SelFile Class Reference

#include <selfile.h>

Public Member Functions
Selfile constructors
	SelFile ()
	SelFile (const FileName &name)
void	reserve (int n)
void	clear ()
Managing files in disk
void	read (const FileName &name, int overrinding=1)
void	append (const FileName &name)
void	merge (const FileName &name)
void	merge (SelFile &sel)
SelFile	operator+ (SelFile &sel)
void	split_in_two (SelFile &sel1, SelFile &sel2)
void	split_in_N (int N, std::vector< SelFile > &parts)
void	mpi_select_part (int rank, int size, int &num_img_tot)
void	mpi_select_part2 (int rank, int size, int &num_img_tot, int mpi_job_size)
void	chooseSubset (int firstImage, int lastImage, SelFile &SFsubset)
void	write (const FileName &sel_name="")
Moving the current line pointer
void	go_beginning ()
void	go_first_ACTIVE ()
const std::string &	operator() (int i)
const std::string &	NextImg (SelLine::Label label=SelLine::ACTIVE)
void	next ()
void	jump (int count, SelLine::Label label=SelLine::ACTIVE)
bool	jump_lines (int count)
void	search (const std::string &img_name)
int	eof ()
Getting information
FileName	name () const
int	Is_ACTIVE () const
int	Is_DISCARDED () const
int	Is_COMMENT () const
const SelLine &	current ()
void	get_current (SelLine &_SL)
int	exists (const std::string &img_name)
int	ImgNo (SelLine::Label label=SelLine::ACTIVE) const
int	LineNo ()
void	ImgSize (int &Ydim, int &Xdim)
FileName	FileExtension ()
int	MaxFileNameLength ()
const std::string	get_current_file ()
SelLine	get_current_line ()
const std::string	get_file_number (int i)
Modifying the selection file
void	remove (const std::string &img_name)
void	remove_current ()
void	set (const std::string &img_name, SelLine::Label label)
void	set_current (SelLine::Label label)
void	set_current_filename (const FileName &fn_new)
void	insert (const std::string &img_name, SelLine::Label label=SelLine::ACTIVE)
void	insert (const SelLine &_selline)
void	insert_comment (const std::string &comment)
void	clean ()
void	clean_comments ()

Some operators
void	assign (const SelFile &sel)
std::ostream &	operator<< (std::ostream &o, const SelFile &sel)

Helpful procedures
SelFile	sort_by_filenames ()
SelFile	randomize ()
SelFile	randomSubset (int subsetN, bool withReplacement=true)
SelFile	random_discard (int N)
SelFile	compare (SelFile &SF1, SelFile &SF2, const int mode)

Detailed Description

Selection File

The SelFile is an object which keeps in memory all the information associated to a .sel file.

Definition at line 168 of file selfile.h.

Constructor & Destructor Documentation

SelFile() [1/2]

SelFile::SelFile

(

)

Empty constructor.

There is no file associated yet.

SelFile sel;

Definition at line 122 of file selfile.cpp.

{
    fn_sel       = "Unnamed";
    no_imgs      = 0;
    current_line = text_line.begin();
}

SelFile() [2/2]

SelFile::SelFile ( const FileName &  name )

inline

Constructor with filename, read from disk.

The given name is loaded (method read) as a selection file.

SelFile sel("g1t.sel");

Definition at line 209 of file selfile.h.

    {
        read(name);
    }

Member Function Documentation

append()

void SelFile::append ( const FileName &  name )

inline

Append a file from disk to an already read one.

The old information on the variable is not lost. All lines in the selection file to be read are appened at the end of the already read one without any kind of check.

sel.read("g1t.sel");
sel.append("g2t.sel");

Definition at line 280 of file selfile.h.

    {
        read(name, 0);
    }

assign()

void SelFile::assign

(

const SelFile &

sel

)

Another function for assigment.

Definition at line 139 of file selfile.cpp.

{
    *this = SF;
}

chooseSubset()

void SelFile::chooseSubset	(	int	firstImage,
		int	lastImage,
		SelFile &	SFsubset
	)

Choose a subset of images.

Definition at line 489 of file selfile.cpp.

{
    SFsubset.clear();
    go_beginning();
    jump(firstImage);
    for (int i=firstImage; i<=lastImage; i++)
    {
        if (!eof()) SFsubset.insert(current());
        next();
    }
}

clean()

void SelFile::clean

(

)

Deletes all DISCARDED images from the selection file.

The current line "pointer" is moved to the beginning of the file.

sel.clean();

Definition at line 158 of file selfile.cpp.

{
    std::vector<SelLine>::iterator current = text_line.begin();
    while (current != text_line.end())
    {
        if ((*current).line_type == SelLine::DATALINE &&
            (*current).label == SelLine::DISCARDED)
        {
            text_line.erase(current);
        }
        else
            current++;
    }
    current_line = text_line.begin();
}

clean_comments()

void SelFile::clean_comments

(

)

Deletes all comments from the selection file.

The current line "pointer" is moved to the beginning of the file.

sel.clean_comments();

Definition at line 175 of file selfile.cpp.

{
    std::vector<SelLine>::iterator current = text_line.begin();
    std::vector<SelLine>::iterator last    = text_line.end();
    std::vector<SelLine>::iterator temp;
    while (current != last)
    {
        if ((*current).line_type == SelLine::COMMENT)
        {
            temp = current;
            temp++;
            text_line.erase(current);
            current = temp;
        }
        else
            current++;
    }
    current_line = text_line.begin();
}

clear()

void SelFile::clear

(

)

Empties the object.

sel.clear();

Definition at line 130 of file selfile.cpp.

{
    fn_sel = "Unnamed";
    text_line.erase(text_line.begin(), text_line.end());
    no_imgs = 0;
    current_line = text_line.begin();
}

current()

const SelLine& SelFile::current ( )

inline

Returns current line as a Sel Line.

Definition at line 526 of file selfile.h.

    {
        return *current_line;
    }

eof()

int SelFile::eof ( )

inline

True if current line "pointer" is at the end of file.

if (sel.eof())
    std::cout << "The selection file is over\n";

Definition at line 485 of file selfile.h.

    {
        return current_line == text_line.end();
    }

exists()

int SelFile::exists ( const std::string &  img_name )

inline

True if the image name is inside the selection file.

The current line "pointer" is not modified. If an image is discarded in the selection file, this function still will say that it exists, although it is discarded.

if (sel.exists("g1ta0001"))
    std::cout << "g1ta0001 exists in the selection file\n";

Definition at line 549 of file selfile.h.

    {
        return find(img_name) != text_line.end();
    }

FileExtension()

FileName SelFile::FileExtension

(

)

Returns the extension of the files inside.

This function returns the extension of the first active file.

Definition at line 620 of file selfile.cpp.

{
    std::vector<SelLine>::iterator aux = current_line;
    go_first_ACTIVE();
    FileName ext = (*current_line).text;
    ext = ext.getExtension();
    current_line = aux;
    return ext;
}

get_current()

void SelFile::get_current ( SelLine &  _SL )

inline

Another function to get a SelLine.

Definition at line 533 of file selfile.h.

    {
        _SL = current();
    }

get_current_file()

const std::string SelFile::get_current_file

(

)

Get the filename of the current line.

If the current line "pointer" is at the end of the file or is pointing to a comment then an empty string is returned.

fn = sel.get_current_file();

Definition at line 647 of file selfile.cpp.

{
    if (current_line == text_line.end())
        return "";
    if ((*current_line).line_type != SelLine::DATALINE)
        return "";
    return (*current_line).text;
}

get_current_line()

SelLine SelFile::get_current_line ( )

inline

Get current line.

Definition at line 621 of file selfile.h.

    {
        return *current_line;
    }

get_file_number()

const std::string SelFile::get_file_number

(

int

i

)

Get the filename at the ACTIVE line number i.

The first file is number 0. If i is greater than the total number of ACTIVE files, then "" is returned.

fn = sel.get_file_number(i);

Definition at line 657 of file selfile.cpp.

{
    if (i < 0)
        return "";
    std::vector<SelLine>::iterator current = text_line.begin();
    std::vector<SelLine>::iterator last    = text_line.end();

    int currenti = 0;
    while (current != last)
    {
        if ((*current).line_type == SelLine::DATALINE &&
            (*current).label == SelLine::ACTIVE)
            currenti++;
        if (currenti > i)
            return (*current).text;
        current++;
    }
    return "";
}

go_beginning()

void SelFile::go_beginning ( )

inline

Go to the beginning of the file.

Moves the pointer to the first line of the file either it is a comment, an active image or a discarded one.

sel.go_beginning();

Definition at line 364 of file selfile.h.

    {
        current_line = text_line.begin();
    }

go_first_ACTIVE()

void SelFile::go_first_ACTIVE ( )

inline

Go to the first ACTIVE image.

Moves the pointer to the first active image in the file.

sel.go_first_ACTIVE();

Definition at line 377 of file selfile.h.

    {
        go_beginning();
        adjust_to_label(SelLine::ACTIVE);
    }

ImgNo()

int SelFile::ImgNo

(

SelLine::Label

label = SelLine::ACTIVE

)

const

Number of images inside a selection file with a certain label.

This function returns the number of images inside the selection file with a given label. By default this label is ACTIVE.

std::cout << "There are " << sel.ImgNo() << " active images\n";
std::cout << "There are " << sel.ImgNo(SelLine::ACTIVE)
    << " active images\n";
std::cout << "There are " << sel.ImgNo(SelLine::DISCARDED)
    << " discarded images\n";

Definition at line 589 of file selfile.cpp.

{
    int N = 0;
    std::vector<SelLine>::const_iterator current = text_line.begin();
    std::vector<SelLine>::const_iterator last    = text_line.end();
    while (current != last)
    {
        if ((*current).line_type == SelLine::DATALINE &&
            (*current).label == label)
            N++;
        current++;
    }
    return N;
}

ImgSize()

void SelFile::ImgSize	(	int &	Ydim,
		int &	Xdim
	)

Returns the size of the images inside.

The filenames within a selection file are supposed to be for SPIDER images, this function opens one of the images (an active one) and returns the size of that image, supposed to be the same for the rest of the images in the selection file.

An exception is thrown if the first valid image in the selfile, doesn't exist in the disk or it is not a XMIPP image.

sel.ImgSize(y, x);

insert() [1/2]

void SelFile::insert	(	const std::string &	img_name,
		SelLine::Label	label = SelLine::ACTIVE
	)

Insert image before current line.

There is no checking for the previous existence of the img. The current line is still pointing to the same line as it was before entering the function.

sel.insert("g1ta0000");
sel.insert("g1ta0000", SelLine::DISCARDED);

Definition at line 759 of file selfile.cpp.

{
    SelLine temp;
    temp.line_type = SelLine::DATALINE;
    temp.text = img_name;
    temp.label = label;
    if (label != SelLine::DISCARDED)
        no_imgs++;

    // Insert and updates current_line
    current_line = text_line.insert(current_line, temp);
    current_line++;
}

insert() [2/2]

void SelFile::insert

(

const SelLine &

_selline

)

Insert line before current line.

It is checked that the line is either a comment or data line, in this case that the label is right, too. The current line is still pointing to the same line as it was before entering the function.

sel.insert("g1ta0000", SelLine::ACTIVE);

Definition at line 774 of file selfile.cpp.

{
    if (_selline.line_type != SelLine::DATALINE &&
        _selline.line_type != SelLine::COMMENT)
        REPORT_ERROR(ERR_SELFILE, "SelFile::insert(SelLine): SelLine type not valid");
    if (_selline.line_type == SelLine::DATALINE)
        if (_selline.label != SelLine::DISCARDED &&
            _selline.label != SelLine::ACTIVE)
            REPORT_ERROR(ERR_SELFILE, "SelFile::insert(SelLine): SelLine label not valid");

    // Sjors 18sep06: added next line
    if (_selline.label != SelLine::DISCARDED)
        no_imgs++;

    // Insert and updates current_line
    current_line = text_line.insert(current_line, _selline);
    current_line++;
}

insert_comment()

void SelFile::insert_comment

(

const std::string &

comment

)

Insert a comment before the current line.

Comments must not start with any special character since a "#" is automatically added at the beginning of the line. The current line is still pointing to the same line as it was before entering the function.

sel.insert_comment("This is a comment");

Definition at line 794 of file selfile.cpp.

{
    SelLine temp;
    temp.line_type = SelLine::COMMENT;
    temp.text = "# " + comment;
    temp.label = SelLine::DISCARDED;

    // Insert and updates current_line
    current_line = text_line.insert(current_line, temp);
    current_line++;
}

Is_ACTIVE()

int SelFile::Is_ACTIVE ( ) const

inline

True if current line is a data line and it is active.

Definition at line 503 of file selfile.h.

    {
        return current_line->Is_data() &&
               current_line->get_label() == SelLine::ACTIVE;
    }

Is_COMMENT()

int SelFile::Is_COMMENT ( ) const

inline

True if current line is a comment.

Definition at line 519 of file selfile.h.

    {
        return current_line->Is_comment();
    }

Is_DISCARDED()

int SelFile::Is_DISCARDED ( ) const

inline

True if current line is a data line and it is active.

Definition at line 511 of file selfile.h.

    {
        return current_line->Is_data() &&
               current_line->get_label() == SelLine::DISCARDED;
    }

jump()

void SelFile::jump	(	int	count,
		SelLine::Label	label = SelLine::ACTIVE
	)

Jump over a number of lines with a given label.

Starting from the current_line "pointer" this function skips a given number of entries with a certain label. For instance, jump over 1 active image is to jump to the next active image. Jump over 2 active images is to jump to the next of the next active image, and so on. You can give as label DISCARDED, too. The number of images to jump must always be positive, the jump cannot be done backwards.

sel.jump(2); // Jump over 2 active images
sel.jump(2, SelLine::ACTIVE); // The same
sel.jump(2, SelLine::DISCARDED) // Jump over 2 discarded images

Definition at line 540 of file selfile.cpp.

{
    adjust_to_label(label);
    for (int i = 0; i < how_many; i++)
        if (current_line != text_line.end())
        {
            current_line++;
            adjust_to_label(label);
        }
}

jump_lines()

bool SelFile::jump_lines

(

int

count

)

Jump over a number of lines disregarding the label.

Returns false if the end of the line is reached before the number of requested line jumps

Definition at line 527 of file selfile.cpp.

{
    for (int i = 0; i < how_many; i++)
    {
        if (current_line != text_line.end())
            current_line++;
        else
            return false;
    }
    return true;
}

LineNo()

int SelFile::LineNo

(

)

Returns the number of lines within a file.

This function gives the total number of lines (including comments) within a file.

std::cout << "There are " << sel.LineNo() << " lines in this file\n";

Definition at line 605 of file selfile.cpp.

{
    int N = 0;
    std::vector<SelLine>::iterator current = text_line.begin();
    std::vector<SelLine>::iterator last    = text_line.end();
    while (current != last)
    {
        N++;
        current++;
    }
    return N;
}

MaxFileNameLength()

int SelFile::MaxFileNameLength

(

)

Returns the maximum length of an active filename inside the selfile.

The current pointer is not moved.

Definition at line 632 of file selfile.cpp.

{
    std::vector<SelLine>::iterator aux = current_line;
    size_t max_length = 0;
    go_first_ACTIVE();
    while (!eof())
    {
        FileName fn = NextImg();
        max_length = XMIPP_MAX(max_length, fn.length());
    }
    current_line = aux;
    return max_length;
}

merge() [1/2]

void SelFile::merge

(

const FileName &

name

)

Merge a file from disk with an already read one.

All lines (except the comments) in the selection file to be read are either added at the end of the already read one if they are not present at them, either ignored if they are already present, or marked with a comment if the corresponding image name is present in both files but with different labels (active, discarded), in this case the image remains active but a comment in the preceding line informs you of the situation.

sel.read("g1t.sel");
sel.merge("g2t.sel");

Definition at line 254 of file selfile.cpp.

{
    SelFile SF(sel_name);
    *this = *this + SF;
    go_first_ACTIVE();
}

merge() [2/2]

void SelFile::merge

(

SelFile &

sel

)

Merge this file with another selfile.

Definition at line 308 of file selfile.cpp.

{
    std::vector<SelLine>::iterator current = SF.text_line.begin();
    std::vector<SelLine>::iterator last    = SF.text_line.end();
    std::vector<SelLine>::iterator found;

    SelLine discrepancy;
    discrepancy.line_type = SelLine::COMMENT;
    discrepancy.text = "# There were discrepancy in the tags for next line, the "
                       "ACTIVE state is kept";

    while (current != last)
    {
        if ((*current).line_type != SelLine::DATALINE)
        {
            current++;
            continue;
        }
        if ((found = find((*current).text)) == text_line.end())
        {
            // New image not found in the whole Sel File.
            // Add it if it is not discarded
            if ((*current).label != SelLine::DISCARDED)
            {
                text_line.push_back(*current);
                no_imgs++;
            }
        }
        else
            // New image is found, check that its line is not going
            // to be removed, if it is add it again; else, check if
            // there is a discrepancy between them
            if ((*found).label != (*current).label)
            {
                if ((*found).label < (*current).label)
                {
                    (*found).label = SelLine::ACTIVE;
                    no_imgs++;
                }
                text_line.insert(found, 1, discrepancy);
            }
        current++;
    }
}

mpi_select_part()

void SelFile::mpi_select_part	(	int	rank,
		int	size,
		int &	num_img_tot
	)

For MPI-parallelized runs: select relevant part of selfile for that rank.

Definition at line 425 of file selfile.cpp.

{

    (*this).clean_comments();
    (*this).clean();
    num_img_tot = (*this).ImgNo();
    int remaining = num_img_tot % size;
    int Npart = (int)(num_img_tot - remaining) / size;
    int myFirst, myLast;
    if (rank < remaining)
    {
        myFirst = rank * (Npart + 1);
        myLast = myFirst + Npart;
    }
    else
    {
        myFirst = rank * Npart + remaining;
        myLast = myFirst + Npart - 1;
    }
    // Now discard all images in Selfile that are outside myFirst-myLast
    (*this).go_beginning();
    SelFile  SFpart = *this;
    SFpart.clear();
    for (int nr = myFirst; nr <= myLast; nr++)
    {
        (*this).go_beginning();
        (*this).jump_lines(nr);
        SFpart.insert((*this).current());
    }
    *this = SFpart;

}

mpi_select_part2()

void SelFile::mpi_select_part2	(	int	rank,
		int	size,
		int &	num_img_tot,
		int	mpi_job_size
	)

For MPI-parallelized runs: select relevant part of selfile for that rank. I just do not undertand the previous function

Definition at line 458 of file selfile.cpp.

{   // jobNumber process number
    // total number of processes
    // total number of images

    (*this).clean_comments();
    (*this).clean();
    totalNumImg = (*this).ImgNo();
    int myFirst = jobNumber * mpi_job_size;
    int myLast  = myFirst + mpi_job_size-1;
    while ((myLast+1) > totalNumImg)
    {
        myLast = totalNumImg-1;
    }
    // Now discard all images in Selfile that are outside myFirst-myLast
    (*this).go_beginning();
    SelFile  SFpart = *this;
    SFpart.clear();
    for (int nr = myFirst; nr <= myLast; nr++)
    {
        (*this).go_beginning();
        (*this).jump_lines(nr);
        SFpart.insert((*this).current());
    }
    *this = SFpart;
}

name()

FileName SelFile::name ( ) const

inline

Returns the name of the file.

Definition at line 496 of file selfile.h.

    {
        return fn_sel;
    }

next()

void SelFile::next ( )

inline

Move the current pointer to the next image, disregarding its label.

It doesn't matter if next image is ACTIVE or DISCARDED, this function moves the current pointer to it.

sel.go_beginning();
while (!sel.eof())
{
    std::cout << sel.current();
    sel.next();
}

Definition at line 429 of file selfile.h.

    {
        current_line++;
    }

NextImg()

const std::string & SelFile::NextImg

(

SelLine::Label

label = SelLine::ACTIVE

)

Returns the name of the next image with a certain label.

The default label is ACTIVE, ie, by default this function returns the name of the next ACTIVE image. But you can give as label DISCARDED and the function will return the name of the next DISCARDED image starting at the current position of the current_line "pointer". If the file is at the end of the selection file, "" is returned. After this function the "pointer" is actually pointing to the next line following the returned image name.

name = sel.NextImg();  // Next active image
name = sel.NextImg(SelLine::DISCARDED); // Next discarded image

Definition at line 516 of file selfile.cpp.

{
    adjust_to_label(label);
    static const std::string emptyString;
    if (current_line != text_line.end())
        return (*current_line++).text;
    else
        return emptyString;
}

operator()()

const std::string& SelFile::operator() ( int  i )

inline

Get the line i in the file

Get the line i in the file for reading. If it is a comment the comment text is returned. If it is an image, the image filename. The first line in the file is number 0.

std::cout << sel(0) << std::endl;

Definition at line 393 of file selfile.h.

    {
        return text_line[i].text;
    }

operator+()

SelFile SelFile::operator+

(

SelFile &

sel

)

Merge two already read files.

sel1.read("g1t.sel");
sel2.merge("g2t.sel");
sel1 = sel1 + sel2;

Definition at line 355 of file selfile.cpp.

{
    SelFile result;
    result = *this;
    result.merge(SF);
    return result;
}

random_discard()

SelFile SelFile::random_discard

(

int

N

)

Discard randomly N images.

A set of N images are discarded from the actual selection file. If N is equal or greater than the actual number of images within the file, all images are discarded. Comments are kept at their original positions. The current line of the resulting selection file is placed at the beginning of the file.

sel2 = sel1.random_discard(3);

Definition at line 844 of file selfile.cpp.

{
    SelFile  result;
    int      i, rnd_indx;

    SelLine::Label label = SelLine::ACTIVE;
    result = *this;
    N = std::min(N, no_imgs);
    for (i = 0; i < N; i++)
    {
        // Jump a random number from the beginning
        rnd_indx = (int) rnd_unif(0, result.no_imgs);
        result.go_first_ACTIVE();
        result.jump(rnd_indx, label);

        // Discard that image
        (*(result.current_line)).label = SelLine::DISCARDED;

        // Decrease the number of images such that next time
        result.no_imgs--;
    }

    result.go_beginning();
    return result;
}

randomize()

SelFile SelFile::randomize

(

)

Alter order in sel file.

A new selection file with all the images of the actual object (either they are active or discarded) is created, but this time all images are in a random order. The comments of the original selection file are lost in the new copy. The current line of the resulting selection file is placed at the beginning of the file.

sel2 = sel1.randomize();

Definition at line 816 of file selfile.cpp.

{
    SelFile  result, aux;
    int      i;
    int      rnd_indx;

    randomize_random_generator();
    if (no_imgs == 0)
        return aux;
    aux = *this;
    for (i = no_imgs; i > 0; i--)
    {
        // Jump a random number from the beginning
        rnd_indx = (int) rnd_unif(0, i);
        aux.go_first_ACTIVE();
        aux.jump(rnd_indx);
        result.text_line.push_back(*(aux.current_line));
        (*aux.current_line).line_type = SelLine::NOT_CONSIDERED;
    }

    // Adjust remaining fields
    result.no_imgs = no_imgs;
    result.current_line = result.text_line.begin();
    return result;
}

randomSubset()

SelFile SelFile::randomSubset	(	int	subsetN,
		bool	withReplacement = true
	)

Random subset.

A new selection file is created with the number of images specified in subsetN. The new subset can be created with or without replacement.

read()

void SelFile::read	(	const FileName &	name,
		int	overrinding = 1
	)

Read a file from disk.

The old information on the variable is overwritten. An exception is thrown if the file doesn't exist. Lines which do not fit the comment structure or the "image-label" structure are ignored. The image name is limited to MAX_FILENAME_LENGTH characters. After reading the selfile pointer is moved to the first ACTIVE image.

sel.read("g2t.sel");

Definition at line 196 of file selfile.cpp.

{
    SelLine   temp;
    std::ifstream  fh_sel;
    int       line_no = 1;

    // Empties current SelFile
    if (overriding)
        clear();

    // Open file
    else
    {
        // Read normal selfile
        fh_sel.open(sel_name.c_str(), std::ios::in);
        if (!fh_sel)
            REPORT_ERROR(ERR_IO_NOTEXIST, sel_name);

        // Read each line and keep it in the list of the SelFile object
        fh_sel.peek();
        while (!fh_sel.eof())
        {
            try
            {
                fh_sel >> temp;
            }
            catch (XmippError &e)
            {
                std::cout << "Sel file: Line " << line_no << " is skipped due to an error\n";
            }
            switch (temp.line_type)
            {
            case SelLine::NOT_ASSIGNED: break; // Line with an error
            case SelLine::DATALINE:
                            if (temp.label != SelLine::DISCARDED)
                                no_imgs++;
                text_line.push_back(temp);
                break;
            case SelLine::COMMENT:
                            text_line.push_back(temp);
                break;
            default:
                break;
            }
            line_no++;
            fh_sel.peek();
        }

        // Close file
        fh_sel.close();
    }

    // Set "pointer" to the beginning of the file
    if (overriding)
        fn_sel = sel_name;
    go_first_ACTIVE();
}

remove()

void SelFile::remove

(

const std::string &

img_name

)

Removes an image from the selection file.

This function searches for an image in the selection file, if it is found then the corresponding line is deleted. If the image is actually being pointed by the current line, then the current line is now the following line.

sel.remove("g1ta0001");

Definition at line 678 of file selfile.cpp.

{
    std::vector<SelLine>::iterator aux = find(img_name);
    std::vector<SelLine>::iterator temp;
    if (aux != text_line.end())
    {
        if (aux == current_line)
        {
            temp = current_line;
            temp++;
        }
        else
            temp = current_line;
        if ((*aux).line_type == SelLine::DATALINE)
            no_imgs--;
        text_line.erase(aux);
        current_line = temp;
    }
}

remove_current()

void SelFile::remove_current

(

)

Removes actual line.

This function removes the current line, either it is a comment or an image. The current line "pointer" is moved to the following line in the file.

sel.remove_current();

Definition at line 699 of file selfile.cpp.

{
    if (current_line != text_line.end())
    {
        std::vector<SelLine>::iterator temp;
        temp = current_line;
        temp++;
        if ((*current_line).line_type == SelLine::DATALINE)
            no_imgs--;
        text_line.erase(current_line);
        current_line = temp;
    }
}

reserve()

void SelFile::reserve ( int  n )

inline

Reserve memory for N entries.

It doesn't matter if entries are comments or data. The current line is set to the beginning of the SelFile

Definition at line 219 of file selfile.h.

    {
        text_line.reserve(n);
        current_line = text_line.begin();
    }

search()

void SelFile::search ( const std::string &  img_name )

inline

Move "pointer" to a certain image filename.

This function searches for an image name within the file, and locate the current line "pointer" pointing to that line. If the image name is not present (it is not the same "not present" and "discarded") in the selection file, then the pointer is moved to the end of the selection file. You can check this situation using eof(). It doesn't matter if the current line "pointer" before the function call is after the line where the image name is, this function makes a search all over the file, regardless the previous situation of the current line "pointer".

sel.search("g1ta0001");

Definition at line 473 of file selfile.h.

    {
        current_line = find(img_name);
    }

set()

void SelFile::set	(	const std::string &	img_name,
		SelLine::Label	label
	)

Set label of an image.

This function searches for an image inside the selection file and sets its label to the given label. If the image is not found in the file, then it is added at the end with the given label. The current line pointer is not modified.

sel.set("g1ta0001", SelLine::ACTIVE);

Definition at line 714 of file selfile.cpp.

{
    SelLine temp;
    std::vector<SelLine>::iterator aux = find(img_name);
    if (aux == text_line.end())
    {
        temp.line_type = SelLine::DATALINE;
        temp.text = img_name;
        temp.label = label;
        text_line.push_back(temp);
        if (label != SelLine::DISCARDED)
            no_imgs++;
    }
    else
    {
        if ((*aux).label != label)
        {
            (*aux).label = label;
            if (label != SelLine::DISCARDED)
                no_imgs++;
        }
    }
}

set_current()

void SelFile::set_current

(

SelLine::Label

label

)

Set the label of the current file.

The same as the previous function but the label is set to the file currently pointed.

Definition at line 739 of file selfile.cpp.

{
    if ((*current_line).label != label)
    {
        (*current_line).label = label;
        if (label != SelLine::DISCARDED)
            no_imgs++;
    }
}

set_current_filename()

void SelFile::set_current_filename

(

const FileName &

fn_new

)

Change current filename.

This function changes the current filename to a new one if it is not a comment. If it is a comment line, nothing is done.

Definition at line 750 of file selfile.cpp.

{
    if ((*current_line).line_type == SelLine::DATALINE)
    {
        (*current_line).text = fn_new;
    }
}

sort_by_filenames()

SelFile SelFile::sort_by_filenames

(

)

Sort images in ascending order.

All images are sorted in ascending order either they are active or discarded. All comments are gathered at the end of the resulting selection file. The current line of the resulting selection file is placed at the beginning of the file.

sel2 = sel1.sort_by_filenames();

Definition at line 807 of file selfile.cpp.

{
    SelFile result(*this);
    sort(result.text_line.begin(), result.text_line.end());
    result.current_line = result.text_line.begin();
    return result;
}

split_in_N()

void SelFile::split_in_N	(	int	N,
		std::vector< SelFile > &	parts
	)

Split this file into N random parts of approximately the same size.

Definition at line 391 of file selfile.cpp.

{
    // Randomize input data
    SelFile  SFtmp, SFrnd;
    SFrnd = *this;
    SFtmp = SFrnd.randomize();
    SFtmp.go_beginning();
    int Nimg = SFtmp.ImgNo();
    SF.clear();

    // Create space for all SelFiles
    for (int n = 0; n < N; n++)
    {
        auto *ptr_SF = new SelFile;
        ptr_SF->reserve(CEIL(Nimg / N));
        SF.push_back(*ptr_SF);
    }

    // Distribute images
    int n = 0;
    for (int i = 0;i < Nimg; i++)
    {
        SF[n].insert(SFtmp.current());
        n = (n + 1) % N;
        if (i < Nimg - 1)
            SFtmp.NextImg();
    }

    // Sort the Selfiles
    for (int n = 0; n < N; n++)
        SF[n] = SF[n].sort_by_filenames();
}

split_in_two()

void SelFile::split_in_two	(	SelFile &	sel1,
		SelFile &	sel2
	)

Split this file into two random halves.

Definition at line 364 of file selfile.cpp.

{
    SelFile  SFtmp;
    SF1 = *this;
    SFtmp = SF1.randomize();
    SF1.clear();
    int N = SFtmp.ImgNo();
    SF1.reserve(N);
    SF2.reserve(N);
    int half = N / 2;
    SFtmp.go_beginning();
    for (int i = 0;i < N; i++)
    {
        if (i < half)
            SF1.insert(SFtmp.current());
        else
            SF2.insert(SFtmp.current());
        if (i < N - 1)
            SFtmp.NextImg();
    }
    SFtmp = SF1.sort_by_filenames();
    SF1 = SFtmp;
    SFtmp = SF2.sort_by_filenames();
    SF2 = SFtmp;
}

write()

void SelFile::write

(

const FileName &

sel_name = ""

)

Write a selection file to disk.

If you give a name then it becomes like a "Save as ..." and from this point on the name of the selection file has changed.

sel.write(); // Save
sel.write("g3t.sel"); // Save as ...

Definition at line 262 of file selfile.cpp.

{
    std::ofstream    fh_sel;
    std::vector<SelLine>::iterator current = text_line.begin();
    std::vector<SelLine>::iterator last    = text_line.end();

    if (strcmp(sel_name.c_str(), "") != 0)
        fn_sel = sel_name;
    // Don't use sel_name=="" because it wastes memory
#ifdef NEVEREVERROB
    if (sel_name.find(IMAGIC_TAG) == 0)
    {
        // Write Imagic selfile
        const FileName hed_fname = sel_name.substr(IMAGIC_TAG_LEN);
        std::vector<Image *> imgs;
        for (; current != last; current++)
        {
            Image *img;
            if (current->Is_data() && (current->get_label() == SelLine::ACTIVE) &&
                (img = Image::LoadImage(current->get_text())))
                imgs.push_back(img);
        }
        if (!ImagicWriteImagicFile(hed_fname, imgs))
            REPORT_ERROR(1553, "Error writing selfile to Imagic file " + sel_name);
        for (std::vector<Image *>::iterator i = imgs.begin(); i != imgs.end(); i++)
            delete(*i);
    }
    else
#endif
    {
        // Write Xmipp selfile
        // Open file
        fh_sel.open(fn_sel.c_str(), std::ios::out);
        if (!fh_sel)
            REPORT_ERROR(ERR_IO_NOWRITE, fn_sel);

        // Read each line and keep it in the list of the SelFile object
        while (current != last)
            fh_sel << *(current++);

        // Close file
        fh_sel.close();
    }
}

Friends And Related Function Documentation

compare

SelFile compare ( SelFile &  SF1, SelFile &  SF2, const int  mode  )

friend

Compare two selection files.

The result is another selection file. At the beginning of it there is information about the number of active and discarded images on both input selection files, about the number of matching files (a file is said to match if it is active in both selection files), the number of active files which are only in the first selection file, and the number of active files which are only in the second. Then goes the list of matching files, the list of files only in SF1 and the list of files only in SF2. There are comments enough to know where things start and finish, and what the numbers are at the beginning. If a file is active in a file and discarded in the other, then it is said to match and it is kept as active, a preceding comment warns of this situation.

if mode<0, output file will contain all information if mode=0, output will be selfile with overlapping images if mode=1, output will be selfile with images only in file 1 if mode=2, output will be selfile with images only in file 2

sel3 = compare(sel1, sel2);

Definition at line 872 of file selfile.cpp.

{
    std::vector<SelLine>     only_in_SF1;
    std::vector<SelLine>     only_in_SF2;
    std::vector<SelLine>     in_both;
    SelFile           result;
    SelLine           temp;
    int               SF1_discarded = 0, SF2_discarded = 0;
    int maxLen = 15;
    char              str[maxLen + 1]; // + terminating null character

    // Search in File 1
    std::vector<SelLine>::iterator current = SF1.text_line.begin();
    std::vector<SelLine>::iterator last    = SF1.text_line.end();
    std::vector<SelLine>::iterator last_SF = SF2.text_line.end();
    std::vector<SelLine>::iterator found;

    while (current != last)
    {
        // Skip if not active
        if ((*current).line_type != SelLine::DATALINE)
        {
            current++;
            continue;
        }
        if ((*current).label == SelLine::DISCARDED)
        {
            SF1_discarded++;
            current++;
            continue;
        }

        // Try to find this archive into Sel File 2
        found = SF2.find((*current).text);
        if (found == last_SF)
            only_in_SF1.push_back(*current);
        else
            if ((*found).label == SelLine::DISCARDED)
                only_in_SF1.push_back(*current);
            else
                in_both.push_back(*current);
        current++;
    }

    // Search in File 2
    current = SF2.text_line.begin();
    last    = SF2.text_line.end();

    while (current != last)
    {
        // Skip if not active
        if ((*current).line_type != SelLine::DATALINE)
        {
            current++;
            continue;
        }
        if ((*current).label == SelLine::DISCARDED)
        {
            SF2_discarded++;
            current++;
            continue;
        }

        // Try to find this archive into Sel File 2
        found = find(in_both, (*current).text);
        if (found != in_both.end())
        {
            current++;
            continue;
        }
        only_in_SF2.push_back(*current);
        current++;
    }

    // Write Statistics
    if (mode < 0)
    {
    temp.line_type = SelLine::COMMENT;
    temp.label = SelLine::DISCARDED;
    temp.text = "# Statistics of comparison";
    result.text_line.push_back(temp);
    temp.text = "# -------------------------------------------------------------";
    result.text_line.push_back(temp);
    snprintf(str, maxLen, "%6d", SF1.no_imgs);
    temp.text = "# File 1: " + SF1.fn_sel + "(VALID: " + str;
    snprintf(str, maxLen, "%6d", SF1_discarded);
    temp.text += (std::string) " DISCARDED: " + str + ")";
    result.text_line.push_back(temp);
    snprintf(str, maxLen, "%6d", SF2.no_imgs);
    temp.text = "# File 2: " + SF2.fn_sel + "(VALID: " + str;
    snprintf(str, maxLen, "%6d", SF2_discarded);
    temp.text += (std::string) " DISCARDED: " + str + ")";
    result.text_line.push_back(temp);
    temp.text = "";
    result.text_line.push_back(temp);
    snprintf(str, maxLen, "%6lu", (unsigned long int)in_both.size());
    temp.text = (std::string)"# Matching Files: " + str;
    result.text_line.push_back(temp);
    snprintf(str, maxLen, "%6lu", (unsigned long int)only_in_SF1.size());
    temp.text = (std::string)"# Only in file 1: " + str;
    result.text_line.push_back(temp);
    snprintf(str, maxLen, "%6lu", (unsigned long int)only_in_SF2.size());
    temp.text = (std::string)"# Only in file 2: " + str;
    result.text_line.push_back(temp);
    temp.text = "# -------------------------------------------------------------";
    result.text_line.push_back(temp);

    // Write files in both
    temp.text = "";
    result.text_line.push_back(temp);
    temp.text = "# Files in both .sel files";
    result.text_line.push_back(temp);
    }
    if (mode<0 || mode==0) 
    {
    current = in_both.begin();
    last    = in_both.end();
    while (current != last)
        result.text_line.push_back(*current++);
    }

    if (mode<0)
    {
    // Write files only in Sel File 1
    temp.text = "";
    result.text_line.push_back(temp);
    temp.text = "# Files only in the first file";
    result.text_line.push_back(temp);
    }
    if (mode<0 || mode==1) 
    {
    current = only_in_SF1.begin();
    last    = only_in_SF1.end();
    while (current != last)
        result.text_line.push_back(*current++);
    }

    if (mode<0)
    {
    // Write files only in Sel File 2
    temp.text = "";
    result.text_line.push_back(temp);
    temp.text = "# Files only in the second file";
    result.text_line.push_back(temp);
    }
    if (mode<0 || mode==2) 
    {
    current = only_in_SF2.begin();
    last    = only_in_SF2.end();
    while (current != last)
        result.text_line.push_back(*current++);
    }
    // Adjust the remaining fields
    if (mode<0) 
    result.no_imgs = in_both.size() + only_in_SF1.size() + only_in_SF2.size();
    else if (mode==0) 
    result.no_imgs = in_both.size();
    else if (mode==1)
    result.no_imgs = only_in_SF1.size();
    else if (mode==2)
    result.no_imgs = only_in_SF2.size();
    result.current_line = result.text_line.begin();

    return result;
}

operator<<

std::ostream& operator<< ( std::ostream &  o, const SelFile &  sel  )

friend

Show a selection file.

Shows all the lines either they are comments, active images or discarded images. A new line is printed at the end.

std::cout << sel;

Definition at line 145 of file selfile.cpp.

{
    std::vector<SelLine>::const_iterator current = SF.text_line.begin();
    std::vector<SelLine>::const_iterator last    = SF.text_line.end();
    while (current != last)
    {
        o << *current;
        current++;
    }
    return o;
}

---

The documentation for this class was generated from the following files:

• xmipp/libraries/interface/selfile.h
• xmipp/libraries/interface/selfile.cpp