text_alignment 0.2.3 → 0.2.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 1ff237d2f6a69a7328f334b39dd355e2f35c45552219d874d4630957d80f3516
-  data.tar.gz: 133f403758f32ce67fa340cedab21bd02945525a9a004519e245340f9cd2f9fb
+  metadata.gz: c61a711930d19f5a72bd4a4128f5f36038a185eeb203b9ca68afce143694dbd3
+  data.tar.gz: 173fa1ed0277f0384f0804e7a1ea02c3bd1a1eda20a5d3f57523028f815e69ac
 SHA512:
-  metadata.gz: 050dc4978efd7e63a9862f4a8f5ac58e439e30c737d5a07bd6f5cd6dd9157be36324c12cec0957f4fe8b700265c54147701dd7bfd1d0107ecef84b31a20a95e7
-  data.tar.gz: e542e4a90a02e9d8063355edb47fba2284ffc29e02090c9adce691012d1130b7782f624665b2f7bf772140983e548a130476a7d604a2166f0b835a48f245051b
+  metadata.gz: 8f2a018145a07ec1d9a5c2277fd4d1e4bdb694726cf0936c92e8f25addb325e485545b7a1373a265e3b60fe2da0e39cdccbad586968aef093ed2154070242173
+  data.tar.gz: 488390234e0e3b9d7d67389f303aed6e59bb0e16b1ad94908c8020be64b9c0bac9a9f404e42944fd37ece188377c930fd43596678a80ee24f3b00174c9e33aa4

data/bin/align_annotations

@@ -0,0 +1,31 @@
+#!/usr/bin/env ruby
+require 'text_alignment'
+
+require 'json'
+require 'pp'
+
+unless ARGV.length == 2
+	warn "align_annotations target_annotations(.json) reference_annotations(.json)"
+	exit
+end
+
+anns1 = JSON.parse File.read(ARGV[0].strip), :symbolize_names => true
+anns2 = JSON.parse File.read(ARGV[1].strip), :symbolize_names => true
+
+str1 = anns1[:text]
+str2 = anns2[:text]
+
+denotations = anns1[:denotations]
+
+align = TextAlignment::TextAlignment.new(str1, str2, TextAlignment::MAPPINGS)
+puts TextAlignment::sdiff2cdiff(align.sdiff)
+puts "\n=====\n\n"
+
+pp denotations
+puts "-----"
+new_denotations = align.transform_hdenotations(denotations)
+pp new_denotations
+
+# p align.common_elements
+# puts "---------------"
+# p align.mapped_elements

data/lib/text_alignment/text_alignment.rb

@@ -27,6 +27,7 @@ class TextAlignment::TextAlignment
     ## preprocessing
     str1 = str1.dup
     str2 = str2.dup
+    mappings = mappings.dup
 
     ## find the first nomatch character
     TextAlignment::NOMATCH_CHARS.each_char do |c|
@@ -181,35 +182,3 @@ class TextAlignment::TextAlignment
     @position_map_end = posmap_end.sort.to_h
   end
 end
-
-if __FILE__ == $0
-  require 'json'
-  require 'pp'
-
-  str1 = "TI  - Identification of a region which directs the monocytic activity of the\n      colony-stimulating factor 1 (macrophage colony-stimulating factor) receptor\n      promoter and binds PEBP2/CBF (AML1)."
-  str2 = "Identification of a region which directs the monocytic activity of the colony-stimulating factor 1 (macrophage colony-stimulating factor) receptor promoter and binds PEBP2/CBF (AML1).\nThe receptor for the macrophage colony-stimulating factor (or colony-stimulating factor 1 [CSF-1]) is expressed from different promoters in monocytic cells and placental trophoblasts. We have demonstrated that the monocyte-specific expression of the CSF-1 receptor is regulated at the level of transcription by a tissue-specific promoter whose activity is stimulated by the monocyte/B-cell-specific transcription factor PU.1 (D.-E. Zhang, C.J. Hetherington, H.-M. Chen, and D.G. Tenen, Mol. Cell. Biol. 14:373-381, 1994). Here we report that the tissue specificity of this promoter is also mediated by sequences in a region II (bp -88 to -59), which lies 10 bp upstream from the PU.1-binding site. When analyzed by DNase footprinting, region II was protected preferentially in monocytic cells. Electrophoretic mobility shift assays confirmed that region II interacts specifically with nuclear proteins from monocytic cells. Two gel shift complexes (Mono A and Mono B) were formed with separate sequence elements within this region. Competition and supershift experiments indicate that Mono B contains a member of the polyomavirus enhancer-binding protein 2/core-binding factor (PEBP2/CBF) family, which includes the AML1 gene product, while Mono A is a distinct complex preferentially expressed in monocytic cells. Promoter constructs with mutations in these sequence elements were no longer expressed specifically in monocytes. Furthermore, multimerized region II sequence elements enhanced the activity of a heterologous thymidine kinase promoter in monocytic cells but not other cell types tested. These results indicate that the monocyte/B-cell-specific transcription factor PU.1 and the Mono A and Mono B protein complexes act in concert to regulate monocyte-specific transcription of the CSF-1 receptor."
-
-
-  if ARGV.length == 2
-    anns1 = JSON.parse File.read(ARGV[0].strip), :symbolize_names => true
-    anns2 = JSON.parse File.read(ARGV[1].strip), :symbolize_names => true
-  end
-
-  str1 = anns1[:text]
-  str2 = anns2[:text]
-
-  denotations = anns1[:denotations]
-
-  align = TextAlignment::TextAlignment.new(str1, str2, TextAlignment::MAPPINGS)
-  puts TextAlignment::sdiff2cdiff(align.sdiff)
-  puts "\n=====\n\n"
-
-  pp denotations
-  puts "-----"
-  new_denotations = align.transform_hdenotations(denotations)
-  pp new_denotations
-
-  # p align.common_elements
-  # puts "---------------"
-  # p align.mapped_elements
-end

data/lib/text_alignment/version.rb

@@ -1,3 +1,3 @@
 class TextAlignment
-  VERSION = '0.2.3'
+  VERSION = '0.2.4'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: text_alignment
 version: !ruby/object:Gem::Version
-  version: 0.2.3
+  version: 0.2.4
 platform: ruby
 authors:
 - Jin-Dong Kim
@@ -49,7 +49,8 @@ description: |-
                            of two character strings and annotations made to them.
 email:
 - jdkim@dbcls.rois.ac.jp
-executables: []
+executables:
+- align_annotations
 extensions: []
 extra_rdoc_files: []
 files:
@@ -58,6 +59,7 @@ files:
 - Gemfile.lock
 - LICENSE.txt
 - README.md
+- bin/align_annotations
 - lib/text_alignment.rb
 - lib/text_alignment/approximate_fit.rb
 - lib/text_alignment/find_divisions.rb