grucrawler 0.0.1 → 0.0.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/README.md +3 -1
- data/lib/grucrawler/queue.rb +3 -4
- data/lib/grucrawler/version.rb +1 -1
- data/lib/grucrawler.rb +2 -1
- metadata +1 -1
checksums.yaml
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
---
|
|
2
2
|
SHA1:
|
|
3
|
-
metadata.gz:
|
|
4
|
-
data.tar.gz:
|
|
3
|
+
metadata.gz: c5723aa43bafd68af0aa0499c9f748bfca57a580
|
|
4
|
+
data.tar.gz: 3e7046a91e6594c17e987bfb8d934402ee2826dc
|
|
5
5
|
SHA512:
|
|
6
|
-
metadata.gz:
|
|
7
|
-
data.tar.gz:
|
|
6
|
+
metadata.gz: b17e8fdd06816a6e4a1294a9dbf384d95519d1f4046d2d1c4bbf2e4c5821b4839f26e4290e4d655431a1f23beb9fb918c729ae34c6d56c4a6e71e0daf67a6520
|
|
7
|
+
data.tar.gz: 1f253a2fe81b1ae537a0c4de4ed9626261f8f2b72567959839550b883cf41b0c952e164ca1b4ff7a8d1a8378cb7a76a6acfcee9573da48219164069659da67df
|
data/README.md
CHANGED
|
@@ -2,13 +2,15 @@
|
|
|
2
2
|
|
|
3
3
|
```ruby
|
|
4
4
|
require 'grucrawler'
|
|
5
|
+
require 'colorize'
|
|
5
6
|
|
|
6
7
|
class ItalianCrawler
|
|
7
8
|
def options
|
|
8
9
|
{
|
|
9
10
|
visit_urls_only_once: true,
|
|
10
11
|
follow_redirects: true,
|
|
11
|
-
concurrency: 5
|
|
12
|
+
concurrency: 5,
|
|
13
|
+
domain_wait: 20 # seconds between visits to the same domain
|
|
12
14
|
}
|
|
13
15
|
end
|
|
14
16
|
|
data/lib/grucrawler/queue.rb
CHANGED
|
@@ -7,13 +7,14 @@ class GruCrawler
|
|
|
7
7
|
DOMAIN_VISITS_KEY = 'domain_visits'
|
|
8
8
|
QUEUE_KEY = 'queue'
|
|
9
9
|
|
|
10
|
-
def initialize(namespace, visit_once)
|
|
10
|
+
def initialize(namespace, visit_once, domain_wait)
|
|
11
11
|
@redis = Redis.new
|
|
12
12
|
@rns = namespace + ':'
|
|
13
13
|
@concurrent_requests = 0
|
|
14
14
|
@tmp_block = {}
|
|
15
15
|
@domains_throttle = Hash.new(0.0)
|
|
16
16
|
@visit_once = visit_once
|
|
17
|
+
@domain_wait = domain_wait
|
|
17
18
|
end
|
|
18
19
|
|
|
19
20
|
def reset
|
|
@@ -41,15 +42,13 @@ class GruCrawler
|
|
|
41
42
|
url
|
|
42
43
|
end
|
|
43
44
|
|
|
44
|
-
MIN_TIME_TO_WAIT = 20
|
|
45
|
-
|
|
46
45
|
def can_visit_now(url)
|
|
47
46
|
return false if @tmp_block[url]
|
|
48
47
|
|
|
49
48
|
last_visit = last_visit_to_domain(url)
|
|
50
49
|
time_passed = Time.now.to_f - last_visit
|
|
51
50
|
|
|
52
|
-
time_passed >
|
|
51
|
+
time_passed > @domain_wait
|
|
53
52
|
end
|
|
54
53
|
|
|
55
54
|
def started(url)
|
data/lib/grucrawler/version.rb
CHANGED
data/lib/grucrawler.rb
CHANGED
|
@@ -12,7 +12,8 @@ class GruCrawler
|
|
|
12
12
|
def initialize(rules)
|
|
13
13
|
@crawler = rules
|
|
14
14
|
@options = @crawler.options()
|
|
15
|
-
|
|
15
|
+
domain_wait = @options[:domain_wait] || 20
|
|
16
|
+
@queue = GruCrawler::Queue.new(@crawler.class.name, @options[:visit_urls_only_once], domain_wait)
|
|
16
17
|
|
|
17
18
|
@crawler.on_init(self)
|
|
18
19
|
end
|