]> git.llucax.com Git - z.facultad/75.06/jacu.git/blobdiff - src/statichuff/statichuff.c
Se quita repeticion de pagesize en cada bloque comprimido, grabandose una sola vez...
[z.facultad/75.06/jacu.git] / src / statichuff / statichuff.c
index 32b3e8d7b2670460b318aad41aca54ca2b80b2d7..000f0bab7f1c0820f1555f75e420c2d172e19d32 100644 (file)
@@ -1,7 +1,7 @@
 
 #include "statichuff.h"
-#include "../vfile/vfile.h"
 #include <stdlib.h>
+#include <string.h>
 
 void putbit(char bit, char restart, char flush, VFILE *fp)
 {
@@ -70,16 +70,21 @@ int shuff_scanfreq_chunk(HUFF_STATE *chunkshuff, char* chunk, int chunksize)
        int i = 0;
        unsigned char symbol = 0;       
                
-       /* Contamos las frecuencias del chunk*/ 
-       for (i = 0; i < chunksize; ++i) {                               
-               symbol = chunk[i];              
-               chunkshuff->freqtable[symbol] += 1;
-               chunkshuff->sumfreq += 1;
+       /* Contamos las frecuencias del chunk a menos que se use un canonico */ 
+       if (!chunkshuff->canonic) {
+               for (i = 0; i < chunksize; ++i) {                               
+                       symbol = chunk[i];              
+                       chunkshuff->freqtable[symbol] += 1;
+                       chunkshuff->sumfreq += 1;
                                
-               /* Si llegue al tope de freq acumulada, halve em */
-               if (chunkshuff->sumfreq == 14930352) 
-                       chunkshuff->sumfreq = shuff_rescalefreq(chunkshuff->freqtable);
-       }               
+                       /* Si llegue al tope de freq acumulada, halve em */
+                       if (chunkshuff->sumfreq == 14930352) 
+                               chunkshuff->sumfreq = shuff_rescalefreq(chunkshuff->freqtable);
+               }
+       }
+       
+       /* Dumpeamos el chunk en el temporal homero */
+       fwrite(chunk,chunksize,1,chunkshuff->coderfp);
                
        return 1;
 }
@@ -231,7 +236,8 @@ int shuff_encode_symbols(HUFF_STATE *shuff, SHUFFCODE *ctable)
        char bit;
        SHUFFCODE symbolcode;
                
-       /* Abrimos el file */
+       /* Abrimos el source y el destino */
+       if (shuff->coderfp != NULL) fclose(shuff->coderfp); /* close bychunk temp file */
        if ((fpsource = fopen(shuff->sourcefile,"r")) == NULL) return 0;
        if ((fpdest = vfopen(shuff->targetfile,"w",shuff->volsize)) == NULL) return 0;
                
@@ -263,101 +269,21 @@ int shuff_encode_symbols(HUFF_STATE *shuff, SHUFFCODE *ctable)
        return 1;       
 }
 
-int shuff_encode_chunk(HUFF_STATE *chunkshuff, char *chunk, int chunksize, int process)
-{
-       int i;
-       
-       if (!process) {
-               /* Scaneo las frecuencias de este chunk */
-               shuff_scanfreq_chunk(chunkshuff,chunk,chunksize);
-       }
-       else {          
-               /* Ya tengo la tabla y el archivo temporal, paso al huffman by file */          
-               return 0;
-       }
-       
-       return 1; 
-}
-
-HUFF_STATE *shuff_init_static_byfile(char *inputfile, char *outputfile, long volsize)
-{
-       /* Locals */
-       HUFF_STATE *fshuff = (HUFF_STATE*)malloc(sizeof(HUFF_STATE));                   
-       int i;
-       
-       /* Inicializo la estructura para trabajar con Huff Static by File */
-       fshuff->tmpfp = NULL;
-       fshuff->sourcefile = (char*)malloc(sizeof(char)*(strlen(inputfile)+1));
-       fshuff->targetfile = (char*)malloc(sizeof(char)*(strlen(outputfile)+1));
-       strcpy(fshuff->sourcefile,inputfile);   
-       strcpy(fshuff->targetfile,outputfile);
-       fshuff->preloadfreq = 0;
-       fshuff->volsize = volsize;
-       fshuff->sumfreq = 0;
-       fshuff->freqtable = (t_freq*)malloc(sizeof(t_freq)*256);
-       for (i = 0; i < 256; ++i) fshuff->freqtable[i] = 0;             
-       
-       return fshuff;
-}
-
-HUFF_STATE *shuff_init_static_bychunk(char *outputfile, long volsize)
-{
-       /* Locals */
-       HUFF_STATE *cshuff = (HUFF_STATE*)malloc(sizeof(HUFF_STATE));           
-       FILE *fp;
-       int i;
-       
-       /* Inicializo la estructura para trabajar con Huff Static by Chunks */
-       cshuff->sourcefile = (char*)malloc(sizeof(char)*(strlen(outputfile)+2));
-       cshuff->targetfile = (char*)malloc(sizeof(char)*(strlen(outputfile)+1));
-       strcpy(cshuff->targetfile,outputfile);  
-       strcpy(cshuff->sourcefile,outputfile);
-       strcat(cshuff->sourcefile,"~"); 
-       cshuff->volsize = volsize;
-       cshuff->preloadfreq = 1;        
-       cshuff->sumfreq = 0;
-       cshuff->freqtable = (t_freq*)malloc(sizeof(t_freq)*256);
-       for (i = 0; i < 256; ++i) cshuff->freqtable[i] = 0;             
-       
-       /* Abrimos un archivo temporal para ir tirando los chunks */    
-       if ((cshuff->tmpfp = fopen(cshuff->sourcefile,"w")) == NULL) return NULL;       
-
-       return cshuff;
-}
-
-void shuff_deinit_static_byfile(HUFF_STATE *fshuff)
-{
-       /* Libero mallocs y cierro archivos */
-       free(fshuff->freqtable);
-       free(fshuff->sourcefile);
-       free(fshuff->targetfile);               
-}
-
-void shuff_deinit_static_bychunk(HUFF_STATE *cshuff)
-{
-       /* Libero mallocs y cierro archivos */
-       free(cshuff->freqtable);
-       free(cshuff->sourcefile);
-       free(cshuff->targetfile);
-       fclose(cshuff->tmpfp);
-}
-
 int shuff_encode_file(HUFF_STATE *shuff)
 {
-       /* Locals */    
-       SHUFFNODE *codetree;
+       /* Locals */            
        SHUFFCODE *codetable = (SHUFFCODE*)malloc(sizeof(SHUFFCODE)*256);
-       int i;
        
        /* Veo si debo armar una freqtable o si esta preloaded */
-       if (!shuff->preloadfreq) if (!shuff_scanfreq(shuff->sourcefile,shuff->freqtable)) return 0;
+       if ((!shuff->canonic) && (!shuff->bychunk)) 
+               if (!shuff_scanfreq(shuff->sourcefile,shuff->freqtable)) return 0;
        
        /* Genero el arbol de huffman */
-       codetree = shuff_buildtree(shuff->freqtable);
+       shuff->codetree = shuff_buildtree(shuff->freqtable);
 
        /* Armo la tabla de codigos prefijos para el encoder */
        shuff_zerocodes(codetable);
-       shuff_buildcodes(codetable,codetree,0,0);
+       shuff_buildcodes(codetable,shuff->codetree,0,0);
        /*shuff_printcodes(codetable,shuff->freqtable);*/
 
        /* Encodeo byte per byte */
@@ -366,8 +292,6 @@ int shuff_encode_file(HUFF_STATE *shuff)
        /* Free up memory baby yeah */  
        free(codetable);
        
-       /* Destruyo el arbol recursivamente TODO */
-       
        return 1;
 }
 
@@ -388,45 +312,209 @@ SHUFFNODE *shuff_decode_symbols(SHUFFNODE *entrynode, unsigned long int buffer,
        else return shuff_decode_symbols(entrynode->rchild,buffer,bitsleft,symbol);
 }
 
+int shuff_decode_chunk(HUFF_STATE *shuff, char *chunk, int chunksize, int *decodedbytes)
+{
+       SHUFFNODE *currnode = shuff->codetree;  
+       unsigned short int decoded_symbol;      
+       *decodedbytes = 0;
+       
+       while (!vfeof(shuff->decoderfp) && (shuff->bytesleft > 0) && (*decodedbytes < chunksize)) {
+               
+               /* Leo un buffer de 32 bits si es que quedo vacio el anterior */
+               if (shuff->bitsleft == 0) {
+                       if (vfread(&(shuff->codebuffer),sizeof(unsigned long int),1,shuff->decoderfp) != 1) continue;
+                       shuff->bitsleft = sizeof(unsigned long int) * 8;
+               }
+               
+               /* Proceso el buffer sacando simbolos till se me agote el buffer, file o chunk */
+               while ((shuff->bitsleft > 0) && (shuff->bytesleft > 0) && (*decodedbytes < chunksize)) {        
+                       currnode = shuff_decode_symbols(currnode,shuff->codebuffer,&(shuff->bitsleft),&decoded_symbol);
+                       /* Si obtuve un symbolo valido lo emito*/
+                       if (decoded_symbol != 256) {                            
+                               chunk[(*decodedbytes)++] = decoded_symbol;
+                               currnode = shuff->codetree;                             
+                               --(shuff->bytesleft);
+                       }                                               
+               }               
+       }
+       
+       if (shuff->bytesleft == 0) return 0;
+       else return 1;
+}
+
 int shuff_decode_file(HUFF_STATE *shuff)
 {      
-       SHUFFNODE *codetree,*currnode;  
-       unsigned long int bytesleft,codebuffer;
-       VFILE *fpsource;
+       SHUFFNODE *currnode;    
+       unsigned long int codebuffer;   
        FILE *fpdest;
        unsigned short int decoded_symbol;
-       int bitsleft,freqcount = 0;     
+       int bitsleft;   
        
-       /* Levanto cuantos bytes decodeo y la freq table */
-       if ((fpsource = vfopen(shuff->sourcefile,"r",0)) == NULL) return 0;
-       if ((fpdest = fopen(shuff->targetfile,"w")) == NULL) return 0;
-       vfread(&bytesleft,sizeof(unsigned long int),1,fpsource);
-       vfread(shuff->freqtable,sizeof(unsigned long int),256,fpsource);                
-       codetree = shuff_buildtree(shuff->freqtable);
-       currnode = codetree;
+       /* Comienzo a decodificar, pues la tabla ya la levante en el decinit */
+       if ((fpdest = fopen(shuff->targetfile,"w")) == NULL) return 0;  
+       currnode = shuff->codetree;
        
-       while (!vfeof(fpsource) && (bytesleft > 0)) {
+       while (!vfeof(shuff->decoderfp) && (shuff->bytesleft > 0)) {
                
                /* Leo un buffer de 32 bits */
-               if (vfread(&codebuffer,sizeof(unsigned long int),1,fpsource) != 1) continue;
+               if (vfread(&codebuffer,sizeof(unsigned long int),1,shuff->decoderfp) != 1) continue;
                bitsleft = sizeof(unsigned long int) * 8;
                
                /* Proceso el buffer sacando simbolos hasta que se me agote */
-               while ((bitsleft > 0) && (bytesleft > 0)) {     
+               while ((bitsleft > 0) && (shuff->bytesleft > 0)) {      
                        currnode = shuff_decode_symbols(currnode,codebuffer,&bitsleft,&decoded_symbol);
                        /* Si obtuve un symbolo valido lo emito*/
                        if (decoded_symbol != 256) {
                                fputc(decoded_symbol,fpdest);
-                               currnode = codetree;
-                               --bytesleft;                            
+                               currnode = shuff->codetree;
+                               --(shuff->bytesleft);
                        }                                               
                }               
        }
-               
-       vfclose(fpsource);
-       fclose(fpdest);
-       
-       /* Libero el arbol recursivamente TODO */       
+                       
+       /* Close destination */
+       fclose(fpdest); 
        
        return 1;
 }
+
+HUFF_STATE *shuff_init_decoder(char *inputfile, char *outputfile)
+{
+       /* Locals */
+       HUFF_STATE *shuff = (HUFF_STATE*)malloc(sizeof(HUFF_STATE));                    
+       shuff->freqtable = (t_freq*)malloc(sizeof(t_freq)*256); 
+       
+       /* Init fields */
+       shuff->codebuffer = 0;
+       shuff->bitsleft = 0;
+       shuff->coderfp = NULL;
+       shuff->targetfile = NULL;
+       shuff->sourcefile = (char*)malloc(sizeof(char)*(strlen(inputfile)+1));
+       strcpy(shuff->sourcefile,inputfile);
+       if (outputfile != NULL) {
+               shuff->targetfile = (char*)malloc(sizeof(char)*(strlen(outputfile)+1));
+               strcpy(shuff->targetfile,outputfile);
+       }       
+       
+       /* Levanto cuantos bytes debo decodificar y la freqtable */
+       if ((shuff->decoderfp = vfopen(shuff->sourcefile,"r",0)) == NULL) return NULL;  
+       vfread(&(shuff->bytesleft),sizeof(unsigned long int),1,shuff->decoderfp);
+       vfread(shuff->freqtable,sizeof(t_freq),256,shuff->decoderfp);           
+       /* Armo el arbol de huffman que uso para decodificar */
+       shuff->codetree = shuff_buildtree(shuff->freqtable);
+       
+       return shuff;
+}
+
+HUFF_STATE *shuff_init_encoder_byfile(char *inputfile, char *outputfile, long volsize)
+{
+       /* Locals */
+       HUFF_STATE *fshuff = (HUFF_STATE*)malloc(sizeof(HUFF_STATE));                   
+       int i;
+       
+       /* Inicializo la estructura para trabajar con Huff Static by File */
+       fshuff->coderfp = NULL; 
+       fshuff->decoderfp = NULL;       
+       fshuff->sourcefile = (char*)malloc(sizeof(char)*(strlen(inputfile)+1));
+       fshuff->targetfile = (char*)malloc(sizeof(char)*(strlen(outputfile)+1));
+       strcpy(fshuff->sourcefile,inputfile);   
+       strcpy(fshuff->targetfile,outputfile);
+       fshuff->volsize = volsize;
+       fshuff->bychunk = 0;
+       fshuff->canonic = 0;    
+       fshuff->freqtable = (t_freq*)malloc(sizeof(t_freq)*256);
+       for (i = 0; i < 256; ++i) fshuff->freqtable[i] = 0;     
+       fshuff->sumfreq = 0;            
+       fshuff->codetree = NULL;
+       
+       return fshuff;
+}
+
+HUFF_STATE *shuff_init_encoder_bychunk(char *outputfile, long volsize)
+{
+       /* Locals */
+       HUFF_STATE *cshuff = (HUFF_STATE*)malloc(sizeof(HUFF_STATE));                   
+       int i;
+       
+       /* Inicializo la estructura para trabajar con Huff Static by Chunks */          
+       cshuff->decoderfp = NULL;
+       cshuff->sourcefile = (char*)malloc(sizeof(char)*(strlen(outputfile)+2));
+       cshuff->targetfile = (char*)malloc(sizeof(char)*(strlen(outputfile)+1));
+       strcpy(cshuff->targetfile,outputfile);  
+       strcpy(cshuff->sourcefile,outputfile);
+       strcat(cshuff->sourcefile,"~"); 
+       cshuff->volsize = volsize;
+       cshuff->bychunk = 1;
+       cshuff->canonic = 0;
+       cshuff->freqtable = (t_freq*)malloc(sizeof(t_freq)*256);        
+       for (i = 0; i < 256; ++i) cshuff->freqtable[i] = 0;     
+       cshuff->sumfreq = 0;    
+       cshuff->codetree = NULL;        
+       
+       /* Abrimos un archivo temporal para ir tirando los chunks */    
+       if ((cshuff->coderfp = fopen(cshuff->sourcefile,"w")) == NULL) return NULL;     
+
+       return cshuff;
+}
+
+int shuff_loadmodel(HUFF_STATE *shuff, char *modelfile) {
+
+       FILE *fp;
+       
+       if ((shuff) && (shuff->freqtable) && (modelfile)) {
+               /* Cargo el modelo de disco */
+               if ((fp = fopen(modelfile,"r")) == NULL) return 0;
+               if (fread(shuff->freqtable,sizeof(t_freq),256,fp) != 256) return 0;
+               shuff->canonic = 1;
+               if (fp) fclose(fp);             
+               return 1;
+       }       
+       return 0;       
+}
+
+int shuff_savemodel(HUFF_STATE *shuff) {
+
+       FILE *fp;
+       char *auxfilename;
+       char *stopchar;
+       
+       if ((shuff) && (shuff->targetfile) && (shuff->freqtable)) {
+               /* Preparo el nombre del archivo con la tabla */
+               auxfilename = (char*)malloc(strlen(shuff->targetfile)+1);               
+               stopchar = strrchr(shuff->targetfile,'.');              
+               strncpy(auxfilename,shuff->targetfile,stopchar - shuff->targetfile);
+               auxfilename[stopchar - shuff->targetfile] = 0;
+               strcat(auxfilename,".ftb");
+               
+               /* Lo creamos y dumpeamos la tabla de frecuencias (modelo) */
+               if ((fp = fopen(auxfilename,"w")) == NULL) return 0;
+               fwrite(shuff->freqtable,sizeof(t_freq),256,fp);
+               if (fp) fclose(fp);
+                               
+               return 1;
+       }       
+       return 0;
+}
+
+void shuff_deinit_encoder(HUFF_STATE *shuff)
+{
+       /* Libero mallocs y cierro archivos */
+       if (shuff->freqtable) free(shuff->freqtable);
+       if (shuff->coderfp) fclose(shuff->coderfp);
+       if (shuff->bychunk) unlink(shuff->sourcefile);
+       if (shuff->sourcefile) free(shuff->sourcefile); 
+       if (shuff->targetfile) free(shuff->targetfile);                         
+       
+       /* Destruyo recursivamente el arbol de codigos */
+}
+
+void shuff_deinit_decoder(HUFF_STATE *shuff)
+{
+       /* Libero mallocs y cierro archivos */  
+       if (shuff->freqtable) free(shuff->freqtable);
+       if (shuff->sourcefile != NULL) free(shuff->sourcefile);
+       if (shuff->targetfile != NULL) free(shuff->targetfile);
+       if (shuff->decoderfp != NULL) vfclose(shuff->decoderfp);
+               
+       /* Destruyo recursivamente el arbol de codigos */
+}