Add a -M option to orgasmi

src/buffer.c

@@ -8,10 +8,11 @@
 
 #include "orgasm.h"
 
-buffer_t *newBuffer(size_t maxsize,size_t readSize)
+buffer_t *newBuffer(size_t maxsize,size_t readSize, size_t maxread)
 {
 #define MASK  ((size_t)0xF)
 #define CLEAN ((size_t)~MASK)
+#define MIN(x,y) (((x) > (y)) ? (y):(x))
 
 	buffer_t* rep;
 	size_t rsize=round16(readSize)/4 + 2 * sizeof(uint32_t);
@@ -19,6 +20,9 @@ buffer_t *newBuffer(size_t maxsize,size_t readSize)
 	if (((maxsize-48) / rsize) > (size_t)0xFFFFFFFF)
 		maxsize = (size_t)0xFFFFFFFF * rsize + 48;
 
+	if (maxread>0)
+		maxsize = MIN(maxread*rsize+48,maxsize);
+
 
 	rep = MALLOC(sizeof(buffer_t));
 	rep->arena  = MALLOC(maxsize);

src/buildexpand8bits.c

@@ -36,12 +36,12 @@ int main(int argc, char *argv[])
 		for (i=0; i < 4; i++)
 			nuc[3-i] = (icode >> (i*2))  & 3;
 
-#ifdef LITTLE_END
-		*((uint32_t*)nuc) = *((uint32_t*)nuc) >> 24 | \
-				*((uint32_t*)nuc) << 24 | \
-			  ((*((uint32_t*)nuc) >> 8) & 0x0000FF00) | \
-			  ((*((uint32_t*)nuc) << 8) & 0x00FF0000);
-#endif
+//#ifdef LITTLE_END
+//		*((uint32_t*)nuc) = *((uint32_t*)nuc) >> 24 | \
+//				*((uint32_t*)nuc) << 24 | \
+//			  ((*((uint32_t*)nuc) >> 8) & 0x0000FF00) | \
+//			  ((*((uint32_t*)nuc) << 8) & 0x00FF0000);
+//#endif
 
 		if (code < 0xFF)
 			fprintf(stdout,"    0x%08XUL,  // %04X\n", *((uint32_t*)nuc),code);

src/buildindex.c

@@ -9,6 +9,7 @@
 buffer_t *buildIndex(const char *forwardFileName, const char* reverseFileName,
 					    const char *indexname,
 		 	 	 	 	uint32_t minword,
+		 	 	 	 	size_t maxread,
 					    size_t maxbuffersize
 					   )
 {
@@ -32,7 +33,7 @@ buffer_t *buildIndex(const char *forwardFileName, const char* reverseFileName,
 	if (reverse == NULL)
 	FATALERROR("Cannot open file %s",reverseFileName)
 
-	reads = loadPairedFastq(forward,reverse,maxbuffersize);
+	reads = loadPairedFastq(forward,reverse,maxread,maxbuffersize);
 
 	fclose(forward);
 	fclose(reverse);

src/fastq.c

@@ -140,6 +140,7 @@ uint8_t checkACGT(const char* src, size_t size)
 
 buffer_t * loadPairedFastq(FILE* forward,
  			 		       FILE* reverse,
+ 			 	 	 	   size_t maxread,
 					       size_t maxbuffersize
 		                  )
 {
@@ -178,7 +179,7 @@ buffer_t * loadPairedFastq(FILE* forward,
 
 	readLength = readPairedFastq(reads,forward,reverse,buffersize);
 
-	seqbuffer  = newBuffer(maxbuffersize,readLength);
+	seqbuffer  = newBuffer(maxbuffersize,readLength,maxread);
 	encodedseqs= (uint32_t*)(seqbuffer->records);
 
 	while (readLength && (seqbuffer->readCount+2) <= seqbuffer->maxrecord)

src/load.c

@@ -47,7 +47,7 @@ buffer_t *loadIndexedReads(const char *indexname)
 	maxsize = (tmp.recordSize  + \
 			   2 * sizeof(uint32_t)) * tmp.readCount + 48;
 
-	reads = newBuffer(maxsize,tmp.readSize);
+	reads = newBuffer(maxsize,tmp.readSize,0);
 
 	reads->readCount = tmp.readCount;
 

src/orgasm.h

@@ -105,16 +105,18 @@ char *decodeSequence(pnuc buffer, uint32_t begin, int32_t length, char *dest);
 
 buffer_t *loadPairedFastq(FILE* forward,
 	       	   	   	      FILE* reverse,
+  		 	 	 	 	  size_t maxread,
 	       	   	   	      size_t maxbuffersize
 		                 );
 
 buffer_t *buildIndex(const char *forwardFileName, const char* reverseFileName,
 					    const char *indexname,
 		 	 	 	 	uint32_t minword,
+		 	 	 	 	size_t maxread,
 					    size_t maxbuffersize
 					   );
 
-buffer_t *newBuffer(size_t maxsize,size_t readSize);
+buffer_t *newBuffer(size_t maxsize,size_t readSize,size_t maxread);
 void freeBuffer(buffer_t *buffer);
 void swapOrder(buffer_t *buffer);
 

src/orgasmi.c

@@ -6,6 +6,7 @@
  */
 
 #include <stdio.h>
+#include <stdlib.h>
 #include <string.h>
 #include <getopt.h>
 
@@ -39,6 +40,7 @@ static void PrintHelp()
         PP      "            <index>.orx : contains reverse index\n");
         PP      "            <index>.opx : contains read pairing data\n\n");
         PP      "        The assembler will need all these file to process assembling\n\n");
+        PP      "-M    : If specified the count in million of reads to index");
         PP      "-h    : [H]elp - print <this> help\n\n");
         PP      "\n");
         PP      "------------------------------------------\n");
@@ -73,6 +75,7 @@ int main(int argc, char *argv[])
 {
 	buffer_t* reads;
 	char* indexfiles=NULL;
+	size_t toread=0;
 
 	int           carg;
 	int       errflag=0;
@@ -85,7 +88,7 @@ int main(int argc, char *argv[])
 
 
 
-while ((carg = getopt(argc, argv, "ho:")) != -1) {
+while ((carg = getopt(argc, argv, "hM:o:")) != -1) {
 
  switch (carg) {
                             /* -------------------- */
@@ -102,7 +105,13 @@ while ((carg = getopt(argc, argv, "ho:")) != -1) {
     	strcpy(indexfiles,(const char*)optarg);
     	break;
 
-	break;
+    					   /* -------------------- */
+    case 'M':              /* index name           */
+    					   /* -------------------- */
+    	toread = atoll((const char*)optarg) * 1000000;
+    	break;
+
+break;
 
 	case '?':               /* bad option           */
                             /* -------------------- */
@@ -138,7 +147,7 @@ if (errflag)
 
 #define MAXREAD 2000000000
 
-	reads = buildIndex(forwardFileName,reverseFileName,indexfiles,95,MAXREAD);
+	reads = buildIndex(forwardFileName,reverseFileName,indexfiles,95,toread,MAXREAD);
 
 	freeBuffer(reads);