rbbt-sources 3.1.14 → 3.1.15
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/etc/allowed_biomart_archives +1 -0
- data/lib/rbbt/sources/pubmed.rb +3 -2
- data/lib/rbbt/sources/tfacts.rb +75 -0
- metadata +11 -11
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 968b6c9ab9794fa92209ba2c8f36537bfc942711
|
4
|
+
data.tar.gz: 6a5501945048ef2a52872d54e35a85d8637b25a0
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 8ded417d4f6a04e6ff5ba21da1e454d54a0cf73aab95783a1a60ada25ea39faad2f8d4edcae462ae44f03de0ff488ea99737dc235d4472493d309342debe3391
|
7
|
+
data.tar.gz: 91b3fbe0852496b15cae6b512f150e5cf86c1742d57d41104b2247df3e75661dc5dce5aa4bae6e3eb86a3bd427addb842f97617a32a6f977a31fc0385daded3a
|
data/lib/rbbt/sources/pubmed.rb
CHANGED
@@ -227,14 +227,15 @@ module PubMed
|
|
227
227
|
result_files = FileCache.cache_online_elements(pmids, 'pubmed-{ID}.xml') do |ids|
|
228
228
|
result = {}
|
229
229
|
values = []
|
230
|
-
Misc.divide(ids, (ids.length /
|
230
|
+
Misc.divide(ids, (ids.length / 20) + 1).each do |list|
|
231
231
|
begin
|
232
232
|
Misc.try3times do
|
233
233
|
url = "http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi"
|
234
234
|
|
235
235
|
postdata = "db=pubmed&retmode=xml&id=#{list* ","}"
|
236
236
|
xml = TmpFile.with_file(postdata) do |postfile|
|
237
|
-
Open.read(url, :quiet => true, :nocache => true, :nice => @@pubmed_lag, :nice_key => "PubMed", "--post-file=" => postfile)
|
237
|
+
#Open.read(url, :quiet => true, :nocache => true, :nice => @@pubmed_lag, :nice_key => "PubMed", "--post-file=" => postfile)
|
238
|
+
Open.read(url+'?'+postdata, :quiet => true, :nocache => true, :nice => @@pubmed_lag, :nice_key => "PubMed", "--__post-file=" => postfile)
|
238
239
|
end
|
239
240
|
|
240
241
|
values += xml.scan(/(<PubmedArticle>.*?<\/PubmedArticle>)/smu).flatten
|
data/lib/rbbt/sources/tfacts.rb
CHANGED
@@ -47,6 +47,81 @@ module TFacts
|
|
47
47
|
TFacts.targets.tsv.reorder("Transcription Factor (Associated Gene Name)").to_s
|
48
48
|
end
|
49
49
|
|
50
|
+
TFacts.claim TFacts.tf_tg, :proc do
|
51
|
+
require 'spreadsheet'
|
52
|
+
book = Spreadsheet.open TFacts[".source"]["Catalogues.xls"].produce.find
|
53
|
+
|
54
|
+
tsv = TSV.setup({}, :key_field => "Transcription Factor (Associated Gene Name)", :fields => ["Target Gene (Associated Gene Name)", "Sign", "Species", "Source", "PMID"], :namespace => "Hsa", :type => :double)
|
55
|
+
|
56
|
+
sheet = book.worksheet 1
|
57
|
+
sheet.each do |row|
|
58
|
+
target, tf, sign, source, species, pmids = row.values_at(0, 1, 2, 3, 4, 5).collect{|v| v.to_s.gsub(/\s/, ' ') }
|
59
|
+
pmids = "" if pmids.nil? or pmids == "0"
|
60
|
+
next if tf =~ /OFFICIAL_/
|
61
|
+
|
62
|
+
species = species.split(";").compact.reject{|s| s.empty?}.collect{|s|
|
63
|
+
case s
|
64
|
+
when /Homo sap/i
|
65
|
+
"Homo sapiens"
|
66
|
+
when /muscul/i
|
67
|
+
"Mus musculus"
|
68
|
+
else
|
69
|
+
s
|
70
|
+
end
|
71
|
+
}*";"
|
72
|
+
source = source.split(";").compact.reject{|s| s.empty?}*";"
|
73
|
+
pmids = pmids.split(";").compact.reject{|s| s.empty?}.collect{|s| s.sub(/\s*pmid:\s*/,'').strip}*";"
|
74
|
+
|
75
|
+
values = [target, sign, species, source, pmids]
|
76
|
+
values = values.collect{|v| [v]}
|
77
|
+
tsv.zip_new(tf, values)
|
78
|
+
end
|
79
|
+
|
80
|
+
sheet = book.worksheet 0
|
81
|
+
sheet.each do |row|
|
82
|
+
target, tf, source, species, pmids = row.values_at(0, 1, 2, 3, 4, 5).collect{|v| v.to_s.gsub(/\s/, ' ') }
|
83
|
+
next if tf =~ /OFFICIAL_/
|
84
|
+
|
85
|
+
species = species.split(";").compact.reject{|s| s.empty?}.collect{|s|
|
86
|
+
case s
|
87
|
+
when /Homo sap/i
|
88
|
+
"Homo sapiens"
|
89
|
+
when /muscul/i
|
90
|
+
"Mus musculus"
|
91
|
+
else
|
92
|
+
s
|
93
|
+
end
|
94
|
+
}*";"
|
95
|
+
source = source.split(";").compact.reject{|s| s.empty?}*";"
|
96
|
+
pmids = "" if pmids.nil? or pmids == "0"
|
97
|
+
pmids = pmids.split(";").compact.reject{|s| s.empty?}.collect{|s| s.sub(/\s*pmid:\s*/,'').strip}*";"
|
98
|
+
|
99
|
+
current = tsv[tf]
|
100
|
+
if current and current[0].include? target
|
101
|
+
new = []
|
102
|
+
Misc.zip_fields(current).each do |ntarget,nsign,nspecies,nsource,npmids|
|
103
|
+
if target == ntarget
|
104
|
+
if species != nspecies
|
105
|
+
nspecies = (nspecies.split(";") + species.split(";")).uniq * ";"
|
106
|
+
end
|
107
|
+
npmids = (npmids.split(";") + pmids.split(";")).uniq * ";"
|
108
|
+
end
|
109
|
+
new << [ntarget, nsign, nspecies, nsource, npmids]
|
110
|
+
end
|
111
|
+
values = Misc.zip_fields(new)
|
112
|
+
tsv[tf] = values
|
113
|
+
else
|
114
|
+
pmids = "" if pmids.nil?
|
115
|
+
|
116
|
+
values = [target, "", species, source, pmids]
|
117
|
+
tsv.zip_new(tf, values)
|
118
|
+
end
|
119
|
+
|
120
|
+
end
|
121
|
+
|
122
|
+
tsv.to_s
|
123
|
+
end
|
124
|
+
|
50
125
|
end
|
51
126
|
|
52
127
|
if defined? Entity and defined? Gene and Entity === Gene
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: rbbt-sources
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 3.1.
|
4
|
+
version: 3.1.15
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Miguel Vazquez
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2017-
|
11
|
+
date: 2017-04-18 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: rbbt-util
|
@@ -181,19 +181,19 @@ signing_key:
|
|
181
181
|
specification_version: 4
|
182
182
|
summary: Data sources for the Ruby Bioinformatics Toolkit (rbbt)
|
183
183
|
test_files:
|
184
|
-
- test/test_helper.rb
|
185
184
|
- test/rbbt/sources/test_pubmed.rb
|
186
|
-
- test/rbbt/sources/
|
187
|
-
- test/rbbt/sources/test_gscholar.rb
|
185
|
+
- test/rbbt/sources/test_pharmagkb.rb
|
188
186
|
- test/rbbt/sources/test_biomart.rb
|
189
|
-
- test/rbbt/sources/
|
187
|
+
- test/rbbt/sources/test_gscholar.rb
|
188
|
+
- test/rbbt/sources/test_kegg.rb
|
190
189
|
- test/rbbt/sources/test_pina.rb
|
191
190
|
- test/rbbt/sources/test_entrez.rb
|
192
|
-
- test/rbbt/sources/test_pharmagkb.rb
|
193
|
-
- test/rbbt/sources/test_organism.rb
|
194
191
|
- test/rbbt/sources/test_matador.rb
|
195
|
-
- test/rbbt/sources/
|
196
|
-
- test/rbbt/sources/test_kegg.rb
|
192
|
+
- test/rbbt/sources/test_HPRD.rb
|
197
193
|
- test/rbbt/sources/test_synapse.rb
|
194
|
+
- test/rbbt/sources/test_go.rb
|
198
195
|
- test/rbbt/sources/test_stitch.rb
|
199
|
-
- test/rbbt/sources/
|
196
|
+
- test/rbbt/sources/test_organism.rb
|
197
|
+
- test/rbbt/sources/test_string.rb
|
198
|
+
- test/rbbt/sources/test_tfacts.rb
|
199
|
+
- test/test_helper.rb
|