CCL: re-factor and use internal utility ccl_rpn_node_mkbool
[yaz-moved-to-github.git] / src / cclfind.c
index b108b74..9403634 100644 (file)
@@ -100,7 +100,7 @@ struct ccl_rpn_node *ccl_rpn_node_create(enum ccl_rpn_kind kind)
     ccl_assert(p);
     p->kind = kind;
 
-    switch(kind)
+    switch (kind)
     {
     case CCL_RPN_TERM:
         p->u.t.attr_list = 0;
@@ -113,6 +113,71 @@ struct ccl_rpn_node *ccl_rpn_node_create(enum ccl_rpn_kind kind)
     return p;
 }
 
+static struct ccl_rpn_node *ccl_rpn_node_mkbool(struct ccl_rpn_node *l,
+                                                struct ccl_rpn_node *r,
+                                                enum ccl_rpn_kind op)
+{
+    if (l && r)
+    {
+        struct ccl_rpn_node *tmp = ccl_rpn_node_create(op);
+        tmp->u.p[0] = l;
+        tmp->u.p[1] = r;
+        tmp->u.p[2] = 0;
+        return tmp;
+    }
+    else if (r)
+        return r;
+    return l;
+}
+
+static struct ccl_rpn_node *ccl_rpn_dup(struct ccl_rpn_node *rpn)
+{
+    struct ccl_rpn_node *n;
+    struct ccl_rpn_attr *attr, **attrp;
+    if (!rpn)
+        return 0;
+    n = ccl_rpn_node_create(rpn->kind);
+    switch (rpn->kind)
+    {
+    case CCL_RPN_AND:
+    case CCL_RPN_OR:
+    case CCL_RPN_NOT:
+        n->u.p[0] = ccl_rpn_dup(rpn->u.p[0]);
+        n->u.p[1] = ccl_rpn_dup(rpn->u.p[1]);
+        break;
+    case CCL_RPN_TERM:
+        n->u.t.term = xstrdup(rpn->u.t.term);
+        n->u.t.qual = rpn->u.t.qual ? xstrdup(rpn->u.t.qual) : 0;
+        attrp = &n->u.t.attr_list;
+        for (attr = rpn->u.t.attr_list; attr; attr = attr->next)
+        {
+            *attrp = (struct ccl_rpn_attr *) xmalloc(sizeof(**attrp));
+            (*attrp)->kind = attr->kind;
+            (*attrp)->type = attr->type;
+            if (attr->kind == CCL_RPN_ATTR_STRING)
+                (*attrp)->value.str = xstrdup(attr->value.str);
+            else
+                (*attrp)->value.numeric = attr->value.numeric;
+            if (attr->set)
+                (*attrp)->set = xstrdup(attr->set);
+            else
+                (*attrp)->set = 0;
+            attrp = &(*attrp)->next;
+        }
+        *attrp = 0;
+        break;
+    case CCL_RPN_SET:
+        n->u.setname = xstrdup(rpn->u.setname);
+        break;
+    case CCL_RPN_PROX:
+        n->u.p[0] = ccl_rpn_dup(rpn->u.p[0]);
+        n->u.p[1] = ccl_rpn_dup(rpn->u.p[1]);
+        n->u.p[2] = ccl_rpn_dup(rpn->u.p[2]);
+        break;
+    }
+    return n;
+}
+
 /**
  * ccl_rpn_delete: Delete RPN tree.
  * rpn:   Pointer to tree.
@@ -159,7 +224,7 @@ static struct ccl_rpn_node *find_spec(CCL_parser cclp, ccl_qualifier_t *qa);
 
 static int is_term_ok(int look, int *list)
 {
-    for (;*list >= 0; list++)
+    for (; *list >= 0; list++)
         if (look == *list)
             return 1;
     return 0;
@@ -170,9 +235,7 @@ static struct ccl_rpn_node *search_terms(CCL_parser cclp, ccl_qualifier_t *qa);
 static struct ccl_rpn_attr *add_attr_node(struct ccl_rpn_node *p,
                                            const char *set, int type)
 {
-    struct ccl_rpn_attr *n;
-
-    n = (struct ccl_rpn_attr *)xmalloc(sizeof(*n));
+    struct ccl_rpn_attr *n = (struct ccl_rpn_attr *) xmalloc(sizeof(*n));
     ccl_assert(n);
     if (set)
         n->set = xstrdup(set);
@@ -181,7 +244,6 @@ static struct ccl_rpn_attr *add_attr_node(struct ccl_rpn_node *p,
     n->type = type;
     n->next = p->u.t.attr_list;
     p->u.t.attr_list = n;
-
     return n;
 }
 
@@ -195,9 +257,7 @@ static struct ccl_rpn_attr *add_attr_node(struct ccl_rpn_node *p,
 void ccl_add_attr_numeric(struct ccl_rpn_node *p, const char *set,
                           int type, int value)
 {
-    struct ccl_rpn_attr *n;
-
-    n = add_attr_node(p, set, type);
+    struct ccl_rpn_attr *n = add_attr_node(p, set, type);
     n->kind = CCL_RPN_ATTR_NUMERIC;
     n->value.numeric = value;
 }
@@ -205,9 +265,7 @@ void ccl_add_attr_numeric(struct ccl_rpn_node *p, const char *set,
 void ccl_add_attr_string(struct ccl_rpn_node *p, const char *set,
                          int type, char *value)
 {
-    struct ccl_rpn_attr *n;
-
-    n = add_attr_node(p, set, type);
+    struct ccl_rpn_attr *n = add_attr_node(p, set, type);
     n->kind = CCL_RPN_ATTR_STRING;
     n->value.str = xstrdup(value);
 }
@@ -328,9 +386,10 @@ static int append_term(CCL_parser cclp, const char *src_str, size_t src_len,
 
 
 static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp,
+                                             struct ccl_token *lookahead0,
                                              struct ccl_rpn_attr *attr_use,
                                              ccl_qualifier_t *qa,
-                                             size_t no, int term_len,
+                                             size_t no,
                                              int is_phrase,
                                              int auto_group)
 {
@@ -348,11 +407,12 @@ static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp,
     int z3958_trunc = 0;
     int is_ccl_masked = 0;
     char *attset;
-    struct ccl_token *lookahead = cclp->look_token;
+    struct ccl_token *lookahead = lookahead0;
     const char **truncation_aliases;
     const char *t_default[2];
     const char **mask_aliases;
     const char *m_default[2];
+    int term_len = 0;
 
     truncation_aliases =
         ccl_qual_search_special(cclp->bibset, "truncation");
@@ -362,7 +422,6 @@ static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp,
         t_default[0] = "?";
         t_default[1] = 0;
     }
-
     mask_aliases =
         ccl_qual_search_special(cclp->bibset, "mask");
     if (!mask_aliases)
@@ -371,17 +430,17 @@ static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp,
         m_default[0] = "#";
         m_default[1] = 0;
     }
-
-
     for (i = 0; i < no; i++)
     {
         if (has_ccl_masking(lookahead->name, lookahead->len,
                             truncation_aliases,
                             mask_aliases))
             is_ccl_masked = 1;
+
+        term_len += 1 + lookahead->len + lookahead->ws_prefix_len;
         lookahead = lookahead->next;
     }
-    lookahead = cclp->look_token;
+    lookahead = lookahead0;
 
     p = ccl_rpn_node_create(CCL_RPN_TERM);
     p->u.t.attr_list = NULL;
@@ -397,7 +456,7 @@ static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp,
     {
         struct ccl_rpn_attr *attr;
         for (attr = ccl_qual_get_attr(qa[i]); attr; attr = attr->next)
-            if (attr->type != 1 || attr == attr_use)
+            if (attr->type != 1 || !attr_use || attr == attr_use)
             {
                 switch (attr->kind)
                 {
@@ -540,6 +599,102 @@ static struct ccl_rpn_node *ccl_term_one_use(CCL_parser cclp,
     return p;
 }
 
+static struct ccl_rpn_node *ccl_term_multi_use(CCL_parser cclp,
+                                               struct ccl_token *lookahead0,
+                                               ccl_qualifier_t *qa,
+                                               size_t no,
+                                               int is_phrase,
+                                               int auto_group)
+{
+    struct ccl_rpn_node *p = 0;
+    int i;
+    for (i = 0; qa && qa[i]; i++)
+    {
+        struct ccl_rpn_attr *attr;
+        for (attr = ccl_qual_get_attr(qa[i]); attr; attr = attr->next)
+            if (attr->type == 1 && i == 0)
+            {
+                struct ccl_rpn_node *tmp2;
+                tmp2 = ccl_term_one_use(cclp, lookahead0,
+                                        attr, qa, no,
+                                        is_phrase, auto_group);
+                if (!tmp2)
+                {
+                    ccl_rpn_delete(p);
+                    return 0;
+                }
+                p = ccl_rpn_node_mkbool(p, tmp2, CCL_RPN_OR);
+            }
+    }
+    if (!p)
+        p = ccl_term_one_use(cclp, lookahead0,
+                             0 /* attr: no use */, qa, no,
+                             is_phrase, auto_group);
+    return p;
+}
+
+static struct ccl_rpn_node *split_recur(CCL_parser cclp, ccl_qualifier_t *qa,
+                                        struct ccl_rpn_node *parent,
+                                        struct ccl_token **ar, size_t sz)
+{
+    size_t l;
+    struct ccl_rpn_node *p_top = 0;
+    assert(sz > 0);
+    for (l = 1; l <= sz; l++)
+    {
+        struct ccl_rpn_node *p1;
+        struct ccl_rpn_node *p2 = ccl_term_multi_use(cclp, ar[0],
+                                                     qa, l,
+                                                     l > 1,
+                                                     /* auto_group */0);
+        if (!p2)
+            return 0;
+        if (parent)
+        {
+            struct ccl_rpn_node *tmp = ccl_rpn_node_create(CCL_RPN_AND);
+            tmp->u.p[0] = l > 1 ? ccl_rpn_dup(parent) : parent;
+            tmp->u.p[1] = p2;
+            p2 = tmp;
+        }
+        if (sz > l)
+            p1 = split_recur(cclp, qa, p2, ar + l, sz - l);
+        else
+            p1 = p2;
+        p_top = ccl_rpn_node_mkbool(p_top, p1, CCL_RPN_OR);
+    }
+    assert(p_top);
+    return p_top;
+}
+
+static struct ccl_rpn_node *search_term_split_list(CCL_parser cclp,
+                                                   ccl_qualifier_t *qa,
+                                                   int *term_list, int multi)
+{
+    struct ccl_rpn_node *p;
+    struct ccl_token **ar;
+    struct ccl_token *lookahead = cclp->look_token;
+    size_t i, sz;
+    for (sz = 0; is_term_ok(lookahead->kind, term_list); sz++)
+        lookahead = lookahead->next;
+    if (sz == 0)
+    {
+        cclp->error_code = CCL_ERR_TERM_EXPECTED;
+        return 0;
+    }
+    ar = (struct ccl_token **) xmalloc(sizeof(*lookahead) * sz);
+    lookahead = cclp->look_token;
+    for (i = 0; is_term_ok(lookahead->kind, term_list); i++)
+    {
+        ar[i] = lookahead;
+        lookahead = lookahead->next;
+    }
+    p = split_recur(cclp, qa, 0, ar, sz);
+    xfree(ar);
+    for (i = 0; i < sz; i++)
+        ADVANCE;
+    return p;
+}
+
 /**
  * search_term: Parse CCL search term.
  * cclp:   CCL Parser
@@ -564,11 +719,14 @@ static struct ccl_rpn_node *search_term_x(CCL_parser cclp,
         auto_group = 1;
     if (qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_OR_LIST, 0))
         or_list = 1;
+    if (qual_val_type(qa, CCL_BIB1_STR, CCL_BIB1_STR_SPLIT_LIST, 0))
+    {
+        return search_term_split_list(cclp, qa, term_list, multi);
+    }
     while (1)
     {
         struct ccl_rpn_node *p = 0;
         size_t no, i;
-        int len = 0;
         int is_phrase = 0;
         size_t max = 200;
         if (and_list || or_list || !multi)
@@ -595,70 +753,18 @@ static struct ccl_rpn_node *search_term_x(CCL_parser cclp,
             }
             else if (this_is_phrase || no > 0)
                 is_phrase = 1;
-            len += 1+lookahead->len+lookahead->ws_prefix_len;
             lookahead = lookahead->next;
         }
 
-        if (len == 0)
+        if (no == 0)
             break;      /* no more terms . stop . */
-
-        /* go through all attributes and add them to the attribute list */
-        for (i = 0; qa && qa[i]; i++)
-        {
-            struct ccl_rpn_attr *attr;
-
-            for (attr = ccl_qual_get_attr(qa[i]); attr; attr = attr->next)
-                if (attr->type == 1)
-                {
-                    struct ccl_rpn_node *tmp2;
-                    tmp2 = ccl_term_one_use(cclp, attr, qa, no, len,
-                                            is_phrase,
-                                            auto_group);
-                    if (!tmp2)
-                    {
-                        ccl_rpn_delete(p);
-                        return 0;
-                    }
-                    if (!p)
-                        p = tmp2;
-                    else
-                    {
-                        struct ccl_rpn_node *tmp1;
-                        tmp1 = ccl_rpn_node_create(CCL_RPN_OR);
-                        tmp1->u.p[0] = p;
-                        tmp1->u.p[1] = tmp2;
-                        p = tmp1;
-                    }
-                }
-        }
-        if (!p)
-        {
-            p = ccl_term_one_use(cclp, 0 /* attr: no use */, qa, no, len,
-                                 is_phrase, auto_group);
-            if (!p)
-                return 0;
-        }
+        p = ccl_term_multi_use(cclp, cclp->look_token, qa, no,
+                               is_phrase, auto_group);
         for (i = 0; i < no; i++)
             ADVANCE;
-        /* make the top node point to us.. */
-        if (p_top)
-        {
-            struct ccl_rpn_node *tmp;
-
-            if (or_list)
-                tmp = ccl_rpn_node_create(CCL_RPN_OR);
-            else if (and_list)
-                tmp = ccl_rpn_node_create(CCL_RPN_AND);
-            else
-                tmp = ccl_rpn_node_create(CCL_RPN_AND);
-            tmp->u.p[0] = p_top;
-            tmp->u.p[1] = p;
-
-            p_top = tmp;
-        }
-        else
-            p_top = p;
-
+        if (!p)
+            return 0;
+        p_top = ccl_rpn_node_mkbool(p_top, p, or_list ? CCL_RPN_OR : CCL_RPN_AND);
         if (!multi)
             break;
     }
@@ -703,7 +809,6 @@ static struct ccl_rpn_node *search_terms2(CCL_parser cclp,
 }
 
 
-
 static
 struct ccl_rpn_node *qualifiers_order(CCL_parser cclp,
                                       ccl_qualifier_t *ap, char *attset)
@@ -947,16 +1052,7 @@ static struct ccl_rpn_node *qualifier_list(CCL_parser cclp,
                     xfree(ap);
                     return 0;
                 }
-                if (node)
-                {
-                    struct ccl_rpn_node *node_this =
-                        ccl_rpn_node_create(CCL_RPN_OR);
-                    node_this->u.p[0] = node;
-                    node_this->u.p[1] = node_sub;
-                    node = node_this;
-                }
-                else
-                    node = node_sub;
+                node = ccl_rpn_node_mkbool(node, node_sub, CCL_RPN_OR);
                 seq++;
             }
             if (seq == 0)
@@ -1019,16 +1115,7 @@ static struct ccl_rpn_node *qualifier_list(CCL_parser cclp,
                 ccl_rpn_delete(node);
                 break;
             }
-            if (node)
-            {
-                struct ccl_rpn_node *node_this =
-                    ccl_rpn_node_create(CCL_RPN_OR);
-                node_this->u.p[0] = node;
-                node_this->u.p[1] = node_sub;
-                node = node_this;
-            }
-            else
-                node = node_sub;
+            node = ccl_rpn_node_mkbool(node, node_sub, CCL_RPN_OR);
             seq++;
         }
     }
@@ -1161,17 +1248,7 @@ static struct ccl_rpn_node *search_elements(CCL_parser cclp,
                 ccl_rpn_delete(node);
                 return 0;
             }
-            if (node)
-            {
-                struct ccl_rpn_node *node_this =
-                    ccl_rpn_node_create(CCL_RPN_OR);
-                node_this->u.p[0] = node;
-                node_this->u.p[1] = node_sub;
-                node_this->u.p[2] = 0;
-                node = node_this;
-            }
-            else
-                node = node_sub;
+            node = ccl_rpn_node_mkbool(node, node_sub, CCL_RPN_OR);
         }
         if (!node)
             node = search_terms(cclp, 0);
@@ -1187,7 +1264,7 @@ static struct ccl_rpn_node *search_elements(CCL_parser cclp,
  */
 static struct ccl_rpn_node *find_spec(CCL_parser cclp, ccl_qualifier_t *qa)
 {
-    struct ccl_rpn_node *p1, *p2, *pn;
+    struct ccl_rpn_node *p1, *p2;
     if (!(p1 = search_elements(cclp, qa)))
         return NULL;
     while (1)
@@ -1202,11 +1279,7 @@ static struct ccl_rpn_node *find_spec(CCL_parser cclp, ccl_qualifier_t *qa)
                 ccl_rpn_delete(p1);
                 return NULL;
             }
-            pn = ccl_rpn_node_create(CCL_RPN_AND);
-            pn->u.p[0] = p1;
-            pn->u.p[1] = p2;
-            pn->u.p[2] = 0;
-            p1 = pn;
+            p1 = ccl_rpn_node_mkbool(p1, p2, CCL_RPN_AND);
             continue;
         case CCL_TOK_OR:
             ADVANCE;
@@ -1216,11 +1289,7 @@ static struct ccl_rpn_node *find_spec(CCL_parser cclp, ccl_qualifier_t *qa)
                 ccl_rpn_delete(p1);
                 return NULL;
             }
-            pn = ccl_rpn_node_create(CCL_RPN_OR);
-            pn->u.p[0] = p1;
-            pn->u.p[1] = p2;
-            pn->u.p[2] = 0;
-            p1 = pn;
+            p1 = ccl_rpn_node_mkbool(p1, p2, CCL_RPN_OR);
             continue;
         case CCL_TOK_NOT:
             ADVANCE;
@@ -1230,11 +1299,7 @@ static struct ccl_rpn_node *find_spec(CCL_parser cclp, ccl_qualifier_t *qa)
                 ccl_rpn_delete(p1);
                 return NULL;
             }
-            pn = ccl_rpn_node_create(CCL_RPN_NOT);
-            pn->u.p[0] = p1;
-            pn->u.p[1] = p2;
-            pn->u.p[2] = 0;
-            p1 = pn;
+            p1 = ccl_rpn_node_mkbool(p1, p2, CCL_RPN_NOT);
             continue;
         }
         break;