data and scripts

2 files changed, 53 insertions, 0 deletions

diff --git a/scripts/geoquery/extract.rb b/scripts/geoquery/extract.rb
new file mode 100755
index 0000000..ea6e864
--- /dev/null
+++ b/scripts/geoquery/extract.rb
@@ -0,0 +1,33 @@
+#!/usr/bin/env ruby
+
+require 'nlp_ruby'
+require 'xmlsimple'
+
+
+def extract fn='./corpus.xml', lang='en', ids
+  doc = XmlSimple.xml_in(fn)
+  doc['example'].each { |example|
+    next if (!ids.include? example['id']) && ids.size>0
+    if lang == 'funql' || lang == 'geo-prolog'
+      puts example['mrl'][0]['content'].to_s.strip
+    else
+      example['nl'].each { |nl|
+        if nl['lang'] == lang
+          puts nl['content']
+        else
+          next
+        end
+      }
+    end
+  }
+end
+
+def main
+  ids = []
+  ids = ReadFile.readlines_strip ARGV[2]
+  extract ARGV[0], ARGV[1], ids
+end
+
+
+main
+
diff --git a/scripts/geoquery/select.rb b/scripts/geoquery/select.rb
new file mode 100755
index 0000000..fbf8d14
--- /dev/null
+++ b/scripts/geoquery/select.rb
@@ -0,0 +1,20 @@
+#!/usr/bin/env ruby
+
+require 'nlp_ruby'
+
+
+def main
+  ids = []
+  ids = ReadFile.readlines_strip(ARGV[0]).map{ |i| i.strip.to_i } if ARGV[0]
+  delete_ids = []
+  delete_ids = ReadFile.readlines_strip(ARGV[1]).map{ |i| i.strip.to_i } if ARGV[1]
+  i = 0
+  while line = STDIN.gets
+    puts line if ids.include?(i)&&!delete_ids.include?(i)
+    i += 1
+  end
+end
+
+
+main
+