X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Fmath%2Fcategoricals.c;h=ef9c528994428775614adedc1e59ab0949fb1ebc;hb=c360fff4fd3e4a98cfe02441f43c27725cead44b;hp=1f302a0aa78c679d17944ca779e28e84d9e85c84;hpb=46fdaed173d06e42dac6712e142bc9f7898075d7;p=pspp diff --git a/src/math/categoricals.c b/src/math/categoricals.c index 1f302a0aa7..ef9c528994 100644 --- a/src/math/categoricals.c +++ b/src/math/categoricals.c @@ -34,6 +34,9 @@ struct value_node struct hmap_node node; /* Node in hash map. */ union value value; /* The value being labeled. */ double cc; /* The total of the weights of cases with this value */ + + void *user_data; /* A pointer to data which the caller can store stuff */ + int subscript; /* A zero based integer, unique within the variable. Can be used as an index into an array */ }; @@ -84,6 +87,17 @@ struct categoricals /* Missing values to be excluded */ enum mv_class exclude; + + + + /* Function to be called on each update */ + update_func *update; + + /* Auxilliary data to be passed to update */ + void *update_aux; + + /* Function specified by the caller to create user_data */ + user_data_create_func *user_data_create; }; @@ -91,14 +105,18 @@ void categoricals_destroy ( struct categoricals *cat) { int i; - for (i = 0 ; i < cat->n_vp; ++i) - hmap_destroy (&cat->vp[i].map); - - pool_destroy (cat->pool); - free (cat); + if (cat != NULL) + { + for (i = 0 ; i < cat->n_vp; ++i) + hmap_destroy (&cat->vp[i].map); + + pool_destroy (cat->pool); + free (cat); + } } +#if 0 void categoricals_dump (const struct categoricals *cat) { @@ -129,8 +147,8 @@ categoricals_dump (const struct categoricals *cat) for (node = hmap_first (m); node; node = hmap_next (m, node)) { struct string s; - ds_init_empty (&s); const struct value_node *vn = HMAP_DATA (node, struct value_node, node); + ds_init_empty (&s); var_append_value_name (vp->var, &vn->value, &s); printf ("Value: %s; Index %d; CC %g\n", ds_cstr (&s), @@ -145,7 +163,14 @@ categoricals_dump (const struct categoricals *cat) printf ("Number of categorical variables: %d\n", cat->n_vp); printf ("Number of non-empty categorical variables: %d\n", cat->n_vars); printf ("Total number of categories: %d\n", cat->n_cats_total); + + printf ("\nReverse variable map:\n"); + + for (v = 0 ; v < cat->n_cats_total; ++v) + printf ("%d ", cat->reverse_variable_map[v]); + printf ("\n"); } +#endif @@ -168,10 +193,12 @@ lookup_value (const struct hmap *map, const struct variable *var, const union va } - struct categoricals * -categoricals_create (const struct variable **v, size_t n_vars, - const struct variable *wv, enum mv_class exclude) +categoricals_create (const struct variable *const *v, size_t n_vars, + const struct variable *wv, enum mv_class exclude, + user_data_create_func *udf, + update_func *update, void *aux + ) { size_t i; struct categoricals *cat = xmalloc (sizeof *cat); @@ -183,6 +210,9 @@ categoricals_create (const struct variable **v, size_t n_vars, cat->reverse_variable_map = NULL; cat->pool = pool_create (); cat->exclude = exclude; + cat->update = update; + cat->update_aux = aux; + cat->user_data_create = udf; cat->vp = pool_calloc (cat->pool, cat->n_vp, sizeof *cat->vp); @@ -235,10 +265,16 @@ categoricals_update (struct categoricals *cat, const struct ccase *c) cat->n_vars++; node->subscript = cat->vp[i].n_cats++ ; + + if ( cat->user_data_create ) + node->user_data = cat->user_data_create (); } node->cc += weight; cat->vp[i].cc += weight; + + if ( cat->update) + cat->update (node->user_data, cat->wv, var, c, cat->update_aux); } } @@ -250,19 +286,6 @@ categoricals_n_count (const struct categoricals *cat, size_t n) } -/* Return the index for value VAL in the Nth variable */ -int -categoricals_index (const struct categoricals *cat, size_t n, const union value *val) -{ - struct value_node *vn = lookup_value (&cat->vp[n].map, cat->vp[n].var, val); - - if ( vn == NULL) - return -1; - - return vn->subscript; -} - - /* Return the total number of categories */ size_t categoricals_total (const struct categoricals *cat) @@ -284,7 +307,9 @@ categoricals_done (struct categoricals *cat) */ int v; int idx = 0; - cat->reverse_variable_map = pool_calloc (cat->pool, cat->n_cats_total, sizeof *cat->reverse_variable_map); + cat->reverse_variable_map = pool_calloc (cat->pool, + cat->n_cats_total, + sizeof *cat->reverse_variable_map); for (v = 0 ; v < cat->n_vp; ++v) { @@ -303,6 +328,8 @@ categoricals_done (struct categoricals *cat) vp->reverse_value_map[vn->subscript] = vn; } + /* Populate the reverse variable map. + */ for (i = 0; i < vp->n_cats; ++i) cat->reverse_variable_map[idx++] = v; } @@ -311,26 +338,31 @@ categoricals_done (struct categoricals *cat) } +static int +reverse_variable_lookup (const struct categoricals *cat, int subscript) +{ + assert (cat->reverse_variable_map); + assert (subscript >= 0); + assert (subscript < cat->n_cats_total); + + return cat->reverse_variable_map[subscript]; +} + /* Return the categorical variable corresponding to SUBSCRIPT */ const struct variable * categoricals_get_variable_by_subscript (const struct categoricals *cat, int subscript) { - int index; - - assert (cat->reverse_variable_map); - - index = cat->reverse_variable_map[subscript]; + int index = reverse_variable_lookup (cat, subscript); return cat->vp[index].var; } - /* Return the value corresponding to SUBSCRIPT */ const union value * categoricals_get_value_by_subscript (const struct categoricals *cat, int subscript) { - int vindex = cat->reverse_variable_map[subscript]; + int vindex = reverse_variable_lookup (cat, subscript); const struct var_params *vp = &cat->vp[vindex]; const struct value_node *vn = vp->reverse_value_map [subscript - vp->base_subscript]; @@ -341,7 +373,7 @@ categoricals_get_value_by_subscript (const struct categoricals *cat, int subscri double categoricals_get_weight_by_subscript (const struct categoricals *cat, int subscript) { - int vindex = cat->reverse_variable_map[subscript]; + int vindex = reverse_variable_lookup (cat, subscript); const struct var_params *vp = &cat->vp[vindex]; return vp->cc; @@ -350,7 +382,7 @@ categoricals_get_weight_by_subscript (const struct categoricals *cat, int subscr double categoricals_get_sum_by_subscript (const struct categoricals *cat, int subscript) { - int vindex = cat->reverse_variable_map[subscript]; + int vindex = reverse_variable_lookup (cat, subscript); const struct var_params *vp = &cat->vp[vindex]; const struct value_node *vn = vp->reverse_value_map [subscript - vp->base_subscript]; @@ -378,3 +410,15 @@ categoricals_get_n_variables (const struct categoricals *cat) { return cat->n_vars; } + + + +void * +categoricals_get_user_data_by_subscript (const struct categoricals *cat, int subscript) +{ + int vindex = reverse_variable_lookup (cat, subscript); + const struct var_params *vp = &cat->vp[vindex]; + + const struct value_node *vn = vp->reverse_value_map [subscript - vp->base_subscript]; + return vn->user_data; +}