This file is indexed.

/usr/include/anfo/logdom.h is in libanfo0-dev 0.98-4.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
//    Copyright 2009 Udo Stenzel
//    This file is part of ANFO
//
//    ANFO is free software: you can redistribute it and/or modify
//    it under the terms of the GNU General Public License as published by
//    the Free Software Foundation, either version 3 of the License, or
//    (at your option) any later version.
//
//    Anfo is distributed in the hope that it will be useful,
//    but WITHOUT ANY WARRANTY; without even the implied warranty of
//    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
//    GNU General Public License for more details.
//
//    You should have received a copy of the GNU General Public License
//    along with Anfo.  If not, see <http://www.gnu.org/licenses/>.

#ifndef INCLUDED_LOGDOM_H
#define INCLUDED_LOGDOM_H

#include <utility>
#include <math.h>
#include <stdint.h>

//! \brief representation of floating point number in log domain
//! This class is supposed to behave like ordinary floating point
//! numbers, but the representation is in the log domain, so we don't
//! lose precision due to multiplication of many small values.  For ease
//! of conversion, we'll use the phred scale.  That means more work in
//! additions, but less on conversion.  Assuming \f$ q_2 \geq q_1 \f$
//! the formula for additions is:
//!
//! \f[ q_s = q_1 - \frac{10}{\ln 10} \ln \left( 1 + \exp \left( \frac{\ln 10}{10} (q_1-q_2) \right) \right) \f]
//!
//! It turns out that IEEE floating point handles corner cases well:
//! from_float(0) gives an internal representation of \f$ -\inf \f$,
//! which works fine in calculations, except that adding two such
//! numbers gives NaN (so don't do that).

class Logdom {
	private:
		double v_ ;

		explicit Logdom( double v ) : v_(v) {}

	public:
		//! \brief constructs a value of one(!)
		//! Why one and not zero?  Because we tend to multiply things,
		//! and one is the neutral element in that case.
		Logdom() : v_(0) {}

		void swap( Logdom &rhs ) throw() { std::swap( v_, rhs.v_ ) ; }
		bool is_finite() const { return isfinite( v_ ) ; }

		//! \brief calculates log( 1 + exp x )
		//! Calculation is done in a way that preserves accuracy, the
		//! base of the logarithm is chosen to coincide with the Phred
		//! scale.
		//! \todo This function need to be accurate only for x < 0; it
		//!       could probably be approximated using a Taylor series.
		static double ld1pexp10( double x ) { return -10.0 / log(10.0) * log1p(  exp( -log(10.0)/10.0*x ) ) ; }

		//! \brief calculates log( 1 - exp x )
		//! \see ld1pexp10()
		static double ld1mexp10( double x ) { return -10.0 / log(10.0) * log1p( -exp( -log(10.0)/10.0*x ) ) ; }

		static Logdom null() { return from_float( 0 ) ; }
		static Logdom one() { return from_float( 1 ) ; }

		//! \brief converts an ordinary number to log domain
		static Logdom from_float( double v ) { return Logdom( -10 * log(v) / log(10.0) ) ; }

		//! \brief converts a Phred-quality score to log domain
		//! This will normally result in a very small number; the Phred
		//! score represents the probability of something being wrong.
		static Logdom from_phred( int p ) { return Logdom( p ) ; }

		int to_phred() const { return int( v_ + 0.5 ) ; }
		uint8_t to_phred_byte() const { int p = to_phred() ; return p > 255 ? 255 : p ; }
		double to_float() const { return exp( -log(10.0) * v_ / 10.0 ) ; }
		Logdom sqrt() const { return Logdom( v_ / 2 ) ; }

		// multiplication is addition, division is subtraction
		Logdom& operator *= ( Logdom b ) { v_ += b.v_ ; return *this ; }
		Logdom& operator /= ( Logdom b ) { v_ -= b.v_ ; return *this ; }

		Logdom operator * ( Logdom b ) const { return Logdom( v_ + b.v_ ) ; }
		Logdom operator / ( Logdom b ) const { return Logdom( v_ - b.v_ ) ; }

		// addition w/o sacrificing precision is a bit harder
		Logdom operator + ( Logdom b ) const
		{
			if( b == null() ) return *this ;
			if( *this == null() ) return b ;

			return Logdom( v_ <= b.v_
					?   v_ + ld1pexp10( b.v_ -   v_ ) 
					: b.v_ + ld1pexp10(   v_ - b.v_ ) ) ;
		}

		Logdom operator - ( Logdom b ) const
		{
			if( b == null() ) return *this ;

			return Logdom( v_ <= b.v_
					?   v_ + ld1mexp10( b.v_ -   v_ ) 
					: b.v_ + ld1mexp10(   v_ - b.v_ ) ) ; 
		}

		Logdom& operator += ( Logdom b )
		{
			if( *this == null() ) {
					v_ = b.v_ ;
			} 
			else if( b != null() ) {
				v_ = v_ <= b.v_
					?   v_ + ld1pexp10( b.v_ -   v_ ) 
					: b.v_ + ld1pexp10(   v_ - b.v_ ) ;
			}
			return *this ;
		}

		Logdom& operator -= ( Logdom b )
		{
			if( b != null() ) {
				v_ = v_ <= b.v_
					?   v_ + ld1mexp10( b.v_ -   v_ ) 
					: b.v_ + ld1mexp10(   v_ - b.v_ ) ; 
			}
			return *this ;
		}

		bool operator >  ( Logdom b ) const { return b.v_ >  v_ ; }
		bool operator >= ( Logdom b ) const { return b.v_ >= v_ ; }
		bool operator <  ( Logdom b ) const { return b.v_ <  v_ ; }
		bool operator <= ( Logdom b ) const { return b.v_ <= v_ ; }

		friend inline bool operator == ( Logdom a, Logdom b ) { return a.v_ == b.v_ ; }
		friend inline bool operator != ( Logdom a, Logdom b ) { return a.v_ != b.v_ ; }
} ;

inline Logdom operator + ( double a, Logdom b ) { return Logdom::from_float( a ) + b ; }
inline Logdom operator + ( Logdom a, double b ) { return a + Logdom::from_float( b ) ; }
inline Logdom operator - ( double a, Logdom b ) { return Logdom::from_float( a ) - b ; }
inline Logdom operator - ( Logdom a, double b ) { return a - Logdom::from_float( b ) ; }
inline Logdom operator * ( double a, Logdom b ) { return Logdom::from_float( a ) * b ; }
inline Logdom operator * ( Logdom a, double b ) { return a * Logdom::from_float( b ) ; }
inline Logdom operator / ( double a, Logdom b ) { return Logdom::from_float( a ) / b ; }
inline Logdom operator / ( Logdom a, double b ) { return a / Logdom::from_float( b ) ; }

template <typename A, typename B> 
A lerp( B p, A a, A b )
{ return (1.0-p) * a + p * b ; }

#endif