datahen 0.14.0 → 0.14.5
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/lib/datahen/cli/scraper.rb +5 -3
- data/lib/datahen/cli/scraper_job.rb +8 -2
- data/lib/datahen/cli/scraper_page.rb +1 -1
- data/lib/datahen/client/base.rb +3 -0
- data/lib/datahen/client/job_stat.rb +12 -4
- data/lib/datahen/scraper/executor.rb +9 -3
- data/lib/datahen/version.rb +1 -1
- 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: e81f515b9874f6b9cd2943873e7ce54249d0c4662833e4eb9133bb8827901a3f
|
4
|
+
data.tar.gz: 02f860429555067ab0d034c7e46164ee3537844c34f40ea0949300a2de294187
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 8cffdec9fa56f0d5ee22215b0eb64c0f2de45b1a3f031729f9929625ccf69e270fea2e554a34d583de71e02c7bbc4f0acc82ca6ad5c0f3ea496176277afd8818
|
7
|
+
data.tar.gz: 66b108bc587b3381a6202877f7915653a4ee1c1e058eeeab192aa1ba3614d7843c8dbcb67a7f08d57c6131ac03398a20a09ac7f309acdd0c244f06c46aa84d30
|
data/lib/datahen/cli/scraper.rb
CHANGED
@@ -60,7 +60,7 @@ module Datahen
|
|
60
60
|
desc "show <scraper_name>", "Show a scraper"
|
61
61
|
def show(scraper_name)
|
62
62
|
client = Client::Scraper.new(options)
|
63
|
-
puts "#{client.find(scraper_name
|
63
|
+
puts "#{client.find(scraper_name)}"
|
64
64
|
end
|
65
65
|
|
66
66
|
desc "delete <scraper_name>", "Delete a scraper and related records"
|
@@ -164,12 +164,14 @@ module Datahen
|
|
164
164
|
option :"max-timestamp", type: :string, desc: 'Ending timestamp point in time to query historic stats (inclusive)'
|
165
165
|
option :"limit", type: :numeric, desc: 'Limit stats retrieved'
|
166
166
|
option :"order", type: :numeric, desc: 'Order stats by timestamp [DESC]'
|
167
|
+
option :live, type: :boolean, desc: 'Get data from the live stats history, not cached stats history.'
|
168
|
+
option :filter, type: :string, desc: 'Filter results on `day` or `hour`, if not specified will return all records.'
|
167
169
|
def history(scraper_name)
|
168
170
|
client = Client::JobStat.new(options)
|
169
171
|
if options[:job]
|
170
|
-
puts "#{client.job_stats_history(options[:job])}"
|
172
|
+
puts "#{client.job_stats_history(options[:job], options)}"
|
171
173
|
else
|
172
|
-
puts "#{client.scraper_job_stats_history(scraper_name)}"
|
174
|
+
puts "#{client.scraper_job_stats_history(scraper_name, options)}"
|
173
175
|
end
|
174
176
|
end
|
175
177
|
|
@@ -7,10 +7,16 @@ module Datahen
|
|
7
7
|
end
|
8
8
|
|
9
9
|
desc "show <scraper_name>", "Show a scraper's current job (Defaults to showing data from cached job)"
|
10
|
+
option :job, :aliases => :j, type: :numeric, desc: 'Set a specific job ID'
|
10
11
|
option :live, type: :boolean, desc: 'Get data from the live job, not cached job.'
|
11
12
|
def show(scraper_name)
|
12
|
-
|
13
|
-
|
13
|
+
if options[:job]
|
14
|
+
client = Client::Job.new(options)
|
15
|
+
puts "#{client.find(options[:job], options)}"
|
16
|
+
else
|
17
|
+
client = Client::ScraperJob.new(options)
|
18
|
+
puts "#{client.find(scraper_name, options)}"
|
19
|
+
end
|
14
20
|
end
|
15
21
|
|
16
22
|
|
@@ -230,7 +230,7 @@ module Datahen
|
|
230
230
|
end
|
231
231
|
end
|
232
232
|
|
233
|
-
desc "failedcontent <gid>", "Show a page's failed content in scraper's current job"
|
233
|
+
desc "failedcontent <scraper_name> <gid>", "Show a page's failed content in scraper's current job"
|
234
234
|
option :job, :aliases => :j, type: :numeric, desc: 'Set a specific job ID'
|
235
235
|
def failedcontent(scraper_name, gid)
|
236
236
|
result = nil
|
data/lib/datahen/client/base.rb
CHANGED
@@ -5,6 +5,8 @@ module Datahen
|
|
5
5
|
class Base
|
6
6
|
include HTTParty
|
7
7
|
|
8
|
+
default_timeout 60
|
9
|
+
|
8
10
|
def self.env_auth_token
|
9
11
|
ENV['DATAHEN_TOKEN']
|
10
12
|
end
|
@@ -55,6 +57,7 @@ module Datahen
|
|
55
57
|
query[:"max-timestamp"] = opts[:"max-timestamp"] if opts[:"max-timestamp"]
|
56
58
|
query[:limit] = opts[:limit] if opts[:limit]
|
57
59
|
query[:order] = opts[:order] if opts[:order]
|
60
|
+
query[:filter] = opts[:filter] if opts[:filter]
|
58
61
|
|
59
62
|
if opts[:query]
|
60
63
|
if opts[:query].is_a?(Hash)
|
@@ -18,12 +18,20 @@ module Datahen
|
|
18
18
|
end
|
19
19
|
end
|
20
20
|
|
21
|
-
def job_stats_history(job_id)
|
22
|
-
|
21
|
+
def job_stats_history(job_id, opts={})
|
22
|
+
if opts[:live]
|
23
|
+
self.class.get("/jobs/#{job_id}/stats/history", @options)
|
24
|
+
else
|
25
|
+
self.class.get("/cached/jobs/#{job_id}/stats/history", @options)
|
26
|
+
end
|
23
27
|
end
|
24
28
|
|
25
|
-
def scraper_job_stats_history(scraper_name)
|
26
|
-
|
29
|
+
def scraper_job_stats_history(scraper_name, opts={})
|
30
|
+
if opts[:live]
|
31
|
+
self.class.get("/scrapers/#{scraper_name}/current_job/stats/history", @options)
|
32
|
+
else
|
33
|
+
self.class.get("/cached/scrapers/#{scraper_name}/current_job/stats/history", @options)
|
34
|
+
end
|
27
35
|
end
|
28
36
|
|
29
37
|
end
|
@@ -60,7 +60,12 @@ module Datahen
|
|
60
60
|
|
61
61
|
def init_global_page()
|
62
62
|
client = Client::GlobalPage.new()
|
63
|
-
client.find(gid)
|
63
|
+
global_page = client.find(gid)
|
64
|
+
unless global_page.code == 200
|
65
|
+
raise "GID #{gid} not found. Aborting execution!"
|
66
|
+
else
|
67
|
+
global_page
|
68
|
+
end
|
64
69
|
end
|
65
70
|
|
66
71
|
def get_content(job_id, gid)
|
@@ -287,11 +292,12 @@ module Datahen
|
|
287
292
|
end
|
288
293
|
|
289
294
|
# behave differently if it is a real save
|
295
|
+
save_status = status
|
290
296
|
if save
|
291
297
|
log_msg = "Saving #{log_msgs.join(' and ')}."
|
292
298
|
puts "#{log_msg}"
|
293
299
|
else
|
294
|
-
|
300
|
+
save_status = "#{status}_try"
|
295
301
|
end
|
296
302
|
|
297
303
|
# saving to server
|
@@ -300,7 +306,7 @@ module Datahen
|
|
300
306
|
gid: gid,
|
301
307
|
pages: pages_slice,
|
302
308
|
outputs: outputs_slice,
|
303
|
-
status:
|
309
|
+
status: save_status)
|
304
310
|
|
305
311
|
if response.code == 200
|
306
312
|
if save
|
data/lib/datahen/version.rb
CHANGED
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: datahen
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.14.
|
4
|
+
version: 0.14.5
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Parama Danoesubroto
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date: 2020-
|
11
|
+
date: 2020-08-04 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: thor
|