nhkore 0.3.3 → 0.3.4
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +22 -1
- data/README.md +35 -11
- data/Rakefile +25 -0
- data/lib/nhkore/cli/news_cmd.rb +2 -1
- data/lib/nhkore/cli/search_cmd.rb +2 -2
- data/lib/nhkore/cli/sift_cmd.rb +6 -111
- data/lib/nhkore/datetime_parser.rb +342 -0
- data/lib/nhkore/lib.rb +1 -0
- data/lib/nhkore/util.rb +0 -17
- data/lib/nhkore/version.rb +1 -1
- data/nhkore.gemspec +1 -1
- metadata +6 -5
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 1db415fe5fa2d6f112fae3ef163b79ddaa88d639de565c4cb35b40fa14fb5a66
|
4
|
+
data.tar.gz: c535ba1f719db2c64ba6713c21c6365f51e31361efd6046064cdf700219f30ab
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: af517bada681dc850e60f1c0e275ba35d2177dc413b076612a091217fd653bba82be725c0247c52bb1e342489fe99c731ac7f5efc0b681d54cdcbb02cdb5360b
|
7
|
+
data.tar.gz: 579569f8c5201f3fe8b460281d585ac8b10c50c881d18ca6d6ff342fab0c178f0ad6e532df42200db117b0f8e85dbbace87846c85fcdd5c0c4eac1501f4b24e7
|
data/CHANGELOG.md
CHANGED
@@ -2,7 +2,28 @@
|
|
2
2
|
|
3
3
|
Format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
|
4
4
|
|
5
|
-
## [[Unreleased]](https://github.com/esotericpig/nhkore/compare/v0.3.
|
5
|
+
## [[Unreleased]](https://github.com/esotericpig/nhkore/compare/v0.3.4...master)
|
6
|
+
|
7
|
+
## [v0.3.4] - 2020-04-25
|
8
|
+
|
9
|
+
### Added
|
10
|
+
- DatetimeParser
|
11
|
+
- Extracted from SiftCmd into its own class
|
12
|
+
- Fixed some minor logic bugs from the old code
|
13
|
+
- Added new feature where 1 range can be empty:
|
14
|
+
- `sift ez -d '...2019'` (from = 1924)
|
15
|
+
- `sift ez -d '2019...'` (to = current year)
|
16
|
+
- `sift ez -d '...'` (still an error)
|
17
|
+
- Added `update_core` rake task for dev
|
18
|
+
- Makes pushing a new release much easier
|
19
|
+
- See *Hacking.Releasing* section in *README*
|
20
|
+
|
21
|
+
### Fixed
|
22
|
+
- SiftCmd `parse_sift_datetime()` for `-d/--datetime` option
|
23
|
+
- Didn't work exactly right (as written in *README*) for some special inputs:
|
24
|
+
- `-d '2019...3'`
|
25
|
+
- `-d '3-3'`
|
26
|
+
- `-d '3'`
|
6
27
|
|
7
28
|
## [v0.3.3] - 2020-04-23
|
8
29
|
|
data/README.md
CHANGED
@@ -26,6 +26,8 @@ This is similar to a [core word/vocabulary list](https://www.fluentin3months.com
|
|
26
26
|
- [News Command](#news-command-)
|
27
27
|
- [Using the Library](#using-the-library-)
|
28
28
|
- [Hacking](#hacking-)
|
29
|
+
- [Updating](#updating-)
|
30
|
+
- [Releasing](#releasing-)
|
29
31
|
- [License](#license-)
|
30
32
|
|
31
33
|
## For Non-Power Users [^](#contents)
|
@@ -525,7 +527,7 @@ doc = ss.html_doc()
|
|
525
527
|
doc.css('a').each() do |anchor|
|
526
528
|
link = anchor['href']
|
527
529
|
|
528
|
-
next if ss.ignore_link?(link)
|
530
|
+
next if ss.ignore_link?(link,cleaned: false)
|
529
531
|
|
530
532
|
if link.include?('https://www3.nhk')
|
531
533
|
puts link
|
@@ -564,6 +566,7 @@ end
|
|
564
566
|
|
565
567
|
```Ruby
|
566
568
|
require 'nhkore/article_scraper'
|
569
|
+
require 'time'
|
567
570
|
|
568
571
|
as = NHKore::ArticleScraper.new(
|
569
572
|
'https://www3.nhk.or.jp/news/easy/k10011862381000/k10011862381000.html',
|
@@ -687,6 +690,7 @@ end
|
|
687
690
|
`Sifter` will sift & sort the `News` data into a single file. The data is sorted by frequency in descending order (i.e., most frequent words first).
|
688
691
|
|
689
692
|
```Ruby
|
693
|
+
require 'nhkore/datetime_parser'
|
690
694
|
require 'nhkore/news'
|
691
695
|
require 'nhkore/sifter'
|
692
696
|
require 'time'
|
@@ -698,7 +702,8 @@ sifter = NHKore::Sifter.new(news)
|
|
698
702
|
sifter.caption = 'Sakura Fields Forever!'
|
699
703
|
|
700
704
|
# Filter the data.
|
701
|
-
|
705
|
+
sifter.filter_by_datetime(NHKore::DatetimeParser.parse_range('2019-12-4...7'))
|
706
|
+
sifter.filter_by_datetime([Time.new(2019,12,4),Time.new(2019,12,7)])
|
702
707
|
sifter.filter_by_datetime(
|
703
708
|
from: Time.new(2019,12,4),to: Time.new(2019,12,7)
|
704
709
|
)
|
@@ -727,13 +732,14 @@ if !File.exist?(file)
|
|
727
732
|
end
|
728
733
|
```
|
729
734
|
|
730
|
-
### Util &
|
735
|
+
### Util, UserAgents, & DatetimeParser
|
731
736
|
|
732
737
|
These provide a variety of useful methods/constants.
|
733
738
|
|
734
739
|
Here are some of the most useful ones:
|
735
740
|
|
736
741
|
```Ruby
|
742
|
+
require 'nhkore/datetime_parser'
|
737
743
|
require 'nhkore/user_agents'
|
738
744
|
require 'nhkore/util'
|
739
745
|
|
@@ -759,14 +765,16 @@ puts
|
|
759
765
|
puts '========'
|
760
766
|
puts '[ Time ]'
|
761
767
|
puts '========'
|
762
|
-
puts "JST now: #{Util.jst_now}"
|
768
|
+
puts "JST now: #{Util.jst_now()}"
|
763
769
|
# Drops in JST_OFFSET, does not change hour/min.
|
764
770
|
puts "JST time: #{Util.jst_time(Time.now)}"
|
765
771
|
puts "JST year: #{Util::JST_YEAR}"
|
766
772
|
puts "1999 sane? #{Util.sane_year?(1999)}" # true
|
767
773
|
puts "1776 sane? #{Util.sane_year?(1776)}" # false
|
768
|
-
puts "Guess 5: #{
|
769
|
-
puts "Guess 99: #{
|
774
|
+
puts "Guess 5: #{DatetimeParser.guess_year(5)}" # 2005
|
775
|
+
puts "Guess 99: #{DatetimeParser.guess_year(99)}" # 1999
|
776
|
+
# => [2020-12-01 00:00:00 +0900, 2020-12-31 23:59:59 +0900]
|
777
|
+
puts "Parse: #{DatetimeParser.parse_range('2020-12')}"
|
770
778
|
puts
|
771
779
|
puts "JST timezone offset: #{Util::JST_OFFSET}"
|
772
780
|
puts "JST timezone offset hour: #{Util::JST_OFFSET_HOUR}"
|
@@ -792,9 +800,9 @@ def fmt_jpn()
|
|
792
800
|
end
|
793
801
|
|
794
802
|
puts " #{fmt_jpn{|x| x}}"
|
795
|
-
puts "Hiragana? #{fmt_jpn{|x|
|
796
|
-
puts "Kana? #{fmt_jpn{|x|
|
797
|
-
puts "Kanji? #{fmt_jpn{|x|
|
803
|
+
puts "Hiragana? #{fmt_jpn{|x| Util.hiragana?(x)}}"
|
804
|
+
puts "Kana? #{fmt_jpn{|x| Util.kana?(x)}}"
|
805
|
+
puts "Kanji? #{fmt_jpn{|x| Util.kanji?(x)}}"
|
798
806
|
puts "Reduce: #{Util.reduce_jpn_space("' '")}"
|
799
807
|
puts
|
800
808
|
|
@@ -842,9 +850,25 @@ You can make some changes/fixes to the code and then install your local version:
|
|
842
850
|
|
843
851
|
`$ bundle exec rake install:local`
|
844
852
|
|
845
|
-
###
|
853
|
+
### Updating [^](#contents)
|
854
|
+
|
855
|
+
This will update *core/* for you:
|
856
|
+
|
857
|
+
`$ bundle exec rake update_core`
|
858
|
+
|
859
|
+
### Releasing [^](#contents)
|
846
860
|
|
847
|
-
|
861
|
+
1. Update *CHANGELOG.md*, *version.rb*, & *Gemfile.lock*
|
862
|
+
- *Raketary*: `$ raketary bump -v`
|
863
|
+
- Run: `$ bundle update`
|
864
|
+
2. Run: `$ bundle exec rake update_core`
|
865
|
+
3. Run: `$ bundle exec rake clobber pkg_core`
|
866
|
+
4. Create a new release & tag
|
867
|
+
- Add `pkg/nhkore-core.zip`
|
868
|
+
5. Run: `$ git pull`
|
869
|
+
6. Upload GitHub package
|
870
|
+
- *Raketary*: `$ raketary github_pkg`
|
871
|
+
7. Run: `$ bundle exec rake release`
|
848
872
|
|
849
873
|
## License [^](#contents)
|
850
874
|
|
data/Rakefile
CHANGED
@@ -63,6 +63,31 @@ Rake::TestTask.new() do |task|
|
|
63
63
|
task.warning = true
|
64
64
|
end
|
65
65
|
|
66
|
+
# If you need to run a part after the 1st part,
|
67
|
+
# just type 'n' to not overwrite the file and then 'y' for continue.
|
68
|
+
desc "Update '#{File.join(NHKore::Util::CORE_DIR,'')}' files for release"
|
69
|
+
task :update_core do |task|
|
70
|
+
require 'highline'
|
71
|
+
|
72
|
+
CONTINUE_MSG = "\nContinue (y/n)? "
|
73
|
+
|
74
|
+
cmd = ['ruby','-w','./lib/nhkore.rb','-t','300','-m','10']
|
75
|
+
hl = HighLine.new()
|
76
|
+
|
77
|
+
next unless sh(*cmd,'se','ez','bing')
|
78
|
+
next unless hl.agree(CONTINUE_MSG)
|
79
|
+
puts
|
80
|
+
|
81
|
+
next unless sh(*cmd,'news','-s','100','ez')
|
82
|
+
next unless hl.agree(CONTINUE_MSG)
|
83
|
+
puts
|
84
|
+
|
85
|
+
next unless sh(*cmd,'sift','-e','csv' ,'ez')
|
86
|
+
next unless sh(*cmd,'sift','-e','html','ez')
|
87
|
+
next unless sh(*cmd,'sift','-e','json','ez')
|
88
|
+
next unless sh(*cmd,'sift','-e','yml' ,'ez')
|
89
|
+
end
|
90
|
+
|
66
91
|
YARD::Rake::YardocTask.new() do |task|
|
67
92
|
task.files = [File.join('lib','**','*.rb')]
|
68
93
|
|
data/lib/nhkore/cli/news_cmd.rb
CHANGED
@@ -23,6 +23,7 @@
|
|
23
23
|
|
24
24
|
require 'time'
|
25
25
|
|
26
|
+
require 'nhkore/datetime_parser'
|
26
27
|
require 'nhkore/error'
|
27
28
|
require 'nhkore/missingno'
|
28
29
|
require 'nhkore/news'
|
@@ -57,7 +58,7 @@ module CLI
|
|
57
58
|
date time to use as a fallback in cases when an article doesn't have one;
|
58
59
|
format: YYYY-mm-dd H:M; example: 2020-03-30 15:30
|
59
60
|
EOD
|
60
|
-
value = Time.strptime(value,'%Y-%m-%d %H:%M',&
|
61
|
+
value = Time.strptime(value,'%Y-%m-%d %H:%M',&DatetimeParser.method(:guess_year))
|
61
62
|
value = Util.jst_time(value)
|
62
63
|
value
|
63
64
|
end
|
@@ -83,7 +83,7 @@ module CLI
|
|
83
83
|
key = key.to_s()
|
84
84
|
|
85
85
|
if key.include?('show')
|
86
|
-
raise CLIError
|
86
|
+
raise CLIError,"must specify a sub command for option[#{key}]"
|
87
87
|
end
|
88
88
|
end
|
89
89
|
|
@@ -283,7 +283,7 @@ module CLI
|
|
283
283
|
puts "> Easy: #{BingScraper.build_url(SearchScraper::YASASHII_SITE,count: count)}"
|
284
284
|
puts "> Regular: #{BingScraper.build_url(SearchScraper::FUTSUU_SITE,count: count)}"
|
285
285
|
else
|
286
|
-
raise CLIError
|
286
|
+
raise CLIError,'must specify a sub command for option[show-urls]'
|
287
287
|
end
|
288
288
|
|
289
289
|
return true
|
data/lib/nhkore/cli/sift_cmd.rb
CHANGED
@@ -24,6 +24,7 @@
|
|
24
24
|
require 'date'
|
25
25
|
require 'time'
|
26
26
|
|
27
|
+
require 'nhkore/datetime_parser'
|
27
28
|
require 'nhkore/news'
|
28
29
|
require 'nhkore/sifter'
|
29
30
|
require 'nhkore/util'
|
@@ -41,26 +42,6 @@ module CLI
|
|
41
42
|
DEFAULT_SIFT_YASASHII_FILE = "#{Sifter::DEFAULT_YASASHII_FILE}{search.criteria}{file.ext}"
|
42
43
|
SIFT_EXTS = [:csv,:htm,:html,:json,:yaml,:yml]
|
43
44
|
|
44
|
-
# Order matters.
|
45
|
-
SIFT_DATETIME_FMTS = [
|
46
|
-
'%Y-%m-%d %H:%M',
|
47
|
-
'%Y-%m-%d %H',
|
48
|
-
'%Y-%m-%d',
|
49
|
-
'%m-%d %H:%M',
|
50
|
-
'%Y-%m %H:%M',
|
51
|
-
'%m-%d %H',
|
52
|
-
'%Y-%m %H',
|
53
|
-
'%m-%d',
|
54
|
-
'%Y-%m',
|
55
|
-
'%d %H:%M',
|
56
|
-
'%y %H:%M',
|
57
|
-
'%d %H',
|
58
|
-
'%Y %H',
|
59
|
-
'%H:%M',
|
60
|
-
'%d',
|
61
|
-
'%Y'
|
62
|
-
]
|
63
|
-
|
64
45
|
attr_accessor :sift_datetime_text
|
65
46
|
attr_accessor :sift_search_criteria
|
66
47
|
|
@@ -90,7 +71,11 @@ module CLI
|
|
90
71
|
'9' (9th of Current Year & Month)
|
91
72
|
EOD
|
92
73
|
app.sift_datetime_text = value # Save the original value for the file name
|
93
|
-
|
74
|
+
|
75
|
+
value = DatetimeParser.parse_range(value)
|
76
|
+
|
77
|
+
app.check_empty_opt(:datetime,value) if value.nil?()
|
78
|
+
|
94
79
|
value
|
95
80
|
end
|
96
81
|
option :e,:ext,<<-EOD,argument: :required,default: DEFAULT_SIFT_EXT,transform: -> (value) do
|
@@ -211,96 +196,6 @@ module CLI
|
|
211
196
|
return filename
|
212
197
|
end
|
213
198
|
|
214
|
-
# TODO: This should probably be moved into its own class, into Util, or into Sifter?
|
215
|
-
def parse_sift_datetime(value)
|
216
|
-
value = Util.reduce_space(value).strip() # Don't use unspace_web_str(), want spaces for formats
|
217
|
-
value = value.split('...',2)
|
218
|
-
|
219
|
-
check_empty_opt(:datetime,nil) if value.empty?() # For ''
|
220
|
-
|
221
|
-
# Make a "to" and a "from" date time range.
|
222
|
-
value << value[0].dup() if value.length == 1
|
223
|
-
|
224
|
-
to_day = nil
|
225
|
-
to_hour = 23
|
226
|
-
to_minute = 59
|
227
|
-
to_month = 12
|
228
|
-
to_year = Util::MAX_SANE_YEAR
|
229
|
-
|
230
|
-
value.each_with_index() do |v,i|
|
231
|
-
v = check_empty_opt(:datetime,v) # For '...', '12-25...', or '...12-25'
|
232
|
-
|
233
|
-
has_day = false
|
234
|
-
has_hour = false
|
235
|
-
has_minute = false
|
236
|
-
has_month = false
|
237
|
-
has_year = false
|
238
|
-
|
239
|
-
SIFT_DATETIME_FMTS.each_with_index() do |fmt,i|
|
240
|
-
begin
|
241
|
-
# If don't do this, "%d" values will be parsed using "%d %H".
|
242
|
-
# It seems as though strptime() ignores space.
|
243
|
-
raise ArgumentError if !v.include?(' ') && fmt.include?(' ')
|
244
|
-
|
245
|
-
# If don't do this, "%y" values will be parsed using "%d".
|
246
|
-
raise ArgumentError if fmt == '%d' && v.length > 2
|
247
|
-
|
248
|
-
v = Time.strptime(v,fmt,&Util.method(:guess_year))
|
249
|
-
|
250
|
-
has_day = fmt.include?('%d')
|
251
|
-
has_hour = fmt.include?('%H')
|
252
|
-
has_minute = fmt.include?('%M')
|
253
|
-
has_month = fmt.include?('%m')
|
254
|
-
has_year = fmt.include?('%Y')
|
255
|
-
|
256
|
-
break # No problem; this format worked
|
257
|
-
rescue ArgumentError
|
258
|
-
# Out of formats.
|
259
|
-
raise if i >= (SIFT_DATETIME_FMTS.length - 1)
|
260
|
-
end
|
261
|
-
end
|
262
|
-
|
263
|
-
# "From" date time.
|
264
|
-
if i == 0
|
265
|
-
# Set these so that "2012-7-4...7-9" will use the appropriate year
|
266
|
-
# of "2012" for "7-9".
|
267
|
-
to_day = v.day if has_day
|
268
|
-
to_hour = v.hour if has_hour
|
269
|
-
to_minute = v.min if has_minute
|
270
|
-
to_month = v.month if has_month
|
271
|
-
to_year = v.year if has_year
|
272
|
-
|
273
|
-
v = Time.new(
|
274
|
-
has_year ? v.year : Util::MIN_SANE_YEAR,
|
275
|
-
has_month ? v.month : 1,
|
276
|
-
has_day ? v.day : 1,
|
277
|
-
has_hour ? v.hour : 0,
|
278
|
-
has_minute ? v.min : 0
|
279
|
-
)
|
280
|
-
# "To" date time.
|
281
|
-
else
|
282
|
-
to_hour = v.hour if has_hour
|
283
|
-
to_minute = v.min if has_minute
|
284
|
-
to_month = v.month if has_month
|
285
|
-
to_year = v.year if has_year
|
286
|
-
|
287
|
-
if has_day
|
288
|
-
to_day = v.day
|
289
|
-
# Nothing passed from the "from" date time?
|
290
|
-
elsif to_day.nil?()
|
291
|
-
# Last day of month.
|
292
|
-
to_day = Date.new(to_year,to_month,-1).day
|
293
|
-
end
|
294
|
-
|
295
|
-
v = Time.new(to_year,to_month,to_day,to_hour,to_minute)
|
296
|
-
end
|
297
|
-
|
298
|
-
value[i] = v
|
299
|
-
end
|
300
|
-
|
301
|
-
return value
|
302
|
-
end
|
303
|
-
|
304
199
|
def run_sift_cmd(type)
|
305
200
|
news_name = nil
|
306
201
|
|
@@ -0,0 +1,342 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
# encoding: UTF-8
|
3
|
+
# frozen_string_literal: true
|
4
|
+
|
5
|
+
#--
|
6
|
+
# This file is part of NHKore.
|
7
|
+
# Copyright (c) 2020 Jonathan Bradley Whited (@esotericpig)
|
8
|
+
#
|
9
|
+
# NHKore is free software: you can redistribute it and/or modify
|
10
|
+
# it under the terms of the GNU Lesser General Public License as published by
|
11
|
+
# the Free Software Foundation, either version 3 of the License, or
|
12
|
+
# (at your option) any later version.
|
13
|
+
#
|
14
|
+
# NHKore is distributed in the hope that it will be useful,
|
15
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
16
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
17
|
+
# GNU Lesser General Public License for more details.
|
18
|
+
#
|
19
|
+
# You should have received a copy of the GNU Lesser General Public License
|
20
|
+
# along with NHKore. If not, see <https://www.gnu.org/licenses/>.
|
21
|
+
#++
|
22
|
+
|
23
|
+
|
24
|
+
require 'attr_bool'
|
25
|
+
require 'date'
|
26
|
+
require 'time'
|
27
|
+
|
28
|
+
require 'nhkore/util'
|
29
|
+
|
30
|
+
|
31
|
+
module NHKore
|
32
|
+
###
|
33
|
+
# @author Jonathan Bradley Whited (@esotericpig)
|
34
|
+
# @since 0.3.4
|
35
|
+
###
|
36
|
+
class DatetimeParser
|
37
|
+
# Order matters!
|
38
|
+
FMTS = [
|
39
|
+
'%Y-%m-%d %H:%M',
|
40
|
+
'%Y-%m-%d %H',
|
41
|
+
'%Y-%m-%d',
|
42
|
+
'%m-%d %H:%M',
|
43
|
+
'%Y-%m %H:%M',
|
44
|
+
'%m-%d %H',
|
45
|
+
'%Y-%m %H',
|
46
|
+
'%m-%d',
|
47
|
+
'%Y-%m',
|
48
|
+
'%d %H:%M',
|
49
|
+
'%y %H:%M',
|
50
|
+
'%d %H',
|
51
|
+
'%Y %H',
|
52
|
+
'%H:%M',
|
53
|
+
'%d',
|
54
|
+
'%Y',
|
55
|
+
]
|
56
|
+
|
57
|
+
def self.guess_year(year)
|
58
|
+
if year < 1000
|
59
|
+
century = Util::JST_YEAR / 100 * 100 # 2120 -> 2100
|
60
|
+
millennium = Util::JST_YEAR / 1000 * 1000 # 2120 -> 2000
|
61
|
+
|
62
|
+
# If year <= 23 (2022 -> 23)...
|
63
|
+
if year <= ((Util::JST_YEAR % 100) + 1)
|
64
|
+
# Assume this century.
|
65
|
+
year = century + year
|
66
|
+
elsif year >= 100
|
67
|
+
# If (2000 + 150) <= 2201 (if current year is 2200)...
|
68
|
+
if (millennium + year) <= (Util::JST_YEAR + 1)
|
69
|
+
# Assume this millennium.
|
70
|
+
# So if the current year is 2200, and year is 150,
|
71
|
+
# then it will be 2000 + 150 = 2150.
|
72
|
+
year = millennium + year
|
73
|
+
else
|
74
|
+
# Assume previous millennium (2000 -> 1000),
|
75
|
+
# so year 999 will become 1999.
|
76
|
+
millennium -= 1000 if millennium >= 1000
|
77
|
+
year = millennium + year
|
78
|
+
end
|
79
|
+
else
|
80
|
+
# Assume previous century (2000 -> 1900).
|
81
|
+
century -= 100 if century >= 100
|
82
|
+
year = century + year
|
83
|
+
end
|
84
|
+
end
|
85
|
+
|
86
|
+
return year
|
87
|
+
end
|
88
|
+
|
89
|
+
def self.parse_range(value)
|
90
|
+
# Do not use unspace_web_str(), want spaces for formats.
|
91
|
+
value = Util.strip_web_str(Util.reduce_space(value))
|
92
|
+
values = value.split('...',2)
|
93
|
+
|
94
|
+
return nil if values.empty?() # For '' or '...'
|
95
|
+
|
96
|
+
# For '2020...' or '...2020'.
|
97
|
+
if value.include?('...')
|
98
|
+
# values.length is always 2 because of 2 in split() above.
|
99
|
+
|
100
|
+
# For '2020...'.
|
101
|
+
if Util.empty_web_str?(values[1])
|
102
|
+
values[1] = :infinity
|
103
|
+
# For '...2020'.
|
104
|
+
elsif Util.empty_web_str?(values[0])
|
105
|
+
values[0] = :infinity
|
106
|
+
end
|
107
|
+
end
|
108
|
+
|
109
|
+
datetimes = [
|
110
|
+
DatetimeParser.new(), # "From" date time
|
111
|
+
DatetimeParser.new(), # "To" date time
|
112
|
+
]
|
113
|
+
|
114
|
+
values.each_with_index() do |v,i|
|
115
|
+
dt = datetimes[i]
|
116
|
+
|
117
|
+
# Minimum/Maximum date time for '2020...' or '...2020'.
|
118
|
+
if v == :infinity
|
119
|
+
# "From" date time.
|
120
|
+
if i == 0
|
121
|
+
dt.min!()
|
122
|
+
# "To" date time.
|
123
|
+
else
|
124
|
+
dt.max!()
|
125
|
+
end
|
126
|
+
else
|
127
|
+
v = Util.strip_web_str(v)
|
128
|
+
|
129
|
+
FMTS.each_with_index() do |fmt,i|
|
130
|
+
begin
|
131
|
+
# If don't do this, "%d" values will be parsed using "%d %H".
|
132
|
+
# It seems as though strptime() ignores space.
|
133
|
+
raise ArgumentError if fmt.include?(' ') && !v.include?(' ')
|
134
|
+
|
135
|
+
# If don't do this, "%y..." values will be parsed using "%d...".
|
136
|
+
raise ArgumentError if fmt.start_with?('%d') && v.split(' ')[0].length > 2
|
137
|
+
|
138
|
+
dt.parse!(v,fmt)
|
139
|
+
|
140
|
+
break # No problem; this format worked
|
141
|
+
rescue ArgumentError
|
142
|
+
# Out of formats.
|
143
|
+
raise if i >= (FMTS.length - 1)
|
144
|
+
end
|
145
|
+
end
|
146
|
+
end
|
147
|
+
end
|
148
|
+
|
149
|
+
from = datetimes[0]
|
150
|
+
to = datetimes[1]
|
151
|
+
|
152
|
+
from.autofill!(:from,to)
|
153
|
+
to.autofill!(:to,from)
|
154
|
+
|
155
|
+
return [from.jst_time(),to.jst_time()]
|
156
|
+
end
|
157
|
+
|
158
|
+
attr_accessor :day
|
159
|
+
attr_accessor :hour
|
160
|
+
attr_accessor :min
|
161
|
+
attr_accessor :month
|
162
|
+
attr_accessor :sec
|
163
|
+
attr_accessor :year
|
164
|
+
|
165
|
+
attr_accessor? :has_day
|
166
|
+
attr_accessor? :has_hour
|
167
|
+
attr_accessor? :has_min
|
168
|
+
attr_accessor? :has_month
|
169
|
+
attr_accessor? :has_sec
|
170
|
+
attr_accessor? :has_year
|
171
|
+
|
172
|
+
attr_reader? :min_or_max
|
173
|
+
|
174
|
+
def initialize(year=nil,month=nil,day=nil,hour=nil,min=nil,sec=nil)
|
175
|
+
super()
|
176
|
+
|
177
|
+
set!(year,month,day,hour,min,sec)
|
178
|
+
|
179
|
+
self.has = false
|
180
|
+
@min_or_max = false
|
181
|
+
end
|
182
|
+
|
183
|
+
def autofill!(type,other)
|
184
|
+
case type
|
185
|
+
when :from
|
186
|
+
is_from = true
|
187
|
+
when :to
|
188
|
+
is_from = false
|
189
|
+
else
|
190
|
+
raise ArgumentError,"invalid type[#{type}]"
|
191
|
+
end
|
192
|
+
|
193
|
+
return self if @min_or_max
|
194
|
+
|
195
|
+
has_small = false
|
196
|
+
jst_now = Util.jst_now()
|
197
|
+
|
198
|
+
# Must be from smallest to biggest.
|
199
|
+
|
200
|
+
if @has_sec || other.has_sec?()
|
201
|
+
@sec = other.sec unless @has_sec
|
202
|
+
has_small = true
|
203
|
+
else
|
204
|
+
if has_small
|
205
|
+
@sec = jst_now.sec
|
206
|
+
else
|
207
|
+
@sec = is_from ? 0 : 59
|
208
|
+
end
|
209
|
+
end
|
210
|
+
|
211
|
+
if @has_min || other.has_min?()
|
212
|
+
@min = other.min unless @has_min
|
213
|
+
has_small = true
|
214
|
+
else
|
215
|
+
if has_small
|
216
|
+
@min = jst_now.min
|
217
|
+
else
|
218
|
+
@min = is_from ? 0 : 59
|
219
|
+
end
|
220
|
+
end
|
221
|
+
|
222
|
+
if @has_hour || other.has_hour?()
|
223
|
+
@hour = other.hour unless @has_hour
|
224
|
+
has_small = true
|
225
|
+
else
|
226
|
+
if has_small
|
227
|
+
@hour = jst_now.hour
|
228
|
+
else
|
229
|
+
@hour = is_from ? 0 : 23
|
230
|
+
end
|
231
|
+
end
|
232
|
+
|
233
|
+
if @has_day || other.has_day?()
|
234
|
+
@day = other.day unless @has_day
|
235
|
+
has_small = true
|
236
|
+
else
|
237
|
+
if has_small
|
238
|
+
@day = jst_now.day
|
239
|
+
else
|
240
|
+
@day = is_from ? 1 : :last_day
|
241
|
+
end
|
242
|
+
end
|
243
|
+
|
244
|
+
if @has_month || other.has_month?()
|
245
|
+
@month = other.month unless @has_month
|
246
|
+
has_small = true
|
247
|
+
else
|
248
|
+
if has_small
|
249
|
+
@month = jst_now.month
|
250
|
+
else
|
251
|
+
@month = is_from ? 1 : 12
|
252
|
+
end
|
253
|
+
end
|
254
|
+
|
255
|
+
if @has_year || other.has_year?()
|
256
|
+
@year = other.year unless @has_year
|
257
|
+
has_small = true
|
258
|
+
else
|
259
|
+
if has_small
|
260
|
+
@year = jst_now.year
|
261
|
+
else
|
262
|
+
@year = is_from ? Util::MIN_SANE_YEAR : jst_now.year
|
263
|
+
end
|
264
|
+
end
|
265
|
+
|
266
|
+
# Must be after setting @year & @month.
|
267
|
+
if @day == :last_day
|
268
|
+
@day = Date.new(@year,@month,-1).day
|
269
|
+
end
|
270
|
+
|
271
|
+
return self
|
272
|
+
end
|
273
|
+
|
274
|
+
def max!()
|
275
|
+
@min_or_max = true
|
276
|
+
|
277
|
+
# Ex: 2020-12-31 23:59:59
|
278
|
+
return set!(Util::JST_YEAR,12,31,23,59,59)
|
279
|
+
end
|
280
|
+
|
281
|
+
def min!()
|
282
|
+
@min_or_max = true
|
283
|
+
|
284
|
+
# Ex: 1924-01-01 00:00:00
|
285
|
+
return set!(Util::MIN_SANE_YEAR,1,1,0,0,0)
|
286
|
+
end
|
287
|
+
|
288
|
+
def parse!(value,fmt)
|
289
|
+
value = Time.strptime(value,fmt,&self.class.method(:guess_year))
|
290
|
+
|
291
|
+
@has_day = fmt.include?('%d')
|
292
|
+
@has_hour = fmt.include?('%H')
|
293
|
+
@has_min = fmt.include?('%M')
|
294
|
+
@has_month = fmt.include?('%m')
|
295
|
+
@has_sec = fmt.include?('%S')
|
296
|
+
@has_year = fmt.include?('%Y')
|
297
|
+
|
298
|
+
@day = value.day if @has_day
|
299
|
+
@hour = value.hour if @has_hour
|
300
|
+
@min = value.min if @has_min
|
301
|
+
@month = value.month if @has_month
|
302
|
+
@sec = value.sec if @has_sec
|
303
|
+
@year = value.year if @has_year
|
304
|
+
|
305
|
+
return self
|
306
|
+
end
|
307
|
+
|
308
|
+
def set!(year=nil,month=nil,day=nil,hour=nil,min=nil,sec=nil)
|
309
|
+
@year = year
|
310
|
+
@month = month
|
311
|
+
@day = day
|
312
|
+
@hour = hour
|
313
|
+
@min = min
|
314
|
+
@sec = sec
|
315
|
+
|
316
|
+
return self
|
317
|
+
end
|
318
|
+
|
319
|
+
def has=(value)
|
320
|
+
@has_day = value
|
321
|
+
@has_hour = value
|
322
|
+
@has_min = value
|
323
|
+
@has_month = value
|
324
|
+
@has_sec = value
|
325
|
+
@has_year = value
|
326
|
+
|
327
|
+
return self
|
328
|
+
end
|
329
|
+
|
330
|
+
def jst_time()
|
331
|
+
return Util.jst_time(time())
|
332
|
+
end
|
333
|
+
|
334
|
+
def time()
|
335
|
+
return Time.new(@year,@month,@day,@hour,@min,@sec)
|
336
|
+
end
|
337
|
+
|
338
|
+
def to_s()
|
339
|
+
return "#{@year}-#{@month}-#{@day} #{@hour}:#{@min}:#{@sec}"
|
340
|
+
end
|
341
|
+
end
|
342
|
+
end
|
data/lib/nhkore/lib.rb
CHANGED
data/lib/nhkore/util.rb
CHANGED
@@ -108,23 +108,6 @@ module NHKore
|
|
108
108
|
return !str.match?(/[\/\\]/)
|
109
109
|
end
|
110
110
|
|
111
|
-
def self.guess_year(year)
|
112
|
-
if year < 100
|
113
|
-
# 2021 -> 2000.
|
114
|
-
millennium = JST_YEAR / 100 * 100
|
115
|
-
|
116
|
-
# If year <= (2021 -> 21), assume this century.
|
117
|
-
if year <= (JST_YEAR % 100)
|
118
|
-
year = millennium + year
|
119
|
-
else
|
120
|
-
# Assume previous century (2000 -> 1900).
|
121
|
-
year = (millennium - 100) + year
|
122
|
-
end
|
123
|
-
end
|
124
|
-
|
125
|
-
return year
|
126
|
-
end
|
127
|
-
|
128
111
|
def self.hiragana?(str)
|
129
112
|
return HIRAGANA_REGEX =~ str
|
130
113
|
end
|
data/lib/nhkore/version.rb
CHANGED
data/nhkore.gemspec
CHANGED
@@ -67,7 +67,7 @@ Gem::Specification.new() do |spec|
|
|
67
67
|
spec.add_runtime_dependency 'http-cookie' ,'~> 1.0' # For parsing/setting cookies (BingScraper/Scraper)
|
68
68
|
spec.add_runtime_dependency 'japanese_deinflector' ,'~> 0.0' # For unconjugating Japanese words (plain/dictionary form)
|
69
69
|
spec.add_runtime_dependency 'nokogiri' ,'~> 1.10' # For scraping/hacking
|
70
|
-
spec.add_runtime_dependency 'psychgus' ,'~> 1.
|
70
|
+
spec.add_runtime_dependency 'psychgus' ,'~> 1.3' # For styling Psych YAML
|
71
71
|
spec.add_runtime_dependency 'public_suffix' ,'~> 4.0' # For parsing URL domain names
|
72
72
|
spec.add_runtime_dependency 'rainbow' ,'~> 3.0' # For CLI color output
|
73
73
|
spec.add_runtime_dependency 'rubyzip' ,'~> 2.3' # For extracting Zip files (GetCmd)
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: nhkore
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.3.
|
4
|
+
version: 0.3.4
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Jonathan Bradley Whited (@esotericpig)
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2020-04-
|
11
|
+
date: 2020-04-25 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: attr_bool
|
@@ -128,14 +128,14 @@ dependencies:
|
|
128
128
|
requirements:
|
129
129
|
- - "~>"
|
130
130
|
- !ruby/object:Gem::Version
|
131
|
-
version: '1.
|
131
|
+
version: '1.3'
|
132
132
|
type: :runtime
|
133
133
|
prerelease: false
|
134
134
|
version_requirements: !ruby/object:Gem::Requirement
|
135
135
|
requirements:
|
136
136
|
- - "~>"
|
137
137
|
- !ruby/object:Gem::Version
|
138
|
-
version: '1.
|
138
|
+
version: '1.3'
|
139
139
|
- !ruby/object:Gem::Dependency
|
140
140
|
name: public_suffix
|
141
141
|
requirement: !ruby/object:Gem::Requirement
|
@@ -357,6 +357,7 @@ files:
|
|
357
357
|
- lib/nhkore/cli/news_cmd.rb
|
358
358
|
- lib/nhkore/cli/search_cmd.rb
|
359
359
|
- lib/nhkore/cli/sift_cmd.rb
|
360
|
+
- lib/nhkore/datetime_parser.rb
|
360
361
|
- lib/nhkore/defn.rb
|
361
362
|
- lib/nhkore/dict.rb
|
362
363
|
- lib/nhkore/dict_scraper.rb
|
@@ -389,7 +390,7 @@ metadata:
|
|
389
390
|
changelog_uri: https://github.com/esotericpig/nhkore/blob/master/CHANGELOG.md
|
390
391
|
homepage_uri: https://github.com/esotericpig/nhkore
|
391
392
|
source_code_uri: https://github.com/esotericpig/nhkore
|
392
|
-
post_install_message: " \n NHKore v0.3.
|
393
|
+
post_install_message: " \n NHKore v0.3.4\n \n You can now use [nhkore] on the
|
393
394
|
command line.\n \n Homepage: https://github.com/esotericpig/nhkore\n \n Code:
|
394
395
|
\ https://github.com/esotericpig/nhkore\n Changelog: https://github.com/esotericpig/nhkore/blob/master/CHANGELOG.md\n
|
395
396
|
\ Bugs: https://github.com/esotericpig/nhkore/issues\n \n"
|