publicstorage 0.3.0 → 1.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/README.md +8 -2
- data/lib/publicstorage/cli.rb +9 -3
- data/lib/publicstorage/config.rb +1 -1
- data/lib/publicstorage/crawl.rb +40 -0
- data/lib/publicstorage/crawler.rb +0 -9
- data/lib/publicstorage/dimensions.rb +25 -14
- data/lib/publicstorage/fetch_error.rb +12 -0
- data/lib/publicstorage/price.rb +6 -2
- data/lib/publicstorage/rates.rb +4 -7
- data/lib/publicstorage/version.rb +1 -1
- metadata +7 -4
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: f797ad8918b9cea8da35957e83b8bbf56ab02f1bfae95ee24a0fe1a316d1117d
|
4
|
+
data.tar.gz: e2a976a36b0fc0d5edefde19c7234b778257341abc9c9507029bac2d6e31f1e3
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: eb338e5697de874ba118f47828d19ed3ed817cd136aac67372064fcea2dabde722c34cb39924b55e91bbbca3802b8469b7caf6a5fef5fa64abf559a2c2e38090
|
7
|
+
data.tar.gz: 8ed743ff4a0a59deefe92064d62345d6deaaceb860e597583758b8e7d18eac6f0db6b12f67ef0324212882ef1fea14009501b2ab0566d139241ae4d78d60f7eb
|
data/README.md
CHANGED
@@ -1,4 +1,4 @@
|
|
1
|
-
#
|
1
|
+
# Public Storage
|
2
2
|
|
3
3
|
[![LICENSE](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/ksylvest/publicstorage/blob/main/LICENSE)
|
4
4
|
[![RubyGems](https://img.shields.io/gem/v/publicstorage)](https://rubygems.org/gems/publicstorage)
|
@@ -6,6 +6,8 @@
|
|
6
6
|
[![Yard](https://img.shields.io/badge/docs-site-blue.svg)](https://publicstorage.ksylvest.com)
|
7
7
|
[![CircleCI](https://img.shields.io/circleci/build/github/ksylvest/publicstorage)](https://circleci.com/gh/ksylvest/publicstorage)
|
8
8
|
|
9
|
+
A Ruby library offering both a CLI and API for scraping [Public Storage](https://www.publicstorage.com/) self-storage facilities and prices.
|
10
|
+
|
9
11
|
## Installation
|
10
12
|
|
11
13
|
```bash
|
@@ -31,7 +33,7 @@ require 'publicstorage'
|
|
31
33
|
sitemap = PublicStorage::Facility.sitemap
|
32
34
|
sitemap.links.each do |link|
|
33
35
|
url = link.loc
|
34
|
-
facility =
|
36
|
+
facility = PublicStorage::Facility.fetch(url:)
|
35
37
|
|
36
38
|
puts facility.text
|
37
39
|
|
@@ -48,3 +50,7 @@ end
|
|
48
50
|
```bash
|
49
51
|
publicstorage crawl
|
50
52
|
```
|
53
|
+
|
54
|
+
```bash
|
55
|
+
publicstorage crawl "https://www.publicstorage.com/self-storage-ca-venice/120.html"
|
56
|
+
```
|
data/lib/publicstorage/cli.rb
CHANGED
@@ -21,7 +21,7 @@ module PublicStorage
|
|
21
21
|
command = argv.shift
|
22
22
|
|
23
23
|
case command
|
24
|
-
when 'crawl' then crawl
|
24
|
+
when 'crawl' then crawl(*argv)
|
25
25
|
else
|
26
26
|
warn("unsupported command=#{command.inspect}")
|
27
27
|
exit(Code::ERROR)
|
@@ -30,8 +30,9 @@ module PublicStorage
|
|
30
30
|
|
31
31
|
private
|
32
32
|
|
33
|
-
|
34
|
-
|
33
|
+
# @url [String] optional
|
34
|
+
def crawl(url = nil)
|
35
|
+
Crawl.run(url: url)
|
35
36
|
exit(Code::OK)
|
36
37
|
end
|
37
38
|
|
@@ -52,6 +53,11 @@ module PublicStorage
|
|
52
53
|
|
53
54
|
options.on('-h', '--help', 'help') { help(options) }
|
54
55
|
options.on('-v', '--version', 'version') { version }
|
56
|
+
|
57
|
+
options.separator <<~COMMANDS
|
58
|
+
commands:
|
59
|
+
crawl [url]
|
60
|
+
COMMANDS
|
55
61
|
end
|
56
62
|
end
|
57
63
|
end
|
data/lib/publicstorage/config.rb
CHANGED
@@ -12,7 +12,7 @@ module PublicStorage
|
|
12
12
|
attr_accessor :timeout
|
13
13
|
|
14
14
|
def initialize
|
15
|
-
@user_agent = ENV.fetch('
|
15
|
+
@user_agent = ENV.fetch('PUBLICSTORAGE_USER_AGENT', "publicstorage.rb/#{VERSION}")
|
16
16
|
@timeout = Integer(ENV.fetch('PUBLICSTORAGE_TIMEOUT', 60))
|
17
17
|
end
|
18
18
|
end
|
@@ -0,0 +1,40 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module PublicStorage
|
4
|
+
# Handles the crawl command via CLI.
|
5
|
+
class Crawl
|
6
|
+
def self.run(...)
|
7
|
+
new(...).run
|
8
|
+
end
|
9
|
+
|
10
|
+
# @param stdout [IO] optional
|
11
|
+
# @param stderr [IO] optional
|
12
|
+
# @param url [String] optional
|
13
|
+
def initialize(stdout: $stdout, stderr: $stderr, url: nil)
|
14
|
+
@stdout = stdout
|
15
|
+
@stderr = stderr
|
16
|
+
@url = url
|
17
|
+
end
|
18
|
+
|
19
|
+
def run
|
20
|
+
if @url
|
21
|
+
process(url: @url)
|
22
|
+
else
|
23
|
+
sitemap = Facility.sitemap
|
24
|
+
@stdout.puts("count=#{sitemap.links.count}")
|
25
|
+
@stdout.puts
|
26
|
+
sitemap.links.each { |link| process(url: link.loc) }
|
27
|
+
end
|
28
|
+
end
|
29
|
+
|
30
|
+
def process(url:)
|
31
|
+
@stdout.puts(url)
|
32
|
+
facility = Facility.fetch(url: url)
|
33
|
+
@stdout.puts(facility.text)
|
34
|
+
facility.prices.each { |price| @stdout.puts(price.text) }
|
35
|
+
@stdout.puts
|
36
|
+
rescue FetchError => e
|
37
|
+
@stderr.puts("url=#{url} error=#{e.message}")
|
38
|
+
end
|
39
|
+
end
|
40
|
+
end
|
@@ -5,15 +5,6 @@ module PublicStorage
|
|
5
5
|
class Crawler
|
6
6
|
HOST = 'https://www.publicstorage.com'
|
7
7
|
|
8
|
-
# Raised for unexpected HTTP responses.
|
9
|
-
class FetchError < StandardError
|
10
|
-
# @param url [String]
|
11
|
-
# @param response [HTTP::Response]
|
12
|
-
def initialize(url:, response:)
|
13
|
-
super("url=#{url} status=#{response.status.inspect} body=#{response.body.inspect}")
|
14
|
-
end
|
15
|
-
end
|
16
|
-
|
17
8
|
# @param url [String]
|
18
9
|
# @raise [FetchError]
|
19
10
|
# @return [Nokogiri::HTML::Document]
|
@@ -1,8 +1,10 @@
|
|
1
1
|
# frozen_string_literal: true
|
2
2
|
|
3
3
|
module PublicStorage
|
4
|
-
# The dimensions (width + depth +
|
4
|
+
# The dimensions (width + depth + height) of a price.
|
5
5
|
class Dimensions
|
6
|
+
DEFAULT_HEIGHT = 8.0 # feet
|
7
|
+
|
6
8
|
SELECTOR = '.unit-size'
|
7
9
|
|
8
10
|
# @attribute [rw] depth
|
@@ -13,17 +15,17 @@ module PublicStorage
|
|
13
15
|
# @return [Integer]
|
14
16
|
attr_accessor :width
|
15
17
|
|
16
|
-
# @attribute [rw]
|
17
|
-
#
|
18
|
-
attr_accessor :
|
18
|
+
# @attribute [rw] height
|
19
|
+
# @return [Integer]
|
20
|
+
attr_accessor :height
|
19
21
|
|
20
22
|
# @param depth [Float]
|
21
23
|
# @param width [Float]
|
22
|
-
# @param
|
23
|
-
def initialize(depth:, width:,
|
24
|
+
# @param height [Float]
|
25
|
+
def initialize(depth:, width:, height: DEFAULT_HEIGHT)
|
24
26
|
@depth = depth
|
25
27
|
@width = width
|
26
|
-
@
|
28
|
+
@height = height
|
27
29
|
end
|
28
30
|
|
29
31
|
# @return [String]
|
@@ -31,26 +33,35 @@ module PublicStorage
|
|
31
33
|
props = [
|
32
34
|
"depth=#{@depth.inspect}",
|
33
35
|
"width=#{@width.inspect}",
|
34
|
-
"
|
36
|
+
"height=#{@height.inspect}"
|
35
37
|
]
|
36
38
|
"#<#{self.class.name} #{props.join(' ')}>"
|
37
39
|
end
|
38
40
|
|
39
41
|
# @return [String] e.g. "10' × 10' (100 sqft)"
|
40
42
|
def text
|
41
|
-
"#{format('%g', @width)}' × #{format('%g', @depth)}' (#{
|
43
|
+
"#{format('%g', @width)}' × #{format('%g', @depth)}' (#{sqft} sqft)"
|
44
|
+
end
|
45
|
+
|
46
|
+
# @return [Integer]
|
47
|
+
def sqft
|
48
|
+
Integer(@width * @depth)
|
49
|
+
end
|
50
|
+
|
51
|
+
# @return [Integer]
|
52
|
+
def cuft
|
53
|
+
Integer(@width * @depth * @height)
|
42
54
|
end
|
43
55
|
|
44
|
-
# @param
|
56
|
+
# @param data [Hash]
|
45
57
|
#
|
46
58
|
# @return [Dimensions]
|
47
|
-
def self.parse(
|
48
|
-
match =
|
59
|
+
def self.parse(data:)
|
60
|
+
match = data['dimension'].match(/(?<depth>[\d\.]+)'x(?<width>[\d\.]+)'/)
|
49
61
|
depth = Float(match[:depth])
|
50
62
|
width = Float(match[:width])
|
51
|
-
sqft = Integer(depth * width)
|
52
63
|
|
53
|
-
new(depth:, width:,
|
64
|
+
new(depth:, width:, height: DEFAULT_HEIGHT)
|
54
65
|
end
|
55
66
|
end
|
56
67
|
end
|
@@ -0,0 +1,12 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module PublicStorage
|
4
|
+
# Raised for unexpected HTTP responses.
|
5
|
+
class FetchError < StandardError
|
6
|
+
# @param url [String]
|
7
|
+
# @param response [HTTP::Response]
|
8
|
+
def initialize(url:, response:)
|
9
|
+
super("url=#{url} status=#{response.status.inspect} body=#{response.body.inspect}")
|
10
|
+
end
|
11
|
+
end
|
12
|
+
end
|
data/lib/publicstorage/price.rb
CHANGED
@@ -3,6 +3,8 @@
|
|
3
3
|
module PublicStorage
|
4
4
|
# The price (id + dimensions + rate) for a facility
|
5
5
|
class Price
|
6
|
+
GTM_SELECTOR = 'button[data-gtmdata]'
|
7
|
+
|
6
8
|
# @attribute [rw] id
|
7
9
|
# @return [String]
|
8
10
|
attr_accessor :id
|
@@ -43,8 +45,10 @@ module PublicStorage
|
|
43
45
|
#
|
44
46
|
# @return [Price]
|
45
47
|
def self.parse(element:)
|
46
|
-
|
47
|
-
|
48
|
+
data = JSON.parse(element.at(GTM_SELECTOR).attribute('data-gtmdata'))
|
49
|
+
|
50
|
+
rates = Rates.parse(data:)
|
51
|
+
dimensions = Dimensions.parse(data:)
|
48
52
|
|
49
53
|
new(
|
50
54
|
id: element.attr('data-unitid'),
|
data/lib/publicstorage/rates.rb
CHANGED
@@ -3,9 +3,6 @@
|
|
3
3
|
module PublicStorage
|
4
4
|
# The rates (street + web) for a facility
|
5
5
|
class Rates
|
6
|
-
STREET_SELECTOR = '.unit-prices .unit-pricing .unit-strike-through-price'
|
7
|
-
WEB_SELECTOR = '.unit-prices .unit-pricing .unit-price'
|
8
|
-
|
9
6
|
# @attribute [rw] street
|
10
7
|
# @return [Integer]
|
11
8
|
attr_accessor :street
|
@@ -35,12 +32,12 @@ module PublicStorage
|
|
35
32
|
"$#{@street} (street) | $#{@web} (web)"
|
36
33
|
end
|
37
34
|
|
38
|
-
# @param
|
35
|
+
# @param data [Hash]
|
39
36
|
#
|
40
37
|
# @return [Rates]
|
41
|
-
def self.parse(
|
42
|
-
street =
|
43
|
-
web =
|
38
|
+
def self.parse(data:)
|
39
|
+
street = data['listprice']
|
40
|
+
web = data['saleprice']
|
44
41
|
new(street:, web:)
|
45
42
|
end
|
46
43
|
end
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: publicstorage
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version:
|
4
|
+
version: 1.1.0
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Kevin Sylvestre
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date:
|
11
|
+
date: 2025-01-08 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: http
|
@@ -83,9 +83,11 @@ files:
|
|
83
83
|
- lib/publicstorage/address.rb
|
84
84
|
- lib/publicstorage/cli.rb
|
85
85
|
- lib/publicstorage/config.rb
|
86
|
+
- lib/publicstorage/crawl.rb
|
86
87
|
- lib/publicstorage/crawler.rb
|
87
88
|
- lib/publicstorage/dimensions.rb
|
88
89
|
- lib/publicstorage/facility.rb
|
90
|
+
- lib/publicstorage/fetch_error.rb
|
89
91
|
- lib/publicstorage/geocode.rb
|
90
92
|
- lib/publicstorage/link.rb
|
91
93
|
- lib/publicstorage/price.rb
|
@@ -98,8 +100,9 @@ licenses:
|
|
98
100
|
metadata:
|
99
101
|
rubygems_mfa_required: 'true'
|
100
102
|
homepage_uri: https://github.com/ksylvest/publicstorage
|
101
|
-
source_code_uri: https://github.com/ksylvest/publicstorage
|
102
|
-
changelog_uri: https://github.com/ksylvest/publicstorage
|
103
|
+
source_code_uri: https://github.com/ksylvest/publicstorage/tree/v1.1.0
|
104
|
+
changelog_uri: https://github.com/ksylvest/publicstorage/releases/tag/v1.1.0
|
105
|
+
documentation_uri: https://publicstorage.ksylvest.com/
|
103
106
|
post_install_message:
|
104
107
|
rdoc_options: []
|
105
108
|
require_paths:
|