X-Git-Url: http://git.indexdata.com/?p=yaz-moved-to-github.git;a=blobdiff_plain;f=retrieval%2Fd1_absyn.c;h=7813015cbbdb97cac10acebe7e8117beceb1d973;hp=1b4b44bf383c87e9c391425ffad2735841c835ab;hb=ec29e37cd269b330eb493157dedf9aacf5c6ca46;hpb=4090fc5c7b77769020633d09cf63b325aceaf49a diff --git a/retrieval/d1_absyn.c b/retrieval/d1_absyn.c index 1b4b44b..7813015 100644 --- a/retrieval/d1_absyn.c +++ b/retrieval/d1_absyn.c @@ -1,10 +1,34 @@ /* - * Copyright (c) 1995-1997, Index Data. + * Copyright (c) 1995-1998, Index Data. * See the file LICENSE for details. * Sebastian Hammer, Adam Dickmeiss * * $Log: d1_absyn.c,v $ - * Revision 1.15 1997-12-09 16:18:16 adam + * Revision 1.21 1998-06-09 13:55:07 adam + * Minor changes. + * + * Revision 1.20 1998/05/18 13:07:02 adam + * Changed the way attribute sets are handled by the retriaval module. + * Extended Explain conversion / schema. + * Modified server and client to work with ASN.1 compiled protocol handlers. + * + * Revision 1.19 1998/03/05 08:15:32 adam + * Implemented data1_add_insert_taggeddata utility which is more flexible + * than data1_insert_taggeddata. + * + * Revision 1.18 1998/02/27 14:08:04 adam + * Added const to some char pointer arguments. + * Reworked data1_read_node so that it doesn't create a tree with + * pointers to original "SGML"-buffer. + * + * Revision 1.17 1998/02/11 11:53:34 adam + * Changed code so that it compiles as C++. + * + * Revision 1.16 1997/12/18 10:51:30 adam + * Implemented sub-trees feature for schemas - including forward + * references. + * + * Revision 1.15 1997/12/09 16:18:16 adam * Work on EXPLAIN schema. First implementation of sub-schema facility * in the *.abs files. * @@ -78,6 +102,13 @@ struct data1_absyn_cache_info data1_absyn_cache next; }; +struct data1_attset_cache_info +{ + char *name; + data1_attset *attset; + data1_attset_cache next; +}; + data1_absyn *data1_absyn_search (data1_handle dh, const char *name) { data1_absyn_cache p = *data1_absyn_cache_get (dh); @@ -91,12 +122,24 @@ data1_absyn *data1_absyn_search (data1_handle dh, const char *name) return NULL; } +void data1_absyn_trav (data1_handle dh, void *handle, + void (*fh)(data1_handle dh, void *h, data1_absyn *a)) +{ + data1_absyn_cache p = *data1_absyn_cache_get (dh); + + while (p) + { + (*fh)(dh, handle, p->absyn); + p = p->next; + } +} + data1_absyn *data1_absyn_add (data1_handle dh, const char *name) { char fname[512]; NMEM mem = data1_nmem_get (dh); - data1_absyn_cache p = nmem_malloc (mem, sizeof(*p)); + data1_absyn_cache p = (data1_absyn_cache)nmem_malloc (mem, sizeof(*p)); data1_absyn_cache *pp = data1_absyn_cache_get (dh); sprintf(fname, "%s.abs", name); @@ -107,7 +150,7 @@ data1_absyn *data1_absyn_add (data1_handle dh, const char *name) return p->absyn; } -data1_absyn *data1_get_absyn (data1_handle dh, char *name) +data1_absyn *data1_get_absyn (data1_handle dh, const char *name) { data1_absyn *absyn; @@ -116,8 +159,75 @@ data1_absyn *data1_get_absyn (data1_handle dh, char *name) return absyn; } +data1_attset *data1_attset_search_name (data1_handle dh, const char *name) +{ + data1_attset_cache p = *data1_attset_cache_get (dh); + + while (p) + { + if (!strcmp (name, p->name)) + return p->attset; + p = p->next; + } + return NULL; +} + +data1_attset *data1_attset_search_id (data1_handle dh, int id) +{ + data1_attset_cache p = *data1_attset_cache_get (dh); + + while (p) + { + if (id == p->attset->reference) + return p->attset; + p = p->next; + } + return NULL; +} + +data1_attset *data1_attset_add (data1_handle dh, const char *name) +{ + char fname[512], aname[512]; + NMEM mem = data1_nmem_get (dh); + data1_attset *attset; + + strcpy (aname, name); + sprintf(fname, "%s.att", name); + attset = data1_read_attset (dh, fname); + if (!attset) + { + char *cp; + attset = data1_read_attset (dh, name); + if (attset && (cp = strrchr (aname, '.'))) + *cp = '\0'; + } + if (!attset) + logf (LOG_WARN|LOG_ERRNO, "couldn't load attribute set %s", name); + else + { + data1_attset_cache p = (data1_attset_cache) + nmem_malloc (mem, sizeof(*p)); + data1_attset_cache *pp = data1_attset_cache_get (dh); + + attset->name = p->name = nmem_strdup (mem, aname); + p->attset = attset; + p->next = *pp; + *pp = p; + } + return attset; +} + +data1_attset *data1_get_attset (data1_handle dh, const char *name) +{ + data1_attset *attset; + + if (!(attset = data1_attset_search_name (dh, name))) + attset = data1_attset_add (dh, name); + return attset; +} + data1_esetname *data1_getesetbyname(data1_handle dh, data1_absyn *a, - char *name) + const char *name) { data1_esetname *r; @@ -129,7 +239,7 @@ data1_esetname *data1_getesetbyname(data1_handle dh, data1_absyn *a, data1_element *data1_getelementbytagname (data1_handle dh, data1_absyn *abs, data1_element *parent, - char *tagname) + const char *tagname) { data1_element *r; @@ -150,7 +260,7 @@ data1_element *data1_getelementbytagname (data1_handle dh, data1_absyn *abs, } data1_element *data1_getelementbyname (data1_handle dh, data1_absyn *absyn, - char *name) + const char *name) { data1_element *r; assert (absyn->main_elements); @@ -160,6 +270,30 @@ data1_element *data1_getelementbyname (data1_handle dh, data1_absyn *absyn, return 0; } + +void fix_element_ref (data1_handle dh, data1_absyn *absyn, data1_element *e) +{ + for (; e; e = e->next) + { + if (!e->sub_name) + { + if (e->children) + fix_element_ref (dh, absyn, e->children); + } + else + { + data1_sub_elements *sub_e = absyn->sub_elements; + while (sub_e && strcmp (e->sub_name, sub_e->name)) + sub_e = sub_e->next; + if (sub_e) + e->children = sub_e->elements; + else + logf (LOG_WARN, "Unresolved reference to sub-elements %s", + e->sub_name); + } + } +} + data1_absyn *data1_read_absyn (data1_handle dh, const char *file) { char line[512], *r, cmd[512], args[512]; @@ -171,16 +305,15 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) data1_maptab **maptabp; data1_marctab **marcp; data1_termlist *all = 0; - int level; + int level = 0; - logf (LOG_DEBUG, "begin data1_read_absyn file=%s", file); if (!(f = yaz_path_fopen(data1_get_tabpath (dh), file, "r"))) { logf(LOG_WARN|LOG_ERRNO, "Couldn't open %s", file); return 0; } - res = nmem_malloc(data1_nmem_get(dh), sizeof(*res)); + res = (data1_absyn *)nmem_malloc(data1_nmem_get(dh), sizeof(*res)); res->name = 0; res->reference = VAL_NONE; res->tagset = 0; @@ -213,13 +346,13 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) { data1_element *new_element; int i; - char path[512], name[512], termlists[512], *p; + char path[512], name[512], termlists[512], *p, *sub_p; int type, value; data1_termlist **tp; if (!cur_elements) { - cur_elements = nmem_malloc(data1_nmem_get(dh), + cur_elements = (data1_sub_elements *)nmem_malloc(data1_nmem_get(dh), sizeof(*cur_elements)); cur_elements->next = res->sub_elements; cur_elements->elements = NULL; @@ -253,29 +386,24 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) return 0; } level = i; - if (*p == '$' && level > 0) - { - data1_sub_elements *sub_e = res->sub_elements; - - p++; - while (sub_e && strcmp (p, sub_e->name)) - sub_e = sub_e->next; - if (sub_e) - *ppl[level] = sub_e->elements; - if (level) - level--; - continue; - } new_element = *ppl[level] = - nmem_malloc(data1_nmem_get(dh), sizeof(*new_element)); + (data1_element *)nmem_malloc(data1_nmem_get(dh), sizeof(*new_element)); new_element->next = new_element->children = 0; new_element->tag = 0; new_element->termlists = 0; + new_element->sub_name = 0; tp = &new_element->termlists; ppl[level] = &new_element->next; ppl[level+1] = &new_element->children; - + + /* consider subtree (if any) ... */ + if ((sub_p = strchr (p, ':')) && sub_p[1]) + { + *sub_p++ = '\0'; + new_element->sub_name = + nmem_strdup (data1_nmem_get(dh), sub_p); + } /* well-defined tag */ if (sscanf(p, "(%d,%d)", &type, &value) == 2) { @@ -298,11 +426,11 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) else if (*p) { data1_tag *nt = - new_element->tag = nmem_malloc(data1_nmem_get (dh), + new_element->tag = (data1_tag *)nmem_malloc(data1_nmem_get (dh), sizeof(*new_element->tag)); nt->which = DATA1T_string; nt->value.string = nmem_strdup(data1_nmem_get (dh), p); - nt->names = nmem_malloc(data1_nmem_get(dh), + nt->names = (data1_name *)nmem_malloc(data1_nmem_get(dh), sizeof(*new_element->tag->names)); nt->names->name = nt->value.string; nt->names->next = 0; @@ -316,7 +444,6 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) fclose(f); return 0; } - /* parse termList definitions */ p = termlists; if (*p == '-') @@ -344,7 +471,7 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) } if (*attname == '!') strcpy(attname, name); - *tp = nmem_malloc(data1_nmem_get(dh), sizeof(**tp)); + *tp = (data1_termlist *)nmem_malloc(data1_nmem_get(dh), sizeof(**tp)); (*tp)->next = 0; if (!((*tp)->att = data1_getattbyname(dh, res->attset, attname))) @@ -358,7 +485,7 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) (*tp)->structure = "w"; else { - (*tp)->structure = nmem_malloc (data1_nmem_get (dh), + (*tp)->structure = (char *)nmem_malloc (data1_nmem_get (dh), strlen(structure)+1); strcpy ((*tp)->structure, structure); } @@ -367,7 +494,6 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) while ((p = strchr(p, ',')) && *(++p)); *tp = all; /* append any ALL entries to the list */ } - new_element->name = nmem_strdup(data1_nmem_get (dh), name); } else if (!strcmp(cmd, "section")) @@ -379,7 +505,7 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) file, args); continue; } - cur_elements = nmem_malloc(data1_nmem_get(dh), + cur_elements = (data1_sub_elements *)nmem_malloc(data1_nmem_get(dh), sizeof(*cur_elements)); cur_elements->next = res->sub_elements; cur_elements->elements = NULL; @@ -421,7 +547,7 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) fclose(f); return 0; } - *tp = nmem_malloc(data1_nmem_get(dh), sizeof(**tp)); + *tp = (data1_termlist *)nmem_malloc(data1_nmem_get(dh), sizeof(**tp)); if (!((*tp)->att = data1_getattbyname (dh, res->attset, attname))) { @@ -434,7 +560,7 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) (*tp)->structure = "w"; else { - (*tp)->structure = nmem_malloc (data1_nmem_get (dh), + (*tp)->structure = (char *)nmem_malloc (data1_nmem_get (dh), strlen(structure)+1); strcpy ((*tp)->structure, structure); } @@ -482,7 +608,7 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) fclose(f); return 0; } - if (!(res->attset = data1_read_attset (dh, name))) + if (!(res->attset = data1_get_attset (dh, name))) { logf(LOG_WARN, "Attset failed in %s", file); fclose(f); @@ -534,7 +660,7 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) fclose(f); return 0; } - *esetpp = nmem_malloc(data1_nmem_get(dh), sizeof(**esetpp)); + *esetpp = (data1_esetname *)nmem_malloc(data1_nmem_get(dh), sizeof(**esetpp)); (*esetpp)->name = nmem_strdup(data1_nmem_get(dh), name); (*esetpp)->next = 0; if (*fname == '@') @@ -591,13 +717,14 @@ data1_absyn *data1_read_absyn (data1_handle dh, const char *file) } } fclose(f); - - cur_elements = res->sub_elements; - while (cur_elements && strcmp (cur_elements->name, "main")) - cur_elements = cur_elements->next; - if (cur_elements) - res->main_elements = cur_elements->elements; - + + for (cur_elements = res->sub_elements; cur_elements; + cur_elements = cur_elements->next) + { + if (!strcmp (cur_elements->name, "main")) + res->main_elements = cur_elements->elements; + fix_element_ref (dh, res, cur_elements->elements); + } logf (LOG_DEBUG, "end data1_read_absyn file=%s", file); return res; }