zorki 0.1.20 → 0.1.21
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/lib/zorki/scrapers/post_scraper.rb +22 -0
- data/lib/zorki/scrapers/scraper.rb +7 -2
- data/lib/zorki/version.rb +1 -1
- data/lib/zorki.rb +24 -0
- metadata +2 -2
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: '0228694f5aafe38a9856efe3a4f9f042b94aeecbd5cb7ec919326f48c0b8000e'
|
4
|
+
data.tar.gz: c54da1aa5b98d3166a047a24022b74356538e7dcb974e032bc5236065c5edd4c
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 30b2953e778ce0a6e36f221350c6760974de7279c764111cfe2327f3bcc5f11a93d19a4747d20c35260038e07c06bad2c3d2fc704b4878c23a41ba309ae689d9
|
7
|
+
data.tar.gz: 99ed979bcc6f4f0758bca49d34580333ba42c35d74ce3ccfe52871acb9665d3841e0d54ed68bb3af2cf1ef366ebb1f3f9ea9136784958778751d7142f09e5b38
|
@@ -5,6 +5,28 @@ require "typhoeus"
|
|
5
5
|
module Zorki
|
6
6
|
class PostScraper < Scraper
|
7
7
|
def parse(id)
|
8
|
+
count = 0
|
9
|
+
|
10
|
+
until count == 2
|
11
|
+
puts "Retrieving ID #{id}"
|
12
|
+
|
13
|
+
begin
|
14
|
+
result = attempt_parse(id)
|
15
|
+
break
|
16
|
+
rescue ImageRequestZeroSize
|
17
|
+
debugger
|
18
|
+
# If the image is zero size, we retry
|
19
|
+
puts "Zero sized image found, retrying #{count}"
|
20
|
+
count += 1
|
21
|
+
end
|
22
|
+
end
|
23
|
+
|
24
|
+
raise ImageRequestZeroSize if count == 5
|
25
|
+
|
26
|
+
result
|
27
|
+
end
|
28
|
+
|
29
|
+
def attempt_parse(id)
|
8
30
|
# Stuff we need to get from the DOM (implemented is starred):
|
9
31
|
# - User *
|
10
32
|
# - Text *
|
@@ -172,6 +172,7 @@ module Zorki
|
|
172
172
|
|
173
173
|
def login
|
174
174
|
puts "Attempting to login..."
|
175
|
+
|
175
176
|
# Reset the sessions so that there's nothing laying around
|
176
177
|
# page.driver.browser.close
|
177
178
|
|
@@ -185,7 +186,10 @@ module Zorki
|
|
185
186
|
|
186
187
|
# We don't have to login if we already are
|
187
188
|
begin
|
188
|
-
|
189
|
+
if find_field("Search", wait: 10).present?
|
190
|
+
puts "Already logged in"
|
191
|
+
return
|
192
|
+
end
|
189
193
|
rescue Capybara::ElementNotFound; end
|
190
194
|
|
191
195
|
# Check if we're redirected to a login page, if we aren't we're already logged in
|
@@ -213,9 +217,10 @@ module Zorki
|
|
213
217
|
# Sometimes Instagram just... doesn't let you log in
|
214
218
|
raise "Instagram not accessible" if loop_count == 5
|
215
219
|
|
220
|
+
puts "Login successful"
|
216
221
|
# No we don't want to save our login credentials
|
217
222
|
begin
|
218
|
-
puts "Checking and clearing Save Info button
|
223
|
+
puts "Checking and clearing Save Info button"
|
219
224
|
|
220
225
|
find_button("Save Info").click()
|
221
226
|
rescue Capybara::ElementNotFound; end
|
data/lib/zorki/version.rb
CHANGED
data/lib/zorki.rb
CHANGED
@@ -46,11 +46,34 @@ module Zorki
|
|
46
46
|
end
|
47
47
|
end
|
48
48
|
|
49
|
+
class ImageRequestZeroSize < RetryableError; end
|
50
|
+
|
49
51
|
define_setting :temp_storage_location, "tmp/zorki"
|
50
52
|
|
51
53
|
# Get an image from a URL and save to a temp folder set in the configuration under
|
52
54
|
# temp_storage_location
|
55
|
+
|
56
|
+
# We do this because sometimes the images are coming back sized zero
|
53
57
|
def self.retrieve_media(url)
|
58
|
+
count = 0
|
59
|
+
|
60
|
+
until count == 5
|
61
|
+
temp_file_name = attempt_retrieve_media(url)
|
62
|
+
|
63
|
+
# If it's more than 1kb return properly
|
64
|
+
return temp_file_name if File.size(temp_file_name) > 100
|
65
|
+
|
66
|
+
# Delete the file since we want to retry
|
67
|
+
debugger
|
68
|
+
|
69
|
+
File.delete(temp_file_name)
|
70
|
+
count += 1
|
71
|
+
end
|
72
|
+
|
73
|
+
raise(ImageRequestZeroSize)
|
74
|
+
end
|
75
|
+
|
76
|
+
def self.attempt_retrieve_media(url)
|
54
77
|
response = Typhoeus.get(url)
|
55
78
|
|
56
79
|
# Get the file extension if it's in the file
|
@@ -69,6 +92,7 @@ module Zorki
|
|
69
92
|
# We do this in case the folder isn't created yet, since it's a temp folder we'll just do so
|
70
93
|
self.create_temp_storage_location
|
71
94
|
File.binwrite(temp_file_name, response.body)
|
95
|
+
|
72
96
|
temp_file_name
|
73
97
|
end
|
74
98
|
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: zorki
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.1.
|
4
|
+
version: 0.1.21
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Christopher Guess
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date: 2024-03-
|
11
|
+
date: 2024-03-29 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: capybara
|