size_t uread(void *buffer, size_t count, UFILE *rp)
{
switch( rp->type ) {
case UFT_HTTP:
return hread(buffer,count,rp->handle);
break;
case UFT_FILE:
return fread(buffer,1,count,rp->handle);
break;
default:
return -1;
break;
}
}
/**
* Open file and check if it is Huffile.
* Return NULL if not Huffile or other mistake, else return Huffile *.
*/
bool isHuffile(char *path) {
// open input
Huffile* input = hfopen(path, "r");
if (input == NULL)
{
printf("Could not open %s for reading.\n", path);
return 0;
}
// read in header
Huffeader header;
if (hread(&header, input) == false)
{
hfclose(input);
printf("Could not read header.\n");
return 0;
}
// check for magic number
if (header.magic != MAGIC)
{
hfclose(input);
printf("File was not huffed.\n");
return 0;
}
// check checksum
int checksum = header.checksum;
for (int i = 0; i < SYMBOLS; i++)
{
checksum -= header.frequencies[i];
}
if (checksum != 0)
{
hfclose(input);
printf("File was not huffed.\n");
return 0;
}
// close input
hfclose(input);
return true;
}
int32_t init_ramcoder(struct ramcoder *coder,HUFF *hp,bits256 *seed)
{
int32_t i,precision,numbits = 0;
if ( coder->lastsymbol == 0 )
coder->lastsymbol = 0x100, coder->upper_lastsymbol = (coder->lastsymbol + 1);
coder->cumulativeProb = coder->lower = coder->code = coder->underflowBits = coder->ranges[0] = 0;
for (i=1; i<=coder->upper_lastsymbol; i++)
{
coder->ranges[i] = coder->ranges[i - 1] + 1 + 10*((i <= 0x100) ? (GETBIT(seed->bytes,i-1) != 0) : 0);
//printf("%d ",coder->ranges[i]);
}
for (i=1; i<=coder->upper_lastsymbol; i++)
coder->cumulativeProb += (coder->ranges[i] - coder->ranges[i - 1]);
// printf("cumulative.%d\n",coder->cumulativeProb);
precision = (8 * sizeof(uint16_t));
coder->upper = (1LL << precision) - 1;
if ( hp != 0 )
coder->code = hread(&numbits,precision,hp), coder->code <<= (precision - numbits);
return(numbits);
}
/*
* only directories can be read. everthing else returns EOF.
*/
static void
rread(Fcall *f)
{
long cnt;
Fid *fidp;
cnt = f->count;
f->count = 0;
fidp = newfid(f->fid);
f->data = (char*)rbuf+IOHDRSZ;
if(fidp->open == 0) {
reply(f, "file not open");
return;
}
if ((fidp->node->d.mode&DMDIR) == 0) {
reply(f, 0); /*EOF*/
return;
}
if(cnt > MAXRPC)
cnt = MAXRPC;
if(f->offset == 0)
fidp->dirindex = 0;
switch(fidp->node->d.type) {
case Directory:
case Addrdir:
case Trusted:
f->count = dread(fidp, cnt);
break;
case IPaddr:
case Acctaddr:
f->count = hread(fidp, cnt);
break;
default:
reply(f, "can't read this type of file");
return;
}
reply(f, 0);
}
static ssize_t multipart_read(hFILE *fpv, void *buffer, size_t nbytes)
{
hFILE_multipart *fp = (hFILE_multipart *) fpv;
size_t n;
open_next:
if (fp->currentfp == NULL) {
if (fp->current < fp->nparts) {
const hfile_part *p = &fp->parts[fp->current];
hts_log_debug("Opening part #%zu of %zu: \"%.120s%s\"",
fp->current+1, fp->nparts, p->url,
(strlen(p->url) > 120)? "..." : "");
fp->currentfp = p->headers?
hopen(p->url, "r:", "httphdr:v", p->headers, NULL)
: hopen(p->url, "r");
if (fp->currentfp == NULL) return -1;
}
else return 0; // No more parts, so we're truly at EOF
}
n = fp->currentfp->mobile?
fp->currentfp->backend->read(fp->currentfp, buffer, nbytes)
: hread(fp->currentfp, buffer, nbytes);
if (n == 0) {
// We're at EOF on this part, so set up the next part
hFILE *prevfp = fp->currentfp;
free_part(&fp->parts[fp->current]);
fp->current++;
fp->currentfp = NULL;
if (hclose(prevfp) < 0) return -1;
goto open_next;
}
return n; // Number of bytes read by (or an error from) fp->currentfp
}
mFILE *find_file_url(char *file, char *url) {
char buf[8192], *cp;
mFILE *mf = NULL;
int maxlen = 8190 - strlen(file), len;
hFILE *hf;
/* Expand %s for the trace name */
for (cp = buf; *url && cp - buf < maxlen; url++) {
if (*url == '%' && *(url+1) == 's') {
url++;
cp += strlen(strcpy(cp, file));
} else {
*cp++ = *url;
}
}
*cp++ = 0;
if (!(hf = hopen(buf, "r")))
return NULL;
if (NULL == (mf = mfcreate(NULL, 0)))
return NULL;
while ((len = hread(hf, buf, 8192)) > 0) {
if (mfwrite(buf, len, 1, mf) <= 0) {
hclose_abruptly(hf);
mfdestroy(mf);
return NULL;
}
}
if (hclose(hf) < 0 || len < 0) {
mfdestroy(mf);
return NULL;
}
mrewind(mf);
return mf;
}
// Loads the contents of filename to produced a read-only, in memory,
// immobile hfile. fp is the already opened file. We always close this
// input fp, irrespective of whether we error or whether we return a new
// immobile hfile.
static hFILE *hpreload(hFILE *fp) {
hFILE *mem_fp;
char *buf = NULL;
off_t buf_sz = 0, buf_a = 0, buf_inc = 8192, len;
for (;;) {
if (buf_a - buf_sz < 5000) {
buf_a += buf_inc;
char *t = realloc(buf, buf_a);
if (!t) goto err;
buf = t;
if (buf_inc < 1000000) buf_inc *= 1.3;
}
len = hread(fp, buf+buf_sz, buf_a-buf_sz);
if (len > 0)
buf_sz += len;
else
break;
}
if (len < 0) goto err;
mem_fp = hfile_init_fixed(sizeof(hFILE), "r", buf, buf_sz, buf_a);
if (!mem_fp) goto err;
mem_fp->backend = &mem_backend;
if (hclose(fp) < 0) {
hclose_abruptly(mem_fp);
goto err;
}
return mem_fp;
err:
free(buf);
hclose_abruptly(fp);
return NULL;
}
bool createForest(char *path) {
// create empty forest
f = mkforest();
// open input
Huffile* input = hfopen(path, "r");
if (input == NULL)
{
printf("Could not open %s for reading.\n", path);
return 0;
}
// read in header
Huffeader header;
if (hread(&header, input) == false)
{
hfclose(input);
printf("Could not read header.\n");
return 0;
}
// read symbols and freq from header, create and plant trees in forest
for (int i = 0; i < SYMBOLS; i++)
{
if ( header.frequencies[i] != 0 )
{
Tree *t = mktree();
t->symbol = i;
t->frequency = header.frequencies[i];
plant(f, t);
count++;
}
}
// join trees as siblings
for (int i = 0; i < count - 1; i++)
{
Tree *t = mktree();
t->left = pick(f);
t->right = pick(f);
t->frequency = t->left->frequency + t->right->frequency;
plant(f, t);
}
// close input
hfclose(input);
return true;
}
int main(void)
{
static const int size[] = { 1, 13, 403, 999, 30000 };
char buffer[40000];
char *original;
int c, i;
ssize_t n;
off_t off;
reopen("vcf.c", "test/hfile1.tmp");
while ((c = hgetc(fin)) != EOF) {
if (hputc(c, fout) == EOF) fail("hputc");
}
if (herrno(fin)) { errno = herrno(fin); fail("hgetc"); }
reopen("test/hfile1.tmp", "test/hfile2.tmp");
if (hpeek(fin, buffer, 50) < 0) fail("hpeek");
while ((n = hread(fin, buffer, 17)) > 0) {
if (hwrite(fout, buffer, n) != n) fail("hwrite");
}
if (n < 0) fail("hread");
reopen("test/hfile2.tmp", "test/hfile3.tmp");
while ((n = hread(fin, buffer, sizeof buffer)) > 0) {
if (hwrite(fout, buffer, n) != n) fail("hwrite");
if (hpeek(fin, buffer, 700) < 0) fail("hpeek");
}
if (n < 0) fail("hread");
reopen("test/hfile3.tmp", "test/hfile4.tmp");
i = 0;
off = 0;
while ((n = hread(fin, buffer, size[i++ % 5])) > 0) {
off += n;
buffer[n] = '\0';
check_offset(fin, off, "pre-peek");
if (hputs(buffer, fout) == EOF) fail("hputs");
if ((n = hpeek(fin, buffer, size[(i+3) % 5])) < 0) fail("hpeek");
check_offset(fin, off, "post-peek");
}
if (n < 0) fail("hread");
reopen("test/hfile4.tmp", "test/hfile5.tmp");
n = hread(fin, buffer, 200);
if (n < 0) fail("hread");
else if (n != 200) fail("hread only got %d", (int)n);
if (hwrite(fout, buffer, 1000) != 1000) fail("hwrite");
check_offset(fin, 200, "input/first200");
check_offset(fout, 1000, "output/first200");
if (hseek(fin, 800, SEEK_CUR) < 0) fail("hseek/cur");
check_offset(fin, 1000, "input/seek");
for (off = 1000; (n = hread(fin, buffer, sizeof buffer)) > 0; off += n)
if (hwrite(fout, buffer, n) != n) fail("hwrite");
if (n < 0) fail("hread");
check_offset(fin, off, "input/eof");
check_offset(fout, off, "output/eof");
if (hseek(fin, 200, SEEK_SET) < 0) fail("hseek/set");
if (hseek(fout, 200, SEEK_SET) < 0) fail("hseek(output)");
check_offset(fin, 200, "input/backto200");
check_offset(fout, 200, "output/backto200");
n = hread(fin, buffer, 800);
if (n < 0) fail("hread");
else if (n != 800) fail("hread only got %d", (int)n);
if (hwrite(fout, buffer, 800) != 800) fail("hwrite");
check_offset(fin, 1000, "input/wrote800");
check_offset(fout, 1000, "output/wrote800");
if (hflush(fout) == EOF) fail("hflush");
original = slurp("vcf.c");
for (i = 1; i <= 5; i++) {
char *text;
sprintf(buffer, "test/hfile%d.tmp", i);
text = slurp(buffer);
if (strcmp(original, text) != 0) {
fprintf(stderr, "%s differs from vcf.c\n", buffer);
return EXIT_FAILURE;
}
free(text);
}
free(original);
if (hclose(fin) != 0) fail("hclose(input)");
if (hclose(fout) != 0) fail("hclose(output)");
fout = hopen("test/hfile_chars.tmp", "w");
if (fout == NULL) fail("hopen(\"test/hfile_chars.tmp\")");
for (i = 0; i < 256; i++)
if (hputc(i, fout) != i) fail("chars: hputc (%d)", i);
if (hclose(fout) != 0) fail("hclose(test/hfile_chars.tmp)");
fin = hopen("test/hfile_chars.tmp", "r");
if (fin == NULL) fail("hopen(\"test/hfile_chars.tmp\") for reading");
for (i = 0; i < 256; i++)
if ((c = hgetc(fin)) != i)
fail("chars: hgetc (%d = 0x%x) returned %d = 0x%x", i, i, c, c);
if ((c = hgetc(fin)) != EOF) fail("chars: hgetc (EOF) returned %d", c);
if (hclose(fin) != 0) fail("hclose(test/hfile_chars.tmp) for reading");
fin = hopen("data:hello, world!\n", "r");
if (fin == NULL) fail("hopen(\"data:...\")");
n = hread(fin, buffer, 300);
if (n < 0) fail("hread");
buffer[n] = '\0';
if (strcmp(buffer, "hello, world!\n") != 0) fail("hread result");
if (hclose(fin) != 0) fail("hclose(\"data:...\")");
return EXIT_SUCCESS;
}
int main(int argc, char* argv[])
{
// ensure proper usage
if (argc != 2)
{
printf("Usage: %s input\n", argv[0]);
return 1;
}
// open input
Huffile* input = hfopen(argv[1], "r");
if (input == NULL)
{
printf("Could not open %s for reading.\n", argv[1]);
return 1;
}
// read in header
Huffeader header;
if (hread(&header, input) == false)
{
hfclose(input);
printf("Could not read header.\n");
return 1;
}
// check for magic number
if (header.magic != MAGIC)
{
hfclose(input);
printf("File was not huffed.\n");
return 1;
}
// check checksum
int checksum = header.checksum;
for (int i = 0; i < SYMBOLS; i++)
{
checksum -= header.frequencies[i];
}
if (checksum != 0)
{
hfclose(input);
printf("File was not huffed.\n");
return 1;
}
// determine number of rows to use
const int rows = (int) ceil((LAST - FIRST) / (double) COLS);
// dump frequencies in a nice table
printf("\n");
for (int row = 0; row < rows; row++)
{
for (int col = 0; col < COLS; col++)
{
int index = FIRST + row + rows * col;
if (index > LAST)
{
break;
}
printf("%c %-6d ", index, header.frequencies[index]);
}
printf("\n");
}
printf("\n");
// dump bits contiguously
int bit;
while ((bit = bread(input)) != EOF)
{
printf("%d", bit);
}
printf("\n\n");
// close input
hfclose(input);
// that's all folks!
return 0;
}
int main(int argc, char* argv[])
{
// ensure proper usage
if (argc != 3)
{
printf("Usage: %s input output\n", argv[0]);
return 1;
}
// open input
Huffile* input = hfopen(argv[1], "r");
if (input == NULL)
{
printf("Could not open %s for reading.\n", argv[1]);
return 1;
}
// open outfile
FILE* outfile = fopen(argv[2], "w");
// read in header
Huffeader header;
if (hread(&header, input) == false)
{
hfclose(input);
printf("Could not read header.\n");
return 1;
}
// check for magic number
if (header.magic != MAGIC)
{
hfclose(input);
printf("File was not huffed.\n");
return 1;
}
// check checksum
int checksum = header.checksum;
for (int i = 0; i < SYMBOLS; i++)
{
checksum -= header.frequencies[i];
}
if (checksum != 0)
{
hfclose(input);
printf("File was not huffed.\n");
return 1;
}
// make forest
Forest* forest = mkforest();
// search for symbols that are non-zero frequency
for (int i = 0; i < SYMBOLS; i++)
{
// make and plant the trees in the forest
if (header.frequencies[i] >= 1)
{
Tree* newTree = mktree();
newTree->frequency = header.frequencies[i];
newTree->symbol = i;
newTree->left = NULL;
newTree->right = NULL;
plant(forest, newTree);
}
}
// run loop until there is only one tree left
bool done = false;
while (!done)
{
// pick smallest tree
Tree* a = pick(forest);
// pick second smallest tree
Tree* b = pick(forest);
// if there is only one tree left in the forest, a is the huffman tree
if (b == NULL)
{
done = true;
root = a;
}
// if two trees were succesfully picked
else
{
// combine the two trees by calling the combine function
Tree* combinedTree = combine(a, b);
// plant combined tree back in forest
plant(forest, combinedTree);
}
}
// write message to file
int bit;
Tree* cursor = root;
while ((bit = bread(input)) != EOF)
{
// if bit == 0 -> go left
if (bit == 0)
{
cursor = cursor->left;
}
// if bit == 1 -> go right
else if (bit == 1)
{
cursor = cursor->right;
}
// when you find a leaf
if ((cursor->right == NULL) && (cursor->left == NULL))
{
// print the leaf
fprintf(outfile, "%c", cursor->symbol);
// reset the cursor to root for the next iteration
cursor = root;
}
}
// free root
rmtree(root);
// close forest
rmforest(forest);
// close input
hfclose(input);
// close outfile
fclose(outfile);
// that's all folks!
return 0;
}
/*
* Loads a CRAM .crai index into memory.
*
* Returns 0 for success
* -1 for failure
*/
int cram_index_load(cram_fd *fd, const char *fn) {
char fn2[PATH_MAX];
char buf[65536];
ssize_t len;
kstring_t kstr = {0};
hFILE *fp;
cram_index *idx;
cram_index **idx_stack = NULL, *ep, e;
int idx_stack_alloc = 0, idx_stack_ptr = 0;
size_t pos = 0;
/* Check if already loaded */
if (fd->index)
return 0;
fd->index = calloc((fd->index_sz = 1), sizeof(*fd->index));
if (!fd->index)
return -1;
idx = &fd->index[0];
idx->refid = -1;
idx->start = INT_MIN;
idx->end = INT_MAX;
idx_stack = calloc(++idx_stack_alloc, sizeof(*idx_stack));
idx_stack[idx_stack_ptr] = idx;
sprintf(fn2, "%s.crai", fn);
if (!(fp = hopen(fn2, "r"))) {
perror(fn2);
free(idx_stack);
return -1;
}
// Load the file into memory
while ((len = hread(fp, buf, 65536)) > 0)
kputsn(buf, len, &kstr);
if (len < 0 || kstr.l < 2) {
if (kstr.s)
free(kstr.s);
free(idx_stack);
return -1;
}
if (hclose(fp)) {
if (kstr.s)
free(kstr.s);
free(idx_stack);
return -1;
}
// Uncompress if required
if (kstr.s[0] == 31 && (uc)kstr.s[1] == 139) {
size_t l;
char *s = zlib_mem_inflate(kstr.s, kstr.l, &l);
free(kstr.s);
if (!s) {
free(idx_stack);
return -1;
}
kstr.s = s;
kstr.l = l;
kstr.m = l; // conservative estimate of the size allocated
kputsn("", 0, &kstr); // ensure kstr.s is NUL-terminated
}
// Parse it line at a time
do {
/* 1.1 layout */
if (kget_int32(&kstr, &pos, &e.refid) == -1) {
free(kstr.s); free(idx_stack); return -1;
}
if (kget_int32(&kstr, &pos, &e.start) == -1) {
free(kstr.s); free(idx_stack); return -1;
}
if (kget_int32(&kstr, &pos, &e.end) == -1) {
free(kstr.s); free(idx_stack); return -1;
}
if (kget_int64(&kstr, &pos, &e.offset) == -1) {
free(kstr.s); free(idx_stack); return -1;
}
if (kget_int32(&kstr, &pos, &e.slice) == -1) {
free(kstr.s); free(idx_stack); return -1;
}
if (kget_int32(&kstr, &pos, &e.len) == -1) {
free(kstr.s); free(idx_stack); return -1;
}
e.end += e.start-1;
//printf("%d/%d..%d\n", e.refid, e.start, e.end);
if (e.refid < -1) {
free(kstr.s);
free(idx_stack);
fprintf(stderr, "Malformed index file, refid %d\n", e.refid);
return -1;
}
if (e.refid != idx->refid) {
if (fd->index_sz < e.refid+2) {
size_t index_end = fd->index_sz * sizeof(*fd->index);
fd->index_sz = e.refid+2;
fd->index = realloc(fd->index,
fd->index_sz * sizeof(*fd->index));
memset(((char *)fd->index) + index_end, 0,
fd->index_sz * sizeof(*fd->index) - index_end);
}
idx = &fd->index[e.refid+1];
idx->refid = e.refid;
idx->start = INT_MIN;
idx->end = INT_MAX;
idx->nslice = idx->nalloc = 0;
idx->e = NULL;
idx_stack[(idx_stack_ptr = 0)] = idx;
}
while (!(e.start >= idx->start && e.end <= idx->end)) {
idx = idx_stack[--idx_stack_ptr];
}
// Now contains, so append
if (idx->nslice+1 >= idx->nalloc) {
idx->nalloc = idx->nalloc ? idx->nalloc*2 : 16;
idx->e = realloc(idx->e, idx->nalloc * sizeof(*idx->e));
}
e.nalloc = e.nslice = 0; e.e = NULL;
*(ep = &idx->e[idx->nslice++]) = e;
idx = ep;
if (++idx_stack_ptr >= idx_stack_alloc) {
idx_stack_alloc *= 2;
idx_stack = realloc(idx_stack, idx_stack_alloc*sizeof(*idx_stack));
}
idx_stack[idx_stack_ptr] = idx;
while (pos < kstr.l && kstr.s[pos] != '\n')
pos++;
pos++;
} while (pos < kstr.l);
free(idx_stack);
free(kstr.s);
// dump_index(fd);
return 0;
}
/**
* decompress Huffile and print to file.
*/
bool decompress(char *path1, char *path2) {
// open input
Huffile* input = hfopen(path1, "r");
if (input == NULL)
{
printf("Could not open %s for reading.\n", path1);
return false;
}
// read in header
Huffeader header;
if (hread(&header, input) == false)
{
hfclose(input);
printf("Could not read header.\n");
return false;
}
// open file for writing
FILE *out = fopen(path2, "w");
if (out == NULL)
{
printf("Could not open %s for writing.\n", path2);
return false;
}
// read bits, crawl the tree and write to out
int bit;
char ch;
Tree *p = f->first->tree;
// check single-character file
if ( count == 1 )
{
ch = p->symbol;
if ( fputc(ch, out) == EOF )
{
printf("can't write\n");
return false;
}
return true;
}
while ((bit = bread(input)) != EOF)
{
if ( bit == 0 )
{
p = p->left;
if ( isLeaf(p) )
{
ch = p->symbol;
if ( fputc(ch, out) == EOF )
{
printf("can't write\n");
return false;
}
p = f->first->tree;
}
}
else if ( bit == 1 )
{
p = p->right;
if ( isLeaf(p) )
{
ch = p->symbol;
if ( fputc(ch, out) == EOF )
{
printf("can't write\n");
return false;
}
p = f->first->tree;
}
}
}
return true;
}
HTTPRESULT *http_read(char *url, int maxlen)
{
HTTPRESULT *result = http_new_result();
if ( strncmp(url,"http://",7)==0 )
{
HTTP *fp = hopen(url,maxlen);
if ( fp==NULL )
{
output_error("http_read('%s'): unable to access url", url);
result->status = errno;
}
else
{
size_t len = (int)hfilelength(fp);
size_t hlen = len;
char *buffer = (char*)malloc(len+1);
char *data = NULL;
if ( hread(buffer,len,fp)<=0 )
{
output_error("http_read('%s'): unable to read file", url);
result->status = errno;
}
else
{
buffer[len]='\0';
data = strstr(buffer,"\r\n\r\n");
if ( data!=NULL )
{
hlen = data - buffer;
*data++ = '\0';
result->body.size = (int)(len - hlen - 1);
while ( isspace(*data) ) { data++; result->body.size--; }
// if ( sscanf(data,"%x",&dlen)==1 )
// {
// data = strchr(data,'\n');
// while ( isspace(*data) ) {data++;}
// result->body.size = dlen;
// result->body.data = malloc(dlen+1);
// }
// else
// {
result->body.data = malloc(result->body.size+1);
// }
memcpy(result->body.data,data,result->body.size);
result->body.data[result->body.size] = '\0';
}
else
{
result->body.data = "";
result->body.size = 0;
}
result->header.size = (int)hlen;
result->header.data = malloc(hlen+1);
strcpy(result->header.data,buffer);
result->status = 0;
}
}
hclose(fp);
}
else if ( strncmp(url,"file://",7)==0 )
{
FILE *fp = fopen(url+7,"rt");
if ( fp==NULL )
{
output_error("http_read('%s'): unable to access file", url);
result->status = errno;
}
else
{
result->body.size = filelength(fileno(fp))+1;
result->body.data = malloc(result->body.size);
memset(result->body.data,0,result->body.size);
if ( fread(result->body.data,1,result->body.size,fp)<=0 )
{
output_error("http_read('%s'): unable to read file", url);
result->status = errno;
}
else
result->status = 0;
}
fclose(fp);
}
return result;
}
void
main(int argc, char **argv)
{
int i, j, a, mi, oi, tot, keywords;
double totp, p, xp[MAXTAB];
Hash *hmsg;
Word w;
Stringtab *s, *t;
Biobuf bout;
mbest = 15;
keywords = 0;
ARGBEGIN{
case 'D':
debug = 1;
break;
case 'k':
keywords = 1;
break;
case 'm':
mbest = atoi(EARGF(usage()));
if(mbest > MAXBEST)
sysfatal("cannot keep more than %d words", MAXBEST);
break;
default:
usage();
}ARGEND
for(i=0; i<argc; i++)
if(strcmp(argv[i], "~") == 0)
break;
if(i > MAXTAB)
sysfatal("cannot handle more than %d tables", MAXTAB);
if(i+1 >= argc)
usage();
for(i=0; i<argc; i++){
if(strcmp(argv[i], "~") == 0)
break;
tab[ntab].file = argv[i];
tab[ntab].hash = hread(argv[i]);
s = findstab(tab[ntab].hash, "*nmsg*", 6, 1);
if(s == nil || s->count == 0)
tab[ntab].nmsg = 1;
else
tab[ntab].nmsg = s->count;
ntab++;
}
Binit(&bout, 1, OWRITE);
oi = ++i;
for(a=i; a<argc; a++){
hmsg = hread(argv[a]);
nbest = 0;
for(s=hmsg->all; s; s=s->link){
w.s = s;
tot = 0;
totp = 0.0;
for(i=0; i<ntab; i++){
t = findstab(tab[i].hash, s->str, s->n, 0);
if(t == nil)
w.count[i] = 0;
else
w.count[i] = t->count;
tot += w.count[i];
p = w.count[i]/(double)tab[i].nmsg;
if(p >= 1.0)
p = 1.0;
w.p[i] = p;
totp += p;
}
if(tot < 5){ /* word does not appear enough; give to box 0 */
w.p[0] = 0.5;
for(i=1; i<ntab; i++)
w.p[i] = 0.1;
w.mp = 0.5;
w.mi = 0;
noteword(&w);
continue;
}
w.mp = 0.0;
for(i=0; i<ntab; i++){
p = w.p[i];
p /= totp;
if(p < 0.01)
p = 0.01;
else if(p > 0.99)
p = 0.99;
if(p > w.mp){
w.mp = p;
w.mi = i;
}
w.p[i] = p;
}
noteword(&w);
}
totp = 0.0;
for(i=0; i<ntab; i++){
p = 1.0;
for(j=0; j<nbest; j++)
p *= best[j].p[i];
xp[i] = p;
totp += p;
}
for(i=0; i<ntab; i++)
xp[i] /= totp;
mi = 0;
for(i=1; i<ntab; i++)
if(xp[i] > xp[mi])
mi = i;
if(oi != argc-1)
Bprint(&bout, "%s: ", argv[a]);
Bprint(&bout, "%s %f", tab[mi].file, xp[mi]);
if(keywords){
for(i=0; i<nbest; i++){
Bprint(&bout, " ");
Bwrite(&bout, best[i].s->str, best[i].s->n);
Bprint(&bout, " %f", best[i].p[mi]);
}
}
freehash(hmsg);
Bprint(&bout, "\n");
if(debug){
for(i=0; i<nbest; i++){
Bwrite(&bout, best[i].s->str, best[i].s->n);
Bprint(&bout, " %f", best[i].p[mi]);
if(best[i].p[mi] < best[i].mp)
Bprint(&bout, " (%f %s)", best[i].mp, tab[best[i].mi].file);
Bprint(&bout, "\n");
}
}
}
Bterm(&bout);
}
int main(int argc, char* argv[])
{
// ensure proper usage
if (argc != 3)
{
printf("Usage: %s input output\n", argv[0]);
return 1;
}
// open input
Huffile* input = hfopen(argv[1], "r");
if (input == NULL)
{
printf("Could not open %s for reading.\n", argv[1]);
return 1;
}
// read in header
Huffeader header;
if (hread(&header, input) == false)
{
hfclose(input);
printf("Could not read header.\n");
return 1;
}
// check for magic number
if (header.magic != MAGIC)
{
hfclose(input);
printf("File was not huffed.\n");
return 1;
}
// check checksum
int checksum = header.checksum;
for (int i = 0; i < SYMBOLS; i++)
{
checksum -= header.frequencies[i];
}
if (checksum != 0)
{
hfclose(input);
printf("File was not huffed.\n");
return 1;
}
int index;
Forest* f1 = mkforest();
for(index = 0; index<SYMBOLS;index++)
{
if(header.frequencies[index] != 0)
{
Tree* t1 = mktree();
t1->symbol = (char)index;
t1->frequency = header.frequencies[index];
plant(f1,t1);
}
}
while(true)
{
Tree* temp1 = pick(f1);
Tree* temp2 = pick(f1);
if(temp2 != NULL)
{
Tree* new_tree = mktree();
new_tree->left = temp1;
new_tree->right = temp2;
new_tree->frequency = temp1->frequency + temp2->frequency;
plant(f1,new_tree);
}
else
{
plant(f1,temp1);
break;
}
}
Tree* huffman_tree = pick(f1);
int bit;
Tree* temp = huffman_tree;
FILE *output = fopen(argv[2],"w");
while ((bit = bread(input)) != EOF)
{
if(bit == 1)
{
temp = temp->right;
if(temp->right == NULL && temp->left ==NULL)
{
fprintf(output,"%c",temp->symbol);
temp = huffman_tree;
}
}
else
{
temp = temp->left;
if(temp->right == NULL && temp->left ==NULL)
{
fprintf(output,"%c",temp->symbol);
temp = huffman_tree;
}
}
}
// close input
hfclose(input);
return 0;
}
int main(int argc, char* argv[])
{
// ensure proper usage
if (argc != 3)
{
printf("Usage: %s input\n", argv[0]);
return 1;
}
// open input
Huffile* input = hfopen(argv[1], "r");
if (input == NULL)
{
printf("Could not open %s for reading.\n", argv[1]);
return 1;
}
// open outfile
FILE* outfile = fopen(argv[2], "w");
if (outfile == NULL)
{
fclose(outfile);
fprintf(stderr, "Could not create outfile.\n");
return 2;
}
// read in header
Huffeader header;
if (hread(&header, input) == false)
{
hfclose(input);
printf("Could not read header.\n");
return 1;
}
// check for magic number
if (header.magic != MAGIC)
{
hfclose(input);
printf("File was not huffed.\n");
return 1;
}
// check checksum
int checksum = header.checksum;
for (int i = 0; i < SYMBOLS; i++)
{
checksum -= header.frequencies[i];
}
if (checksum != 0)
{
hfclose(input);
printf("File was not huffed.\n");
return 1;
}
// make forest
Forest* forest = mkforest();
// read in huffeader frequencies
for (int i = 0; i < SYMBOLS; i++)
{
// ignore 0 frequencies
if (header.frequencies[i] > 0)
{
// make new tree for every non-zero frequency occurance
Tree* new_tree = mktree();
new_tree->symbol = i;
new_tree->frequency = header.frequencies[i];
new_tree->left = NULL;
new_tree->right = NULL;
// plant every non-zero frequency tree in forest
plant(forest, new_tree);
}
}
// run loop until there is only one tree left
bool done = false;
while (!done)
{
// pick smallest tree from forest
Tree* a = pick(forest);
// pick second smallest tree from forest
Tree* b = pick(forest);
// if there is no second tree in forest...
if (b == NULL)
{
// break loop
done = true;
// set root to tree 'a' (last remaining) tree
root = a;
}
// else there are at least two remaining trees in the forest
else
{
// combine the two trees into a parent tree
Tree* parent_tree = combine(a, b);
// plant combined tree in forest
plant(forest, parent_tree);
}
}
// write message to outfile
int bit;
Tree* ptr_location = root;
while ((bit = bread(input)) != EOF)
{
// if bit is 0, go left, else go right
if (bit == 0)
ptr_location = ptr_location->left;
// if bit is 1, go right
if (bit == 1)
ptr_location = ptr_location->right;
// leaf is found when both branchs are NULL
if ((ptr_location->left == NULL) && (ptr_location->right == NULL))
{
// write leaf's symbol to outfile
fprintf(outfile, "%c", ptr_location->symbol);
// reset pointer location to root for next iteration of tree
ptr_location = root;
}
}
// free root
rmtree(root);
// close forest
rmforest(forest);
// close input & outfile
hfclose(input);
fclose(outfile);
// that's all folks!
return 0;
}
