include/readFasta.h File Reference

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Functions
void	readFasta_open (char *filename)
int4	readFasta_readSequence ()
void	readFasta_close ()
Variables
char *	readFasta_descriptionBuffer
char *	readFasta_sequenceBuffer
uint4	readFasta_sequenceLength
uint4	readFasta_descriptionLength

---

Function Documentation

void readFasta_close

(

)

Definition at line 159 of file readFasta.c.

References readFasta_descriptionBuffer, readFasta_file, readFasta_lineBuffer, and readFasta_sequenceBuffer.

Referenced by determineDbAlphabetType(), and main().

{
        free(readFasta_lineBuffer);
        free(readFasta_sequenceBuffer);
        free(readFasta_descriptionBuffer);
    fclose(readFasta_file);
}

Here is the caller graph for this function:

void readFasta_open

(

char *

filename

)

Definition at line 23 of file readFasta.c.

References global_malloc(), readFasta_bufferAlloc, readFasta_descriptionBuffer, readFasta_endReached, readFasta_file, readFasta_filename, readFasta_lineBuffer, readFasta_lineLength, readFasta_sequenceBuffer, and readFasta_sequenceBufferAlloc.

Referenced by determineDbAlphabetType(), and main().

{
        readFasta_file = fopen(filename, "r");

    if (readFasta_file == NULL)
    {
        fprintf(stderr, "%s\n", strerror(errno));
                fprintf(stderr, "Error opening file %s for reading\n", filename);
                exit(-1);
    }

    readFasta_filename = filename;

        readFasta_bufferAlloc = 10;
        readFasta_sequenceBufferAlloc = 10;

    // Declare memory for buffers
        readFasta_sequenceBuffer = (char*)global_malloc(sizeof(char) * readFasta_bufferAlloc);
        readFasta_descriptionBuffer = (char*)global_malloc(sizeof(char) * readFasta_bufferAlloc);
        readFasta_lineBuffer = (char*)global_malloc(sizeof(char) * readFasta_bufferAlloc);
    readFasta_lineBuffer[0] = '\0'; readFasta_lineBuffer[1] = '\0';
    readFasta_lineLength = 1;
    readFasta_endReached = 0;
}

Here is the call graph for this function:

Here is the caller graph for this function:

int4 readFasta_readSequence

(

)

Definition at line 49 of file readFasta.c.

References global_realloc(), readFasta_bufferAlloc, readFasta_descriptionBuffer, readFasta_descriptionLength, readFasta_endReached, readFasta_file, readFasta_lineBuffer, readFasta_lineLength, readFasta_sequenceBuffer, readFasta_sequenceBufferAlloc, readFasta_sequenceLength, and uint4.

Referenced by determineDbAlphabetType(), and main().

{
        uint4 pos, pos2;

        // We've reached end of file
        if (readFasta_endReached)
        return 0;

        // Clear sequence buffer
        readFasta_sequenceBuffer[0] = '\0';
    readFasta_sequenceLength = 0;

    // Line buffer contains description for new sequence
        memcpy(readFasta_descriptionBuffer, readFasta_lineBuffer + 1, readFasta_lineLength);
    readFasta_descriptionLength = readFasta_lineLength - 1;

        // Read each line from FASTA file
        while (!readFasta_endReached && fgets(readFasta_lineBuffer, readFasta_bufferAlloc, readFasta_file))
    {
        // If we didn't read an entire line
        readFasta_lineLength = strlen(readFasta_lineBuffer);
        while (readFasta_lineBuffer[readFasta_lineLength - 1] != '\n')
        {
//              printf("Read bit \"%s\"\n", readFasta_lineBuffer);

            // Double size of the buffers and read more
                        readFasta_lineBuffer = (char*)global_realloc(readFasta_lineBuffer,
                                   sizeof(char) * readFasta_bufferAlloc * 2);
                        readFasta_descriptionBuffer = (char*)global_realloc(readFasta_descriptionBuffer,
                                          sizeof(char) * readFasta_bufferAlloc * 2);

                        // Read next line
                        if (!fgets(readFasta_lineBuffer + readFasta_bufferAlloc - 1, readFasta_bufferAlloc, readFasta_file))
                        {
                // If reached end of file
                readFasta_lineLength = strlen(readFasta_lineBuffer);
                readFasta_endReached = 1;
                break;
                        }

            readFasta_bufferAlloc = readFasta_bufferAlloc * 2 - 1;
            readFasta_lineLength = strlen(readFasta_lineBuffer);
        }

        // Remove the trailing \n
        if (readFasta_lineBuffer[readFasta_lineLength - 1] == '\n')
        {
            readFasta_lineLength--;
            readFasta_lineBuffer[readFasta_lineLength] = '\0';
                }

//        printf("Read line \"%s\"\n", readFasta_lineBuffer);

        // If this is a description line
        if (readFasta_lineBuffer[0] == '>')
        {
//              printf("DESCRIPTION!\n");
                        // If the description buffer is not empty
            if (readFasta_descriptionBuffer[0] != '\0')
                        {
                return 1;
            }

//              printf("(First)\n");
            // Else line is description of first sequence
            memcpy(readFasta_descriptionBuffer, readFasta_lineBuffer + 1, readFasta_lineLength);
            readFasta_descriptionLength = readFasta_lineLength - 1;

//            printf("(end copy)\n");
        }
        // Otherwise it is part of a sequence
        else
        {
                // If there is insufficient space in the current sequence buffer
                        if (readFasta_lineLength + readFasta_sequenceLength + 1 > readFasta_sequenceBufferAlloc)
            {
                // Increase size of the buffer
                                readFasta_sequenceBufferAlloc = (readFasta_lineLength + readFasta_sequenceLength + 1) * 2;
                readFasta_sequenceBuffer = (char*)global_realloc(readFasta_sequenceBuffer,
                                           sizeof(char) * readFasta_sequenceBufferAlloc);
                        }

            // Remove non-alphabet characters from the sequence
                        pos = 0; pos2 = 0;
            while (pos < readFasta_lineLength)
            {
                if ((readFasta_lineBuffer[pos] >= 'a' && readFasta_lineBuffer[pos] <= 'z') ||
                    (readFasta_lineBuffer[pos] >= 'A' && readFasta_lineBuffer[pos] <= 'Z'))
                {
                        readFasta_lineBuffer[pos2] = readFasta_lineBuffer[pos];
                    pos2++;
                }
                pos++;
            }
                        readFasta_lineBuffer[pos2] = '\0';
                        readFasta_lineLength = pos2;

            // Append new part of sequence onto the current sequence buffer contents
                        memcpy(readFasta_sequenceBuffer + readFasta_sequenceLength, readFasta_lineBuffer,
                   readFasta_lineLength + 1);
            readFasta_sequenceLength += readFasta_lineLength;
        }
    }

    // End of file. Return contents of last sequence
    readFasta_endReached = 1;
    return 1;
}

Here is the call graph for this function:

Here is the caller graph for this function:

---

Variable Documentation

char* readFasta_descriptionBuffer

Definition at line 4 of file readFasta.h.

Referenced by main(), readFasta_close(), readFasta_open(), and readFasta_readSequence().

uint4 readFasta_descriptionLength

Definition at line 7 of file readFasta.h.

Referenced by main(), and readFasta_readSequence().

char* readFasta_sequenceBuffer

Definition at line 5 of file readFasta.h.

Referenced by determineDbAlphabetType(), main(), readFasta_close(), readFasta_open(), and readFasta_readSequence().

uint4 readFasta_sequenceLength

Definition at line 6 of file readFasta.h.

Referenced by determineDbAlphabetType(), main(), and readFasta_readSequence().

---