sooth 0.3.0 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d6ee402b249e9c9a15a067e007f83a932f1024f6
-  data.tar.gz: 0a815d99956cc8701d3ac094b37d091bbe282137
+  metadata.gz: 27347b90bbbfd21db684190405f633f9932cff4a
+  data.tar.gz: 122ebc71eefabd6d1d95ad8ef15712033bcb9b9f
 SHA512:
-  metadata.gz: c1c1340951c1045048cae75ee46805724b8dac5480493fc50b4e18f189478677725c9dc729c8295677ef9ebdee0a740648f7724f528d4a2c2931157c9d04792d
-  data.tar.gz: a076a7603490c70fd4aa9d87ad3f93dfdb53621f3afdecb3aad9260904065a8112c4a877c31130315be3f34e8ecf66ecc15f35e49c1cd647ae854d4b7f8ca9eb
+  metadata.gz: 4462abb25b5f6c0a719be89b2e20cfe477f990757a54ae69f427cc3280938178f89cdfab01b2e9e605d414b83d17702e6a764c524669fec0b7204a85efad7db0
+  data.tar.gz: fccf92fb019587081ed45645cd2effe4028e81bdd69ee3e38b1528b5e1eca20d1f2d4dc455f2c17b915345c4e3fbaf7f340b627e6fc2f2fe014a58a2c42b7cc5

data/VERSION

@@ -1 +1 @@
-0.3.0
+0.4.0

data/ext/sooth_native/native.c

@@ -174,6 +174,13 @@ method_sooth_native_clear(VALUE self)
 VALUE
 method_sooth_native_load(VALUE self, VALUE filename)
 {
+  SoothPredictor * predictor = NULL;
+  Check_Type(filename, T_STRING);
+  Data_Get_Struct(self, SoothPredictor, predictor);
+  if (!sooth_predictor_load(RSTRING_PTR(filename), predictor))
+  {
+    rb_raise(rb_eIOError, "problem loading predictor");
+  }
   return Qnil;
 }
 
@@ -182,6 +189,13 @@ method_sooth_native_load(VALUE self, VALUE filename)
 VALUE
 method_sooth_native_save(VALUE self, VALUE filename)
 {
+  SoothPredictor * predictor = NULL;
+  Check_Type(filename, T_STRING);
+  Data_Get_Struct(self, SoothPredictor, predictor);
+  if (!sooth_predictor_save(RSTRING_PTR(filename), predictor))
+  {
+    rb_raise(rb_eIOError, "problem saving predictor");
+  }
   return Qnil;
 }
 
@@ -220,7 +234,7 @@ method_sooth_native_count(VALUE self, VALUE bigram)
   Check_Type(RARRAY_PTR(bigram)[1], T_FIXNUM);
   Data_Get_Struct(self, SoothPredictor, predictor);
   uint32_t c_bigram[2] = {NUM2UINT(RARRAY_PTR(bigram)[0]), NUM2UINT(RARRAY_PTR(bigram)[1])};
-  uint64_t count = sooth_predictor_count(predictor, c_bigram);
+  uint32_t count = sooth_predictor_count(predictor, c_bigram);
   return UINT2NUM(count);
 }
 

data/ext/sooth_native/sooth_context.h

@@ -10,7 +10,7 @@
 typedef struct
 {
   uint32_t bigram[2];
-  uint64_t count;
+  uint32_t count;
   uint32_t statistics_size;
   SoothStatistic * statistics;
 }

data/ext/sooth_native/sooth_predictor.c

@@ -1,5 +1,6 @@
 //==============================================================================
 
+#include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 
@@ -7,6 +8,24 @@
 
 //------------------------------------------------------------------------------
 
+void
+sooth_show_predictor(SoothPredictor * predictor)
+{
+  printf("Error Symbol: %u\n", predictor->error_symbol);
+  for (uint32_t i = 0; i < predictor->contexts_size; ++i)
+  {
+    SoothContext context = predictor->contexts[i];
+    printf("  Context %u-%u (%u)\n", context.bigram[0], context.bigram[1], context.count);
+    for (uint32_t j = 0; j < context.statistics_size; ++j)
+    {
+      SoothStatistic statistic = context.statistics[j];
+      printf("    Symbol %u (%u)\n", statistic.symbol, statistic.count);
+    }
+  }
+}
+
+//------------------------------------------------------------------------------
+
 SoothPredictor *
 sooth_predictor_init()
 {
@@ -29,7 +48,7 @@ sooth_predictor_init()
 void
 sooth_predictor_clear(SoothPredictor * predictor)
 {
-  for (uint64_t i = 0; i < predictor->contexts_size; ++i)
+  for (uint32_t i = 0; i < predictor->contexts_size; ++i)
   {
     SoothContext * context = &(predictor->contexts[i]);
     free(context->statistics);
@@ -53,22 +72,81 @@ sooth_predictor_free(SoothPredictor * predictor)
 
 //------------------------------------------------------------------------------
 
-void
+bool
 sooth_predictor_save(const char * const filename, SoothPredictor * predictor)
 {
-}
+  FILE *file = fopen(filename, "wb");
 
-//------------------------------------------------------------------------------
+  if (file == NULL)
+  {
+    return false;
+  }
 
-void sooth_predictor_merge(const char * const filename, SoothPredictor * predictor)
-{
+  fwrite("MH10", 1, 4, file);
+  fwrite(&predictor->error_symbol, 4, 1, file);
+  fwrite(&predictor->contexts_size, 4, 1, file);
+  for (uint32_t i = 0; i < predictor->contexts_size; ++i)
+  {
+    SoothContext context = predictor->contexts[i];
+    fwrite(context.bigram, 4, 2, file);
+    fwrite(&context.count, 4, 1, file);
+    fwrite(&context.statistics_size, 4, 1, file);
+    fwrite(context.statistics, sizeof(SoothStatistic), context.statistics_size, file);
+  }
+
+  fclose(file);
+
+  return true;
 }
+
 //------------------------------------------------------------------------------
 
-void sooth_predictor_load(const char * const filename, SoothPredictor * predictor)
+bool sooth_predictor_load(const char * const filename, SoothPredictor * predictor)
 {
+  FILE *file = fopen(filename, "rb");
+
+  if (file == NULL)
+  {
+    return false;
+  }
+
+  char code[4];
+  fread(code, 1, 4, file);
+  if (strncmp(code, "MH10", 4) != 0)
+  {
+    return false;
+  }
+
   sooth_predictor_clear(predictor);
-  sooth_predictor_merge(predictor, filename);
+
+  fread(&predictor->error_symbol, 4, 1, file);
+  fread(&predictor->contexts_size, 4, 1, file);
+  if (predictor->contexts_size == 0)
+  {
+    return true;
+  }
+  predictor->contexts = malloc(sizeof(SoothContext) * predictor->contexts_size);
+  if (predictor->contexts == NULL)
+  {
+    sooth_predictor_clear(predictor);
+    return NULL;
+  }
+  for (uint32_t i = 0; i < predictor->contexts_size; ++i)
+  {
+    SoothContext * context = &(predictor->contexts[i]);
+    fread(context->bigram, 4, 2, file);
+    fread(&context->count, 4, 1, file);
+    fread(&context->statistics_size, 4, 1, file);
+    context->statistics = malloc(sizeof(SoothStatistic) * context->statistics_size);
+    if (context->statistics == NULL)
+    {
+      sooth_predictor_clear(predictor);
+      return NULL;
+    }
+    fread(context->statistics, sizeof(SoothStatistic), context->statistics_size, file);
+  }
+
+  return true;
 }
 
 //------------------------------------------------------------------------------
@@ -76,12 +154,12 @@ SoothContext *
 sooth_predictor_find_context(SoothPredictor * predictor, uint32_t bigram[2])
 {
   SoothContext * context = NULL;
-  uint64_t mid = 0;
+  uint32_t mid = 0;
 
   if (predictor->contexts_size > 0)
   {
-    uint64_t low = 0;
-    uint64_t high = predictor->contexts_size - 1;
+    uint32_t low = 0;
+    uint32_t high = predictor->contexts_size - 1;
 
     while (low <= high)
     {
@@ -139,12 +217,12 @@ SoothStatistic *
 sooth_predictor_find_statistic(SoothContext * context, uint32_t symbol)
 {
   SoothStatistic * statistic = NULL;
-  uint64_t mid = 0;
+  uint32_t mid = 0;
 
   if (context->statistics_size > 0)
   {
-    uint64_t low = 0;
-    uint64_t high = context->statistics_size - 1;
+    uint32_t low = 0;
+    uint32_t high = context->statistics_size - 1;
 
     while (low <= high)
     {
@@ -220,7 +298,7 @@ sooth_predictor_observe(SoothPredictor * predictor, uint32_t bigram[2], uint32_t
 
 //------------------------------------------------------------------------------
 
-uint64_t
+uint32_t
 sooth_predictor_count(SoothPredictor * predictor, uint32_t bigram[2])
 {
   SoothContext * context = sooth_predictor_find_context(predictor, bigram);
@@ -236,7 +314,7 @@ sooth_predictor_count(SoothPredictor * predictor, uint32_t bigram[2])
 //------------------------------------------------------------------------------
 
 uint32_t
-sooth_predictor_select(SoothPredictor * predictor, uint32_t bigram[2], uint64_t limit)
+sooth_predictor_select(SoothPredictor * predictor, uint32_t bigram[2], uint32_t limit)
 {
   SoothContext * context = sooth_predictor_find_context(predictor, bigram);
 
@@ -245,7 +323,7 @@ sooth_predictor_select(SoothPredictor * predictor, uint32_t bigram[2], uint64_t
     return predictor->error_symbol;
   }
 
-  for (uint64_t i = 0; i < context->statistics_size; ++i)
+  for (uint32_t i = 0; i < context->statistics_size; ++i)
   {
     SoothStatistic statistic = context->statistics[i];
     if (limit > statistic.count)

data/ext/sooth_native/sooth_predictor.h

@@ -4,13 +4,14 @@
 //==============================================================================
 
 #include <stdint.h>
+#include <stdbool.h>
 
 #include "sooth_context.h"
 
 typedef struct
 {
   uint32_t error_symbol;
-  uint64_t contexts_size;
+  uint32_t contexts_size;
   SoothContext * contexts;
 }
 SoothPredictor;
@@ -20,12 +21,11 @@ SoothPredictor;
 SoothPredictor * sooth_predictor_init();
 void sooth_predictor_clear(SoothPredictor * predictor);
 void sooth_predictor_free(SoothPredictor * predictor);
-void sooth_predictor_merge(const char * const filename, SoothPredictor * predictor);
-void sooth_predictor_load(const char * const filename, SoothPredictor * predictor);
-void sooth_predictor_save(const char * const filename, SoothPredictor * predictor);
+bool sooth_predictor_load(const char * const filename, SoothPredictor * predictor);
+bool sooth_predictor_save(const char * const filename, SoothPredictor * predictor);
 uint32_t sooth_predictor_observe(SoothPredictor * predictor, uint32_t bigram[2], uint32_t symbol);
-uint64_t sooth_predictor_count(SoothPredictor * predictor, uint32_t bigram[2]);
-uint32_t sooth_predictor_select(SoothPredictor * predictor, uint32_t bigram[2], uint64_t limit);
+uint32_t sooth_predictor_count(SoothPredictor * predictor, uint32_t bigram[2]);
+uint32_t sooth_predictor_select(SoothPredictor * predictor, uint32_t bigram[2], uint32_t limit);
 
 //==============================================================================
 

data/sooth.gemspec

@@ -2,17 +2,17 @@
 # DO NOT EDIT THIS FILE DIRECTLY
 # Instead, edit Jeweler::Tasks in Rakefile, and run 'rake gemspec'
 # -*- encoding: utf-8 -*-
-# stub: sooth 0.3.0 ruby lib
+# stub: sooth 0.4.0 ruby lib
 # stub: ext/sooth_native/extconf.rb
 
 Gem::Specification.new do |s|
   s.name = "sooth"
-  s.version = "0.3.0"
+  s.version = "0.4.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.require_paths = ["lib"]
   s.authors = ["Jason Hutchens"]
-  s.date = "2014-12-11"
+  s.date = "2014-12-16"
   s.description = "Sooth is a simple stochastic predictive model."
   s.email = "jasonhutchens@gmail.com"
   s.extensions = ["ext/sooth_native/extconf.rb"]

data/spec/predictor_spec.rb

@@ -1,3 +1,5 @@
+require 'tempfile'
+
 require File.expand_path(File.dirname(__FILE__) + '/spec_helper')
 
 describe Sooth::Predictor do
@@ -102,4 +104,34 @@ describe Sooth::Predictor do
     end
 
   end
+
+  describe "#save" do
+
+    it "can save a file and load it back again" do
+      begin
+        file = Tempfile.new('sooth_spec')
+        expect(predictor.observe([1,2], 3)).to eq(1)
+        expect(predictor.observe([2,1], 3)).to eq(1)
+        expect(predictor.observe([1,2], 3)).to eq(2)
+        expect(predictor.observe([1,2], 3)).to eq(3)
+        expect { predictor.save(file.path) } .to_not raise_error
+        expect(predictor.count([1,2])).to eq(3)
+        expect(predictor.count([2,1])).to eq(1)
+        predictor.clear
+        expect(predictor.count([1,2])).to eq(0)
+        expect(predictor.count([2,1])).to eq(0)
+        expect { predictor.load(file.path) }.to_not raise_error
+        expect(predictor.count([1,2])).to eq(3)
+        expect(predictor.count([2,1])).to eq(1)
+        expect(predictor.observe([1,2], 3)).to eq(4)
+        expect(predictor.observe([1,2], 1)).to eq(1)
+        expect(predictor.observe([2,1], 3)).to eq(2)
+        expect(predictor.observe([2,1], 1)).to eq(1)
+      ensure
+        file.close
+        file.unlink
+      end
+    end
+
+  end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sooth
 version: !ruby/object:Gem::Version
-  version: 0.3.0
+  version: 0.4.0
 platform: ruby
 authors:
 - Jason Hutchens
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-12-11 00:00:00.000000000 Z
+date: 2014-12-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec