1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296
|
/* Iterating through the documents in a wi2dvf, according to document type */
/* Copyright (C) 1998, 1999 Andrew McCallum
Written by: Andrew McCallum
This file is part of the Bag-Of-Words Library, `libbow'.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Library General Public License
as published by the Free Software Foundation, version 2.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Library General Public License for more details.
You should have received a copy of the GNU Library General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA */
#include <bow/libbow.h>
#include <math.h>
#include <time.h>
#include <sys/time.h>
#include <unistd.h>
/* This function sets up the data structure so we can step through the word
vectors for each test document easily. */
bow_dv_heap *
bow_test_new_heap (bow_barrel *barrel)
{
return bow_make_dv_heap_from_wi2dvf (barrel->wi2dvf);
}
/* We only need this struct within the following functions. */
typedef struct _bow_tmp_word_struct {
int wi;
int count;
float weight;
} bow_tmp_word_struct;
/* This function takes the heap returned by bow_initialise_test_set and
creates a word vector corresponding to the next document in the test set.
The index of the test document is returned. If the test set is empty, 0
is returned and *wv == NULL. Also, when the test set is exhausted, the
heap is free'd (since it can't be used for anything else anways.
This can't really deal with vectors which are all zero, since they
are not represented explicitly in our data structure. Not sure what
we should/can do. */
int
bow_heap_next_wv_guts (bow_dv_heap *heap, bow_barrel *barrel, bow_wv **wv,
int (*use_if_true)(bow_cdoc*))
{
bow_array *cdocs = barrel->cdocs;
bow_array *word_array;
bow_cdoc *doc;
bow_tmp_word_struct word, *wptr;
int current_di = -1;
int i;
word_array = bow_array_new (50, sizeof(bow_tmp_word_struct), 0);
/* Keep going until we exhaust the heap or we find a test document. */
while ((heap->length > 0) && (word_array->length == 0))
{
current_di = heap->entry[0].current_di;
doc = bow_cdocs_di2doc(cdocs, current_di);
if ((*use_if_true)(doc))
{
/* We have the first entry for the next document */
do
{
word.wi = heap->entry[0].wi;
word.count =
heap->entry[0].dv->entry[heap->entry[0].index].count;
word.weight =
heap->entry[0].dv->entry[heap->entry[0].index].weight;
bow_array_append (word_array, &word);
bow_dv_heap_update (heap);
}
while ((heap->length > 0)
&& (current_di == heap->entry[0].current_di));
}
else
{
/* This document did not pass the test; go on to next document */
do
{
bow_dv_heap_update (heap);
}
while ((heap->length > 0)
&& (current_di == heap->entry[0].current_di));
}
}
/* Here we either have a word_array or else we've run out of documents. */
if (word_array->length != 0)
{
if (*wv)
bow_wv_free (*wv);
/* We now have all the words for this test document in the word
array - need to create a bow_wv */
(*wv) = bow_wv_new (word_array->length);
for (i = 0; i < word_array->length; i++)
{
wptr = bow_array_entry_at_index (word_array, i);
(*wv)->entry[i].wi = wptr->wi;
(*wv)->entry[i].count = wptr->count;
(*wv)->entry[i].weight = wptr->weight;
}
}
else
{
/* Since we've run out of docs, might as well free the test set. */
if (*wv)
bow_wv_free (*wv);
current_di = -1;
(*wv) = NULL;
}
/* Should be finished with the word array now */
#if 1
/* XXX this causes a seg fault - don't know why. */
bow_array_free (word_array);
#else
/* This does the same job for me. */
bow_free (word_array->entries);
bow_free (word_array);
#endif
return current_di;
}
static bow_wv *empty_wv = NULL;
/* Use a heap to efficiently iterate through all the documents that
satisfy the condition USE_IF_TRUE(). Each call to this function
provides the next word vector in *WV, returning the `document index' DI of
the document contained in *WV. This function returns -1 when there
are no more documents left. */
int
bow_heap_next_wv (bow_dv_heap *heap, bow_barrel *barrel, bow_wv **wv,
int (*use_if_true)(bow_cdoc*))
{
int new_di;
bow_cdoc *doc_cdoc;
/* Initialize EMPTY_WV if it hasn't been already. */
if (!empty_wv)
{
empty_wv = bow_wv_new (1);
empty_wv->num_entries = 0;
}
/* This special -2 value set in heap.c */
if (heap->last_di == -2)
{
/* This is the first time this function is being called on this heap */
/* Initialize the *WV to NULL so that the underlying heap function's
free()'ing behavior will work properly */
*wv = NULL;
/* Set this to -1 so that when NEW_DI gets incremented, it becomes 0
and we get the first document. */
new_di = -1;
}
else
{
new_di = heap->last_di;
}
do
{
new_di++;
if (new_di >= barrel->cdocs->length)
{
/* No more satisfying documents left */
heap->heap_wv_di =
bow_heap_next_wv_guts (heap, barrel, &(heap->heap_wv),use_if_true);
assert (heap->heap_wv_di == -1);
bow_free (heap);
return -1;
}
doc_cdoc = bow_array_entry_at_index (barrel->cdocs, new_di);
}
while (!(*use_if_true) (doc_cdoc));
/* NEW_DI is now definitely the document index we will return.
It's just a question of whether that document is empty or not. */
if (heap->heap_wv_di < new_di && heap->heap_wv_di != -1)
heap->heap_wv_di =
bow_heap_next_wv_guts (heap, barrel, &(heap->heap_wv), use_if_true);
assert (heap->heap_wv_di >= new_di || heap->heap_wv_di == -1);
if (heap->heap_wv_di == new_di)
*wv = heap->heap_wv;
else
*wv = empty_wv;
heap->last_di = new_di;
return new_di;
}
int
bow_cdoc_is_train (bow_cdoc *cdoc)
{
return (cdoc->type == bow_doc_train);
}
int
bow_cdoc_is_test (bow_cdoc *cdoc)
{
return (cdoc->type == bow_doc_test);
}
int
bow_cdoc_yes (bow_cdoc *cdoc)
{
return (1);
}
int
bow_doc_yes (bow_doc *doc)
{
return (1);
}
/* Return nonzero iff CDOC has type != TEST */
int
bow_cdoc_is_nontest (bow_cdoc *cdoc)
{
return (cdoc->type != bow_doc_test);
}
int
bow_doc_is_nontest (bow_doc *doc)
{
return (doc->type != bow_doc_test);
}
/* Return nonzero iff CDOC has type == IGNORE */
int
bow_cdoc_is_ignore (bow_cdoc *cdoc)
{
return(cdoc->type == bow_doc_ignore);
}
/* Return nonzero iff CDOC has type == bow_doc_unlabeled */
int
bow_cdoc_is_unlabeled (bow_cdoc *cdoc)
{
return (cdoc->type == bow_doc_unlabeled);
}
/* Return nonzero iff CDOC has type == bow_doc_validation */
int
bow_cdoc_is_validation (bow_cdoc *cdoc)
{
return (cdoc->type == bow_doc_validation);
}
/* Return nonzero iff CDOC has type == bow_doc_pool */
int
bow_cdoc_is_pool (bow_cdoc *cdoc)
{
return (cdoc->type == bow_doc_pool);
}
/* Return nonzero iff CDOC has type == bow_doc_waiting */
int
bow_cdoc_is_waiting (bow_cdoc *cdoc)
{
return (cdoc->type == bow_doc_waiting);
}
int
bow_test_next_wv(bow_dv_heap *heap, bow_barrel *barrel, bow_wv **wv)
{
return (bow_heap_next_wv(heap, barrel, wv, bow_cdoc_is_test));
}
int
bow_nontest_next_wv(bow_dv_heap *heap, bow_barrel *barrel, bow_wv **wv)
{
return (bow_heap_next_wv(heap, barrel, wv, bow_cdoc_is_nontest));
}
int
bow_model_next_wv(bow_dv_heap *heap, bow_barrel *barrel, bow_wv **wv)
{
return (bow_heap_next_wv(heap, barrel, wv, bow_cdoc_is_train));
}
|