From 6d395aa1af961fb29af75fbf655885d971078bc4 Mon Sep 17 00:00:00 2001
From: van51 <vangelis_51@hotmail.com>
Date: Mon, 5 Aug 2013 13:52:11 +0300
Subject: [PATCH] Added HashedDocDotFeatures in modular interfaces

---
 .../features_hasheddocdot_modular.py          | 37 +++++++++++++++++++
 src/interfaces/modular/Features.i             |  2 +
 src/interfaces/modular/Features_includes.i    |  1 +
 src/interfaces/modular/Library.i              |  6 +++
 src/interfaces/modular/Library_includes.i     |  3 ++
 src/shogun/features/HashedDocDotFeatures.cpp  |  5 +++
 6 files changed, 54 insertions(+)
 create mode 100644 examples/undocumented/python_modular/features_hasheddocdot_modular.py

diff --git a/examples/undocumented/python_modular/features_hasheddocdot_modular.py b/examples/undocumented/python_modular/features_hasheddocdot_modular.py
new file mode 100644
index 00000000000..def4aa5a22c
--- /dev/null
+++ b/examples/undocumented/python_modular/features_hasheddocdot_modular.py
@@ -0,0 +1,37 @@
+#!/usr/bin/env python
+strings=['hey','guys','i','am','a','string']
+
+parameter_list=[[strings]]
+
+def features_hasheddocdot_modular(strings):
+	from shogun.Features import StringCharFeatures, RAWBYTE
+	from shogun.Features import HashedDocDotFeatures
+	from shogun.Library import NGramTokenizer
+	from numpy import array
+
+	#create string features
+	f=StringCharFeatures(strings, RAWBYTE)
+	
+	#set the number of bits of the target dimension
+	#means a dim of size 2^5=32
+	num_bits=5
+
+	#create the ngram tokenizer of size 8 to parse the strings
+	tokenizer=NGramTokenizer(8)
+
+	#normalize results
+	normalize=True
+
+	#create HashedDocDot features
+	hddf=HashedDocDotFeatures(num_bits, f, tokenizer, normalize)
+	
+	#should expect 32
+	#print('Feature space dimensionality is', hddf.get_dim_feature_space())
+	
+	#print('Self dot product of string 0', hddf.dot(0, hddf, 0))
+
+	return hddf
+
+if __name__=='__main__':
+	print('HashedDocDotFeatures')
+	features_hasheddocdot_modular(*parameter_list[0])
diff --git a/src/interfaces/modular/Features.i b/src/interfaces/modular/Features.i
index 3ae144b7ccc..0ff0c3418be 100644
--- a/src/interfaces/modular/Features.i
+++ b/src/interfaces/modular/Features.i
@@ -50,6 +50,7 @@
 %rename(AttributeFeatures) CAttributeFeatures;
 %rename(CombinedFeatures) CCombinedFeatures;
 %rename(CombinedDotFeatures) CCombinedDotFeatures;
+%rename(HashedDocDotFeatures) CHashedDocDotFeatures;
 %rename(Labels) CLabels;
 %rename(LabelsFactory) CLabelsFactory;
 
@@ -477,6 +478,7 @@ namespace shogun
 %include <shogun/features/AttributeFeatures.h>
 %include <shogun/features/CombinedFeatures.h>
 %include <shogun/features/CombinedDotFeatures.h>
+%include <shogun/features/HashedDocDotFeatures.h>
 
 %include <shogun/labels/Labels.h>
 %include <shogun/labels/LabelsFactory.h>
diff --git a/src/interfaces/modular/Features_includes.i b/src/interfaces/modular/Features_includes.i
index 3f15c489d67..82421076574 100644
--- a/src/interfaces/modular/Features_includes.i
+++ b/src/interfaces/modular/Features_includes.i
@@ -20,6 +20,7 @@
 #include <shogun/features/Alphabet.h>
 #include <shogun/features/CombinedFeatures.h>
 #include <shogun/features/CombinedDotFeatures.h>
+#include <shogun/features/HashedDocDotFeatures.h>
 #include <shogun/labels/Labels.h>
 #include <shogun/labels/LabelsFactory.h>
 #include <shogun/labels/DenseLabels.h>
diff --git a/src/interfaces/modular/Library.i b/src/interfaces/modular/Library.i
index 90b6b7eec8c..9e860d941c1 100644
--- a/src/interfaces/modular/Library.i
+++ b/src/interfaces/modular/Library.i
@@ -21,6 +21,9 @@
 %rename(Hash) CHash;
 %rename(StructuredData) CStructuredData;
 %rename(DynamicObjectArray) CDynamicObjectArray;
+%rename(Tokenizer) CTokenizer;
+%rename(DelimiterTokenizer) CDelimiterTokenizer;
+%rename(NGramTokenizer) CNGramTokenizer;
 
 %rename(IndexBlock) CIndexBlock;
 %rename(IndexBlockRelation) CIndexBlockRelation;
@@ -458,6 +461,9 @@ namespace shogun
 /* Hash */
 %include <shogun/lib/Hash.h>
 
+%include <shogun/lib/Tokenizer.h>
+%include <shogun/lib/DelimiterTokenizer.h>
+%include <shogun/lib/NGramTokenizer.h>
 %include <shogun/lib/Cache.h>
 %include <shogun/lib/List.h>
 %include <shogun/lib/Signal.h>
diff --git a/src/interfaces/modular/Library_includes.i b/src/interfaces/modular/Library_includes.i
index 77ae0c463be..7304e586384 100644
--- a/src/interfaces/modular/Library_includes.i
+++ b/src/interfaces/modular/Library_includes.i
@@ -26,4 +26,7 @@
 #include <shogun/lib/IndexBlockGroup.h>
 #include <shogun/lib/IndexBlockTree.h>
 #include <shogun/lib/Data.h>
+#include <shogun/lib/Tokenizer.h>
+#include <shogun/lib/DelimiterTokenizer.h>
+#include <shogun/lib/NGramTokenizer.h>
 %}
diff --git a/src/shogun/features/HashedDocDotFeatures.cpp b/src/shogun/features/HashedDocDotFeatures.cpp
index 607ed24555a..6b42130b117 100644
--- a/src/shogun/features/HashedDocDotFeatures.cpp
+++ b/src/shogun/features/HashedDocDotFeatures.cpp
@@ -27,6 +27,11 @@ CHashedDocDotFeatures::CHashedDocDotFeatures(const CHashedDocDotFeatures& orig)
 	init(orig.num_bits, orig.doc_collection, orig.tokenizer, orig.should_normalize);
 }
 
+CHashedDocDotFeatures::CHashedDocDotFeatures(CFile* loader)
+{
+	SG_NOTIMPLEMENTED;
+}
+
 void CHashedDocDotFeatures::init(int32_t hash_bits, CStringFeatures<char>* docs, 
 	CTokenizer* tzer, bool normalize)
 {