StrIdx 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 1079ebe4265ed15d2307d260395e33471ab149bc6cd33ca6e469ff93c960dba8
-  data.tar.gz: 0bdba2496690ad9e1794d653bb5ca047a15fcfee6d2988ff06098c69920ae269
+  metadata.gz: c3e062e9a4d99367ff312ab222e861362e05381a034f513d1581cc2464df53df
+  data.tar.gz: 6088bf73c44fa3487757f2a0d3a60fcc6b00ee0514216a76278f027d83cc7807
 SHA512:
-  metadata.gz: 8a1b2fe9ce6f87a8585f5b9ca86092f21c600575b704fc3bf4ebaa8ae3bd0897f7e272fb580fb98c38e8b65a5b95d0fd8cc1c3baeafdb235fe1d48e51153ae6c
-  data.tar.gz: b4a62b6251a38637905653f2aaa919653465fbc453844a2f02466ce413a8a1de0946de09973535472a70ccbc17151d426e76fe92b3338ccc0af9f83f6821b201
+  metadata.gz: ece52c12cef6f1460e995e296fbaed2911e45a15d46cbe1fd1b0ff21f8ae546198078a83951cb9430bf82ff81d866a22e0244c7472f566efa2376285b8e805b2
+  data.tar.gz: 8b6751d914e41ffbe9c69274054ae398679cabdd159f0ab1c45ca5dc899a9a34c626f782ca4518c68b7d599efe71ea61489e277061b42284bc22345d793b116e

data/README.md

@@ -1,4 +1,4 @@
-# stridx
+# StrIdx
 This library provides fast fuzzy string similarity search and indexing. It has been mainly developed for indexing filepaths, but can be used for other types of strings aswell. It can easily handle fuzzy searches for more than 100,000 filepaths.
 
 The fuzziness means that candidate filepaths do not need to include exact match of the query string. They are considered a good match if they include parts of the query string, and even if those parts are in the wrong order.
@@ -26,7 +26,7 @@ Sum up the scores for each character c and divide by (string length)^2
 For example, if query = "rngnomadriv" 
 and candidate is "./drivers/char/hw_random/nomadik-rng.c", then scores are calculated as follows:
 ```
-    rngnomadriv
+    rngnomadriv (substrings rng=3, nomad=5 and driv=4)
     33355555444 (subscores)
     FFFFFFFFDDD (F=file component, D=dir component)
     score1=(3+3+3+5+5+5+5+5+(4+4+4)*0.7)
@@ -36,6 +36,75 @@ and candidate is "./drivers/char/hw_random/nomadik-rng.c", then scores are calcu
     and minor part by (query string length)*(candidate string length)
     score = score1/(11*11)*0.97 + score1/(11*38)*0.03 = 0.342944
 ```
+
+# Ruby interface
+Install:
+```
+gem install StrIdx
+```
+
+Usage example (see test.rb):
+```ruby
+require "stridx"
+idx = StrIdx::StringIndex.new
+
+t = Time.new
+fn = File.expand_path("flist.txt")
+lines = IO.read(fn).lines.collect { |x| x.strip }
+i = 0
+for x in lines
+  idx.add(x, i)
+  i += 1
+end
+
+idx_time = Time.new 
+puts "\nIndexing time (#{lines.size} files}): #{(idx_time - t).round(4)} seconds"
+
+query = "rngnomadriv"
+res = idx.find(query)
+puts "query: #{query}"
+puts "\nResults:"
+puts "Filename, score"
+puts "==============="
+for id, score in res
+  fn = lines[id]
+  puts "#{fn}, #{score.round(4)}"
+end
+
+query_time = Time.new 
+
+puts "\nSearch time: #{(query_time - idx_time).round(4)} seconds"
+
+```
+
+Output:
+```
+Indexing time (89828 files}): 2.813722207
+query: rngnomadriv
+
+Results:
+Filename, score
+===============
+./drivers/char/hw_random/nomadik-rng.c, 0.3429
+./drivers/pinctrl/nomadik, 0.2714
+./drivers/clk/clk-nomadik.c, 0.2711
+./drivers/gpio/gpio-nomadik.c, 0.2709
+./drivers/i2c/busses/i2c-nomadik.c, 0.2704
+./drivers/clocksource/nomadik-mtu.c, 0.2704
+./drivers/gpu/drm/pl111/pl111_nomadik.h, 0.2701
+./drivers/gpu/drm/pl111/pl111_nomadik.c, 0.2701
+./drivers/pinctrl/nomadik/pinctrl-nomadik.c, 0.2699
+./drivers/input/keyboard/nomadik-ske-keypad.c, 0.2698
+./drivers/pinctrl/nomadik/pinctrl-nomadik-db8500.c, 0.2696
+./drivers/pinctrl/nomadik/pinctrl-nomadik-stn8815.c, 0.2695
+./drivers/char/hw_random/omap-rng.c, 0.2364
+./drivers/char/hw_random/omap3-rom-rng.c, 0.2361
+./include/dt-bindings/pinctrl/nomadik.h, 0.2248
+
+Search time: 0.0488 seconds
+```
+
+
 # C++ API
 See demo.cpp
 ```cpp

data/rubyext/ruby_interf.cpp

@@ -41,7 +41,7 @@ VALUE StringIndexAddSegments(VALUE self, VALUE str, VALUE fileId) {
   return self;
 }
 
-VALUE StringIndexFind(VALUE self, VALUE str, VALUE minChars) {
+VALUE StringIndexFind(VALUE self, VALUE str) {
   VALUE ret;
   std::string s1 = StringValueCStr(str);
 
@@ -50,7 +50,7 @@ VALUE StringIndexFind(VALUE self, VALUE str, VALUE minChars) {
   StringIndex *idx = (StringIndex *)data;
 
   ret = rb_ary_new();
-  const std::vector<std::pair<float, int>> &results = idx->findSimilar(s1, NUM2INT(minChars));
+  const std::vector<std::pair<float, int>> &results = idx->findSimilar(s1, 2);
   int limit = 15;
   int i = 0;
   for (const auto &res : results) {
@@ -68,12 +68,12 @@ VALUE StringIndexFind(VALUE self, VALUE str, VALUE minChars) {
 
 void Init_stridx(void) {
 
-  VALUE cFoo = rb_define_class("CppStringIndex", rb_cObject);
+  VALUE mStrIdx = rb_define_module("StrIdx");
+  VALUE classStringIndex = rb_define_class_under(mStrIdx, "StringIndex", rb_cObject);
 
-  rb_define_alloc_func(cFoo, str_idx_alloc);
-  rb_define_method(cFoo, "add", StringIndexAddSegments, 2);
-  rb_define_method(cFoo, "find", StringIndexFind, 2);
+  rb_define_alloc_func(classStringIndex, str_idx_alloc);
+  rb_define_method(classStringIndex, "add", StringIndexAddSegments, 2);
+  rb_define_method(classStringIndex, "find", StringIndexFind, 1);
 }
 
 } // End extern "C"
-

data/{rubyext/test.rb → test.rb}

@@ -1,34 +1,31 @@
 #!/usr/bin/env ruby
 
-$:.unshift File.dirname(__FILE__)
 require "stridx"
-idx = CppStringIndex.new
+idx = StrIdx::StringIndex.new
 
 t = Time.new
-fn = File.expand_path("../flist.txt")
+fn = File.expand_path("flist.txt")
 lines = IO.read(fn).lines.collect { |x| x.strip }
-i = 1
+i = 0
 for x in lines
   idx.add(x, i)
   i += 1
 end
 
 idx_time = Time.new 
-puts "\nIndexing time: #{idx_time - t}"
-query = "helbind.h"
-res = idx.find(query, 2)
+puts "\nIndexing time (#{lines.size} files}): #{(idx_time - t).round(4)} seconds"
+
+query = "rngnomadriv"
+res = idx.find(query)
 puts "query: #{query}"
 puts "\nResults:"
 puts "Filename, score"
 puts "==============="
-for x in res
-  fn = lines[x[0] - 1]
-  score = x[1]
+for id, score in res
+  fn = lines[id]
   puts "#{fn}, #{score.round(4)}"
-  # pp [lines[x[0] - 1], x[1]]
 end
 
-
 query_time = Time.new 
 
-puts "\nSearch time: #{query_time - idx_time}"
+puts "\nSearch time: #{(query_time - idx_time).round(4)} seconds"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: StrIdx
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
 platform: ruby
 authors:
 - Sami Sieranoja
@@ -52,8 +52,8 @@ files:
 - demo.cpp
 - rubyext/extconf.rb
 - rubyext/ruby_interf.cpp
-- rubyext/test.rb
 - stridx.hpp
+- test.rb
 - unordered_dense.h
 homepage: https://github.com/SamiSieranoja/stridx
 licenses: