InputFile Class Reference

Class for easily reading/writing files without having to worry about file type (uncompressed, gzip, bgzf) when reading. More...

#include <InputFile.h>

Collaboration diagram for InputFile:

[legend]

List of all members.

Public Types
enum	ifileCompression { DEFAULT, UNCOMPRESSED, GZIP, BGZF }
	Compression to use when writing a file & decompression used when reading a file from stdin. More...
Public Member Functions
	InputFile ()
	Default constructor.
	~InputFile ()
	Destructor.
	InputFile (const char *filename, const char *mode, InputFile::ifileCompression compressionMode=InputFile::DEFAULT)
	Constructor for opening a file.
void	bufferReads (unsigned int bufferSize=DEFAULT_BUFFER_SIZE)
	Set the buffer size for reading from files so that bufferSize bytes are read at a time and stored until accessed by another read call.
void	disableBuffering ()
	Disable read buffering.
int	ifclose ()
	Close the file.
int	ifread (void *buffer, unsigned int size)
	Read size bytes from the file into the buffer.
int	ifgetc ()
	Get a character from the file.
void	ifrewind ()
	Reset to the beginning of the file.
int	ifeof ()
	Check to see if we have reached the EOF.
unsigned int	ifwrite (const void *buffer, unsigned int size)
	Write the specified buffer into the file.
bool	isOpen ()
	Returns whether or not the file was successfully opened.
int64_t	iftell ()
	Get current position in the file.
bool	ifseek (int64_t offset, int origin)
	Seek to the specified offset from the origin.
const char *	getFileName () const
	Get the filename that is currently opened.
void	setAttemptRecovery (bool flag=false)
	Enable (default) or disable recovery.
bool	attemptRecoverySync (bool(*checkSignature)(void *data), int length)
bool	openFile (const char *filename, const char *mode, InputFile::ifileCompression compressionMode)
Protected Member Functions
int	readFromFile (void *buffer, unsigned int size)
Protected Attributes
FileType *	myFileTypePtr
unsigned int	myAllocatedBufferSize
char *	myFileBuffer
int	myBufferIndex
int	myCurrentBufferSize
std::string	myFileName
Static Protected Attributes
static const unsigned int	DEFAULT_BUFFER_SIZE = 1048576

---

Detailed Description

Class for easily reading/writing files without having to worry about file type (uncompressed, gzip, bgzf) when reading.

Definition at line 36 of file InputFile.h.

---

Member Enumeration Documentation

enum InputFile::ifileCompression

Compression to use when writing a file & decompression used when reading a file from stdin.

Any other read checks the file to determine how to uncompress it.

Enumerator:

DEFAULT	Check the extension, if it is ".gz", treat as gzip, otherwise treat it as UNCOMPRESSED.
UNCOMPRESSED	uncompressed file.
GZIP	gzip file.
BGZF	bgzf file.

Definition at line 44 of file InputFile.h.

                          {
        DEFAULT,  ///< Check the extension, if it is ".gz", treat as gzip, otherwise treat it as UNCOMPRESSED.
        UNCOMPRESSED,  ///< uncompressed file.
        GZIP,  ///< gzip file.
        BGZF ///< bgzf file.
    };

---

Constructor & Destructor Documentation

InputFile::InputFile	(	const char *	filename,
		const char *	mode,
		InputFile::ifileCompression	compressionMode = InputFile::DEFAULT
	)

Constructor for opening a file.

Parameters:

	filename	file to open
	mode	same format as fopen: "r" for read & "w" for write.
	compressionMode	set the type of file to open for writing or for reading from stdin (when reading files, the compression type is determined by reading the file).

Definition at line 28 of file InputFile.cpp.

{
    // XXX duplicate code
    myAttemptRecovery = false;
    myFileTypePtr = NULL;
    myBufferIndex = 0;
    myCurrentBufferSize = 0;
    myAllocatedBufferSize = DEFAULT_BUFFER_SIZE;
    myFileBuffer = new char[myAllocatedBufferSize];
    myFileName.clear();

    openFile(filename, mode, compressionMode);
}

---

Member Function Documentation

void InputFile::bufferReads

(

unsigned int

bufferSize = DEFAULT_BUFFER_SIZE

)

[inline]

Set the buffer size for reading from files so that bufferSize bytes are read at a time and stored until accessed by another read call.

This improves performance over reading the file small bits at a time. Buffering reads disables the tell call for bgzf files. Any previous values in the buffer will be deleted.

Parameters:

bufferSize

number of bytes to read/buffer at a time, default buffer size is 1048576, and turn off read buffering by setting bufferSize = 1;

Definition at line 84 of file InputFile.h.

Referenced by disableBuffering().

    {
        // If the buffer size is the same, do nothing.
        if(bufferSize == myAllocatedBufferSize)
        {
            return;
        }
        // Delete the previous buffer.
        if(myFileBuffer != NULL)
        {
            delete[] myFileBuffer;
        }
        myBufferIndex = 0;
        myCurrentBufferSize = 0;
        // The buffer size must be at least 1 so one character can be
        // read and ifgetc can just assume reading into the buffer.
        if(bufferSize < 1)
        {
            bufferSize = 1;
        }
        myFileBuffer = new char[bufferSize];
        myAllocatedBufferSize = bufferSize;

        if(myFileTypePtr != NULL)
        {
            if(bufferSize == 1)
            {
                myFileTypePtr->setBuffered(false);
            }
            else
            {
                myFileTypePtr->setBuffered(true);
            }
        }
    }

const char* InputFile::getFileName

(

)

const [inline]

Get the filename that is currently opened.

Returns:

filename associated with this class

Definition at line 405 of file InputFile.h.

Referenced by SamFile::ReadBamIndex().

    {
        return(myFileName.c_str());
    }

int InputFile::ifclose

(

)

[inline]

Close the file.

Returns:

status of the close (0 is success).

Definition at line 134 of file InputFile.h.

Referenced by ifclose().

    {
        if (myFileTypePtr == NULL)
        {
            return EOF;
        }
        int result = myFileTypePtr->close();
        delete myFileTypePtr;
        myFileTypePtr = NULL;
        myFileName.clear();
        return result;
    }

int InputFile::ifeof

(

)

[inline]

Check to see if we have reached the EOF.

Returns:

0 if not EOF, any other value means EOF.

Definition at line 318 of file InputFile.h.

Referenced by ifeof().

    {
        // Not EOF if we are not at the end of the buffer.
        if (myBufferIndex < myCurrentBufferSize)
        {
            // There are still available bytes in the buffer, so NOT EOF.
            return false;
        }
        else
        {
            if (myFileTypePtr == NULL)
            {
                // No myFileTypePtr, so not eof (return 0).
                return 0;
            }
            // exhausted our buffer, so check the file for eof.
            return myFileTypePtr->eof();
        }
    }

int InputFile::ifgetc

(

)

[inline]

Get a character from the file.

Read a character from the internal buffer, or if the end of the buffer has been reached, read from the file into the buffer and return index 0.

Returns:

character that was read or EOF.

Definition at line 282 of file InputFile.h.

Referenced by ifgetc(), and operator>>().

    {
        if (myBufferIndex >= myCurrentBufferSize)
        {
            // at the last index, read a new buffer.
            myCurrentBufferSize = readFromFile(myFileBuffer, myAllocatedBufferSize);
            myBufferIndex = 0;
        }
        // If the buffer index is still greater than or equal to the
        // myCurrentBufferSize, then we failed to read the file - return EOF.
        if (myBufferIndex >= myCurrentBufferSize)
        {
            return(EOF);
        }
        return(myFileBuffer[myBufferIndex++]);
    }

int InputFile::ifread	(	void *	buffer,
		unsigned int	size
	)			[inline]

Read size bytes from the file into the buffer.

Parameters:

	buffer	pointer to memory at least size bytes big to write the data into.
	size	number of bytes to be read

Returns:

number of bytes read, if it is not equal to size, there was either an error or the end of the file was reached, use ifeof to determine which case it was.

Definition at line 154 of file InputFile.h.

Referenced by ifread().

    {
        // There are 2 cases:
        //  1) There are already size available bytes in buffer.
        //  2) There are not size bytes in buffer.

        // Determine the number of available bytes in the buffer.
        unsigned int availableBytes = myCurrentBufferSize - myBufferIndex;
        int returnSize = 0;

        // Case 1: There are already size available bytes in buffer.
        if (size <= availableBytes)
        {
            //   Just copy from the buffer, increment the index and return.
            memcpy(buffer, myFileBuffer+myBufferIndex, size);
            // Increment the buffer index.
            myBufferIndex += size;
            returnSize = size;
        }
        // Case 2: There are not size bytes in buffer.
        else
        {
            // Check to see if there are some bytes in the buffer.
            if (availableBytes > 0)
            {
                // Size > availableBytes > 0
                // Copy the available bytes into the buffer.
                memcpy(buffer, myFileBuffer+myBufferIndex, availableBytes);
            }
            // So far availableBytes have been copied into the read buffer.
            returnSize = availableBytes;
            // Increment myBufferIndex  by what was read.
            myBufferIndex += availableBytes;

            unsigned int remainingSize = size - availableBytes;

            // Check if the remaining size is more or less than the
            // max buffer size.
            if(remainingSize < myAllocatedBufferSize)
            {
                // the remaining size is not the full buffer, but read
                //  a full buffer worth of data anyway.
                myCurrentBufferSize =
                    readFromFile(myFileBuffer, myAllocatedBufferSize);

                // Check for an error.
                if(myCurrentBufferSize <= 0)
                {
                    // No more data was successfully read, so check to see
                    // if any data was copied to the return buffer at all.
                    if( returnSize == 0)
                    {
                        // No data has been copied at all into the
                        // return read buffer, so just return the value
                        // returned from readFromFile.
                        returnSize = myCurrentBufferSize;
                        // Otherwise, returnSize is already set to the
                        // available bytes that was already copied (so no
                        // else statement is needed).
                    }
                    // Set myBufferIndex & myCurrentBufferSize to 0.
                    myCurrentBufferSize = 0;
                    myBufferIndex = 0;
                }
                else
                {
                    // Successfully read more data.
                    // Check to see how much was copied.
                    int copySize = remainingSize;
                    if(copySize > myCurrentBufferSize)
                    {
                        // Not the entire requested amount was read
                        // (either from EOF or there was a partial read due to
                        // an error), so set the copySize to what was read.
                        copySize = myCurrentBufferSize;
                    }

                    // Now copy the rest of the bytes into the buffer.
                    memcpy((char*)buffer+availableBytes, 
                           myFileBuffer, copySize);

                    // set the buffer index to the location after what we are
                    // returning as read.
                    myBufferIndex = copySize;
                
                    returnSize += copySize;
                }
            }
            else
            {
                // More remaining to be read than the max buffer size, so just
                // read directly into the output buffer.
                int readSize = readFromFile((char*)buffer + availableBytes,
                                            remainingSize);

                // Already used the buffer, so "clear" it.
                myCurrentBufferSize = 0;
                myBufferIndex = 0;
                if(readSize <= 0)
                {
                    // No more data was successfully read, so check to see
                    // if any data was copied to the return buffer at all.
                    if(returnSize == 0)
                    {
                        // No data has been copied at all into the
                        // return read buffer, so just return the value
                        // returned from readFromFile.
                        returnSize = readSize;
                        // Otherwise, returnSize is already set to the
                        // available bytes that was already copied (so no
                        // else statement is needed).
                    }
                }
                else
                {
                    // More data was read, so increment the return count.
                    returnSize += readSize;
                }
            }
        }
        return(returnSize);
    }

bool InputFile::ifseek	(	int64_t	offset,
		int	origin
	)			[inline]

Seek to the specified offset from the origin.

Parameters:

	offset	offset into the file to move to (must be from a tell call)
	origin	can be any of the following: Note: not all are valid for all filetypes. SEEK_SET - Beginning of file SEEK_CUR - Current position of the file pointer SEEK_END - End of file

Returns:

true on successful seek and false on a failed seek.

Definition at line 389 of file InputFile.h.

Referenced by ifseek().

    {
        if (myFileTypePtr == NULL)
        {
            // No myFileTypePtr, so return false - could not seek.
            return false;
        }
        // TODO - may be able to seek within the buffer if applicable.
        // Reset buffering since a seek is being done.
        myBufferIndex = 0;
        myCurrentBufferSize = 0;
        return myFileTypePtr->seek(offset, origin);
    }

int64_t InputFile::iftell

(

)

[inline]

Get current position in the file.

Returns:

current position in the file, -1 indicates an error.

Definition at line 368 of file InputFile.h.

Referenced by iftell().

    {
        if (myFileTypePtr == NULL)
        {
            // No myFileTypePtr, so return false - could not seek.
            return -1;
        }
        int64_t pos = myFileTypePtr->tell();
        pos -= (myCurrentBufferSize - myBufferIndex);
        return(pos);
    }

unsigned int InputFile::ifwrite	(	const void *	buffer,
		unsigned int	size
	)			[inline]

Write the specified buffer into the file.

Parameters:

	buffer	buffer containing size bytes to write to the file.
	size	number of bytes to write

Returns:

number of bytes written We do not buffer the write call, so just leave this as normal.

Definition at line 343 of file InputFile.h.

Referenced by ifwrite().

    {
        if (myFileTypePtr == NULL)
        {
            // No myFileTypePtr, so return 0 - nothing written.
            return 0;
        }
        return myFileTypePtr->write(buffer, size);
    }

bool InputFile::isOpen

(

)

[inline]

Returns whether or not the file was successfully opened.

Returns:

true if the file is open, false if not.

Definition at line 355 of file InputFile.h.

Referenced by ifopen(), SamFile::IsOpen(), GlfHeader::read(), SamRecord::setBufferFromFile(), GlfHeader::write(), and SamRecord::writeRecordBuffer().

    {
        // It is open if the myFileTypePtr is set and says it is open.
        if ((myFileTypePtr != NULL) && myFileTypePtr->isOpen())
        {
            return true;
        }
        // File was not successfully opened.
        return false;
    }

void InputFile::setAttemptRecovery

(

bool

flag = false

)

[inline]

Enable (default) or disable recovery.

When true, we can attach a myFileTypePtr that implements a recovery capable decompressor. This requires that the caller be able to catch the exception XXX "blah blah blah".

Definition at line 417 of file InputFile.h.

Referenced by SamFile::OpenForRead().

    {
        myAttemptRecovery = flag;
    }

---

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

• general/InputFile.h
• general/InputFile.cpp