dwc-archive 0.4.6 → 0.4.7
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.
data/VERSION
CHANGED
@@ -1 +1 @@
|
|
1
|
-
0.4.
|
1
|
+
0.4.7
|
@@ -60,3 +60,4 @@ Feature: Reading of a Darwing Core Archive
|
|
60
60
|
When I create a new DarwinCore instance
|
61
61
|
Then I am able to use DarwinCore#normalize_classification method
|
62
62
|
And get normalized classification in expected format
|
63
|
+
And there are paths, synonyms and vernacular names in normalized classification
|
@@ -170,3 +170,24 @@ Then /^get normalized classification in expected format$/ do
|
|
170
170
|
key = @normalized_classification.keys[0]
|
171
171
|
@normalized_classification[key].class.should == DarwinCore::TaxonNormalized
|
172
172
|
end
|
173
|
+
|
174
|
+
Then /^there are paths, synonyms and vernacular names in normalized classification$/ do
|
175
|
+
@paths_are_generated = false
|
176
|
+
@synonyms_are_generated = false
|
177
|
+
@vernaculars_are_generated = false
|
178
|
+
@normalized_classification.each do |k, v|
|
179
|
+
if v.classification_path.size > 0
|
180
|
+
@paths_are_generated = true
|
181
|
+
end
|
182
|
+
if v.synonyms.size > 0
|
183
|
+
@synonyms_are_generated = true
|
184
|
+
end
|
185
|
+
if v.vernacular_names.size > 0
|
186
|
+
@vernaculars_are_generated = true
|
187
|
+
end
|
188
|
+
break if (@vernaculars_are_generated && @paths_are_generated && @synonyms_are_generated)
|
189
|
+
end
|
190
|
+
@paths_are_generated.should be_true
|
191
|
+
@vernaculars_are_generated.should be_true
|
192
|
+
@synonyms_are_generated.should be_true
|
193
|
+
end
|
@@ -37,9 +37,9 @@ class DarwinCore
|
|
37
37
|
|
38
38
|
def normalize
|
39
39
|
@res = {}
|
40
|
-
|
40
|
+
ingest_core
|
41
41
|
calculate_classification_path
|
42
|
-
|
42
|
+
ingest_extensions
|
43
43
|
@res
|
44
44
|
end
|
45
45
|
|
@@ -76,14 +76,14 @@ class DarwinCore
|
|
76
76
|
@core[:taxonomicstatus] ? row[@core[:taxonomicstatus]] : nil)
|
77
77
|
end
|
78
78
|
|
79
|
-
def
|
79
|
+
def ingest_core
|
80
80
|
raise RuntimeError, "Darwin Core core fields must contain taxon id and scientific name" unless (@core[:id] && @core[:scientificname])
|
81
81
|
puts "Reading core information" if @verbose
|
82
82
|
rows = @dwc.core.read[0]
|
83
|
-
puts "
|
83
|
+
puts "Ingesting information from the core" if @verbose
|
84
84
|
rows.each_with_index do |r, i|
|
85
85
|
count = i + 1
|
86
|
-
puts "
|
86
|
+
puts "Ingesting %s'th record" % count if @verbose and count % @verbose_count == 0
|
87
87
|
#core has AcceptedNameUsageId
|
88
88
|
if @core[:acceptednameusageid] && r[@core[:acceptednameusageid]] && r[@core[:acceptednameusageid]] != r[@core[:id]]
|
89
89
|
add_synonym_from_core(@core[:acceptednameusageid], r)
|
@@ -107,7 +107,7 @@ class DarwinCore
|
|
107
107
|
|
108
108
|
def calculate_classification_path
|
109
109
|
@res.each do |taxon_id, taxon|
|
110
|
-
next if taxon.classification_path
|
110
|
+
next if !taxon.classification_path.empty?
|
111
111
|
begin
|
112
112
|
get_classification_path(taxon)
|
113
113
|
rescue DarwinCore::ParentNotCurrentError
|
@@ -117,7 +117,7 @@ class DarwinCore
|
|
117
117
|
end
|
118
118
|
|
119
119
|
def get_classification_path(taxon)
|
120
|
-
return if taxon.classification_path
|
120
|
+
return if !taxon.classification_path.empty?
|
121
121
|
if DarwinCore.nil_field?(taxon.parent_id)
|
122
122
|
taxon.classification_path << taxon.current_name_canonical
|
123
123
|
else
|
@@ -129,28 +129,28 @@ class DarwinCore
|
|
129
129
|
raise DarwinCore::ParentNotCurrentError, error
|
130
130
|
end
|
131
131
|
if parent_cp
|
132
|
-
taxon.classification_path
|
132
|
+
taxon.classification_path << parent_cp + [taxon.current_name_canonical]
|
133
133
|
else
|
134
134
|
get_classification_path(@res[taxon.parent_id])
|
135
|
-
taxon.classification_path
|
135
|
+
taxon.classification_path << @res[taxon.parent_id].classification_path + [taxon.current_name_canonical]
|
136
136
|
end
|
137
137
|
end
|
138
138
|
end
|
139
139
|
|
140
|
-
def
|
140
|
+
def ingest_extensions
|
141
141
|
@extensions.each do |e|
|
142
142
|
ext, fields = *e
|
143
|
-
|
144
|
-
|
143
|
+
ingest_synonyms(e) if fields.keys.include? :scientificname
|
144
|
+
ingest_vernaculars(e) if fields.keys.include? :vernacularname
|
145
145
|
end
|
146
146
|
end
|
147
147
|
|
148
|
-
def
|
149
|
-
puts "
|
148
|
+
def ingest_synonyms(extension)
|
149
|
+
puts "Ingesting synonyms extension" if @verbose
|
150
150
|
ext, fields = *extension
|
151
151
|
ext.read[0].each_with_index do |r, i|
|
152
152
|
count = i + 1
|
153
|
-
puts "
|
153
|
+
puts "Ingesting %s'th record" % count if @verbose && count % @verbose_count == 0
|
154
154
|
@res[r[fields[:id]]].synonyms << SynonymNormalized.new(
|
155
155
|
r[fields[:scientificname]],
|
156
156
|
canonical_name(r[fields[:scientificname]]),
|
@@ -158,12 +158,12 @@ class DarwinCore
|
|
158
158
|
end
|
159
159
|
end
|
160
160
|
|
161
|
-
def
|
162
|
-
puts "
|
161
|
+
def ingest_vernaculars(extension)
|
162
|
+
puts "Ingesting vernacular names" if @verbose
|
163
163
|
ext, fields = *extension
|
164
164
|
ext.read[0].each_with_index do |r, i|
|
165
165
|
count = i + 1
|
166
|
-
puts "
|
166
|
+
puts "Ingesting %s'th record" % count if @verbose && count % @verbose_count == 0
|
167
167
|
@res[r[fields[:id]]].vernacular_names << VernacularNormalized.new(
|
168
168
|
r[fields[:vernacularname]],
|
169
169
|
fields[:languagecode] ? r[fields[:languagecode]] : nil)
|
metadata
CHANGED
@@ -1,13 +1,13 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: dwc-archive
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
hash:
|
4
|
+
hash: 1
|
5
5
|
prerelease: false
|
6
6
|
segments:
|
7
7
|
- 0
|
8
8
|
- 4
|
9
|
-
-
|
10
|
-
version: 0.4.
|
9
|
+
- 7
|
10
|
+
version: 0.4.7
|
11
11
|
platform: ruby
|
12
12
|
authors:
|
13
13
|
- Dmitry Mozzherin
|
@@ -15,7 +15,7 @@ autorequire:
|
|
15
15
|
bindir: bin
|
16
16
|
cert_chain: []
|
17
17
|
|
18
|
-
date: 2010-09-
|
18
|
+
date: 2010-09-18 00:00:00 -04:00
|
19
19
|
default_executable:
|
20
20
|
dependencies:
|
21
21
|
- !ruby/object:Gem::Dependency
|