
|
249
|
+
end
|
250
|
+
|
251
|
+
def self.get_ths(tr)
|
252
|
+
tr.search("//th")
|
253
|
+
end
|
254
|
+
|
255
|
+
def self.strip_span(text)
|
256
|
+
if text.include?("<span")
|
257
|
+
text.scan(/<span[^>]*>/).each do |replace|
|
258
|
+
text = text.gsub(replace, " ") rescue text
|
259
|
+
end
|
260
|
+
text = text.gsub("</span>", "")
|
261
|
+
else
|
262
|
+
text
|
263
|
+
end
|
264
|
+
end
|
265
|
+
|
266
|
+
def self.strip_a(text)
|
267
|
+
if text.include?("<a")
|
268
|
+
text.scan(/<a[^>]*>/).each do |replace|
|
269
|
+
text = text.gsub(replace, " ") rescue text
|
270
|
+
end
|
271
|
+
text = text.gsub("</a>", "")
|
272
|
+
else
|
273
|
+
text
|
274
|
+
end
|
275
|
+
end
|
276
|
+
|
277
|
+
def self.strip_br(text)
|
278
|
+
text.gsub("<br />", "")
|
279
|
+
end
|
280
|
+
|
281
|
+
def self.strip_nbsp(text)
|
282
|
+
text.gsub(" ", " ")
|
283
|
+
end
|
284
|
+
|
285
|
+
def self.pick_element(text)
|
286
|
+
if text[-1] == "-"
|
287
|
+
text = text[0..-2]
|
288
|
+
elsif text[-2..-1] == " +"
|
289
|
+
text = text[0..-3]
|
290
|
+
elsif (text.scan(/[\d,]+/).size == 2) && (text.scan(/[^\d^,^ ]/) == [] )
|
291
|
+
text = text.scan(/[\d,]+/)[0]
|
292
|
+
end
|
293
|
+
text
|
294
|
+
end
|
295
|
+
|
296
|
+
def self.convert_to_integer(text)
|
297
|
+
if text.gsub(",","").to_i.to_s == text.gsub(",","")
|
298
|
+
text.gsub(",","").to_i
|
299
|
+
else
|
300
|
+
text
|
301
|
+
end
|
302
|
+
end
|
303
|
+
|
304
|
+
def self.get_rows(table)
|
305
|
+
trs = get_trs(table)
|
306
|
+
rows = []
|
307
|
+
trs.each do |tr|
|
308
|
+
tds = get_tds(tr)
|
309
|
+
tds = get_ths(tr) if tds.empty?
|
310
|
+
row = tds.map do |td|
|
311
|
+
if td.search("//span").empty?
|
312
|
+
r = td.inner_html
|
313
|
+
else
|
314
|
+
r = td.search("//span").first.inner_html
|
315
|
+
r = strip_span(r)
|
316
|
+
end
|
317
|
+
r = strip_a(r)
|
318
|
+
r = strip_br(r)
|
319
|
+
r = strip_nbsp(r)
|
320
|
+
r = r.strip
|
321
|
+
r = pick_element(r)
|
322
|
+
r = convert_to_integer(r)
|
323
|
+
end
|
324
|
+
rows << row
|
325
|
+
end
|
326
|
+
rows
|
327
|
+
end
|
328
|
+
|
329
|
+
def self.get_tabs(tab_data)
|
330
|
+
tab_data.search("//a").map{|a|a.inner_html}
|
331
|
+
end
|
332
|
+
|
333
|
+
def self.objectify_table_info(t)
|
334
|
+
table = PCGS::Table.new(t[:coin_type], t[:tab_title], t[:coin_grade_type], t[:table])
|
335
|
+
rows = PCGS.get_rows(t[:table]) rescue []
|
336
|
+
table.add_rows(rows)
|
337
|
+
table
|
338
|
+
end
|
339
|
+
end
|
metadata
ADDED
@@ -0,0 +1,46 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: pcgs
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
version: 1.0.0
|
5
|
+
prerelease:
|
6
|
+
platform: ruby
|
7
|
+
authors:
|
8
|
+
- Benjamin Godlove
|
9
|
+
autorequire:
|
10
|
+
bindir: bin
|
11
|
+
cert_chain: []
|
12
|
+
date: 2014-02-25 00:00:00.000000000 Z
|
13
|
+
dependencies: []
|
14
|
+
description: ''
|
15
|
+
email: bgodlove88@gmail.com
|
16
|
+
executables: []
|
17
|
+
extensions: []
|
18
|
+
extra_rdoc_files: []
|
19
|
+
files:
|
20
|
+
- lib/pcgs.rb
|
21
|
+
homepage: http://github.com/brg8/pcgs
|
22
|
+
licenses:
|
23
|
+
- GNU
|
24
|
+
post_install_message:
|
25
|
+
rdoc_options: []
|
26
|
+
require_paths:
|
27
|
+
- lib
|
28
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
29
|
+
none: false
|
30
|
+
requirements:
|
31
|
+
- - ! '>='
|
32
|
+
- !ruby/object:Gem::Version
|
33
|
+
version: '0'
|
34
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
35
|
+
none: false
|
36
|
+
requirements:
|
37
|
+
- - ! '>='
|
38
|
+
- !ruby/object:Gem::Version
|
39
|
+
version: '0'
|
40
|
+
requirements: []
|
41
|
+
rubyforge_project:
|
42
|
+
rubygems_version: 1.8.23
|
43
|
+
signing_key:
|
44
|
+
specification_version: 3
|
45
|
+
summary: Pull coin data from the PCGS website at http://pcgs.com/prices
|
46
|
+
test_files: []
|