* Europagate, 1995
*
* $Log: cclfind.c,v $
- * Revision 1.12 1995/03/20 15:27:43 adam
+ * Revision 1.13 1995/04/17 09:31:42 adam
+ * Improved handling of qualifiers. Aliases or reserved words.
+ *
+ * Revision 1.12 1995/03/20 15:27:43 adam
* Minor changes.
*
* Revision 1.11 1995/02/23 08:31:59 adam
#include <ccl.h>
+/* current lookahead token */
static struct ccl_token *look_token;
+
+/* holds error no if error occur */
static int ccl_error;
+
+/* current bibset */
static CCL_bibset bibset;
+/* returns type of current lookahead */
#define KIND (look_token->kind)
+
+/* move one token forward */
#define ADVANCE look_token = look_token->next
-#define ADVX(x) x=(x)->next
-static struct ccl_rpn_attr *qual_val (struct ccl_rpn_attr *list, int type)
+/*
+ * qual_val_range: search for attribute of type with range
+ * qa: Attribute array
+ * type: Type of attribute to search for
+ * vmin: Lower bound of value to search for
+ * vmax: Upper bound of value to search for
+ * return: Return pointer to integer of attribute value found; NULL
+ * otherwise.
+ */
+static int *qual_val_range (struct ccl_rpn_attr **qa, int type,
+ int vmin, int vmax)
{
- while (list)
- {
- if (list->type == type)
- return list;
- list = list->next;
- }
+ int i;
+ struct ccl_rpn_attr *q;
+
+ if (!qa)
+ return NULL;
+ for (i = 0; (q=qa[i]); i++)
+ while (q)
+ {
+ if (q->type == type && q->value >= vmin && q->value <= vmax)
+ return &q->value;
+ q = q->next;
+ }
return NULL;
}
-static int qual_val_type (struct ccl_rpn_attr *list, int type, int value)
+/*
+ * qual_val_type: test for existance of attribute type/value pair.
+ * qa: Attribute array
+ * type: Type of attribute to search for
+ * value: Value of attribute to seach for
+ * return: 1 if found; 0 otherwise.
+ */
+static int qual_val_type (struct ccl_rpn_attr **qa, int type, int value)
{
- while (list)
- {
- if (list->type == type && list->value == value)
- return 1;
- list = list->next;
- }
+ int i;
+ struct ccl_rpn_attr *q;
+
+ if (!qa)
+ return 0;
+ for (i = 0; (q=qa[i]); i++)
+ while (q)
+ {
+ if (q->type == type && q->value == value)
+ return 1;
+ q = q->next;
+ }
return 0;
}
+/*
+ * strxcat: concatenate strings.
+ * n: Null-terminated Destination string
+ * src: Source string to be appended (not null-terminated)
+ * len: Length of source string.
+ */
static void strxcat (char *n, const char *src, int len)
{
while (*n)
*n = '\0';
}
+/*
+ * copy_token_name: Return copy of CCL token name
+ * tp: Pointer to token info.
+ * return: malloc(3) allocated copy of token name.
+ */
static char *copy_token_name (struct ccl_token *tp)
{
char *str = malloc (tp->len + 1);
return str;
}
+/*
+ * mk_node: Create RPN node.
+ * kind: Type of node.
+ * return: pointer to allocated node.
+ */
static struct ccl_rpn_node *mk_node (enum rpn_node_kind kind)
{
struct ccl_rpn_node *p;
return p;
}
+/*
+ * ccl_rpn_delete: Delete RPN tree.
+ * rpn: Pointer to tree.
+ */
void ccl_rpn_delete (struct ccl_rpn_node *rpn)
{
struct ccl_rpn_attr *attr, *attr1;
static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa);
static struct ccl_rpn_node *search_terms (struct ccl_rpn_attr **qa);
+/*
+ * add_attr: Add attribute (type/value) to RPN term node.
+ * p: RPN node of type term.
+ * type: Type of attribute
+ * value: Value of attribute
+ */
static void add_attr (struct ccl_rpn_node *p, int type, int value)
{
struct ccl_rpn_attr *n;
p->u.t.attr_list = n;
}
+/*
+ * search_term: Parse CCL search term.
+ * qa: Qualifier attributes already applied.
+ * return: pointer to node(s); NULL on error.
+ */
static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa)
{
struct ccl_rpn_node *p;
- struct ccl_rpn_attr *attr;
struct ccl_token *lookahead = look_token;
int len = 0;
int no, i;
int left_trunc = 0;
int right_trunc = 0;
int mid_trunc = 0;
+ int relation_value = -1;
+ int position_value = -1;
+ int structure_value = -1;
+ int truncation_value = -1;
+ int completeness_value = -1;
if (KIND != CCL_TOK_TERM)
{
ccl_error = CCL_ERR_TERM_EXPECTED;
return NULL;
}
+ /* create the term node, but wait a moment before adding the term */
+ p = mk_node (CCL_RPN_TERM);
+ p->u.t.attr_list = NULL;
+ p->u.t.term = NULL;
+
+ if (!qa)
+ {
+ /* no qualifier(s) applied. Use 'term' if it is defined */
+
+ qa = malloc (2*sizeof(*qa));
+ assert (qa);
+ qa[0] = ccl_qual_search (bibset, "term", 4);
+ qa[1] = NULL;
+ }
+
+ /* 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 = qa[i]; attr; attr = attr->next)
+ if (attr->value > 0)
+ { /* deal only with REAL attributes (positive) */
+ switch (attr->type)
+ {
+ case CCL_BIB1_REL:
+ if (relation_value != -1)
+ continue;
+ relation_value = attr->value;
+ break;
+ case CCL_BIB1_POS:
+ if (position_value != -1)
+ continue;
+ position_value = attr->value;
+ break;
+ case CCL_BIB1_STR:
+ if (structure_value != -1)
+ continue;
+ structure_value = attr->value;
+ break;
+ case CCL_BIB1_TRU:
+ if (truncation_value != -1)
+ continue;
+ truncation_value = attr->value;
+ break;
+ case CCL_BIB1_COM:
+ if (completeness_value != -1)
+ continue;
+ completeness_value = attr->value;
+ break;
+ }
+ add_attr (p, attr->type, attr->value);
+ }
+ }
+ /* go through each TERM token. If no truncation attribute is yet
+ met, then look for left/right truncation markers (?) and
+ set left_trunc/right_trunc/mid_trunc accordingly */
for (no = 0; lookahead->kind == CCL_TOK_TERM; no++)
{
for (i = 0; i<lookahead->len; i++)
- if (lookahead->name[i] == '?')
+ if (truncation_value == -1 && lookahead->name[i] == '?')
{
if (no == 0 && i == 0 && lookahead->len >= 1)
left_trunc = 1;
len += 1+lookahead->len;
lookahead = lookahead->next;
}
- p = mk_node (CCL_RPN_TERM);
+ /* len now holds the number of characters in the RPN term */
+ /* no holds the number of CCL tokens (1 or more) */
+
+ if (structure_value == -1 &&
+ qual_val_type (qa, CCL_BIB1_STR, CCL_BIB1_STR_WP))
+ { /* no structure attribute met. Apply either structure attribute
+ WORD or PHRASE depending on number of CCL tokens */
+ if (no == 1)
+ add_attr (p, CCL_BIB1_STR, 2);
+ else
+ add_attr (p, CCL_BIB1_STR, 1);
+ }
+
+ /* make the RPN token */
p->u.t.term = malloc (len);
assert (p->u.t.term);
- p->u.t.attr_list = NULL;
p->u.t.term[0] = '\0';
for (i = 0; i<no; i++)
{
strxcat (p->u.t.term, src_str, src_len);
ADVANCE;
}
- if (qa)
- {
- int i;
- for (i=0; qa[i]; i++)
- {
- struct ccl_rpn_attr *attr;
-
- for (attr = qa[i]; attr; attr = attr->next)
- if (attr->value > 0)
- add_attr (p, attr->type, attr->value);
- }
- attr = qa[0];
- }
- else
- attr = ccl_qual_search (bibset, "term", 4);
- if (attr && qual_val_type (attr, CCL_BIB1_STR, CCL_BIB1_STR_WP))
- {
- if (no == 1)
- add_attr (p, CCL_BIB1_STR, 2);
- else
- add_attr (p, CCL_BIB1_STR, 1);
- }
if (left_trunc && right_trunc)
{
- if (attr && !qual_val_type (attr, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_BOTH))
+ if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_BOTH))
{
ccl_error = CCL_ERR_TRUNC_NOT_BOTH;
- if (qa)
- free (qa);
+ free (qa);
ccl_rpn_delete (p);
return NULL;
}
}
else if (right_trunc)
{
- if (attr && !qual_val_type (attr, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_RIGHT))
+ if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_RIGHT))
{
ccl_error = CCL_ERR_TRUNC_NOT_RIGHT;
- if (qa)
- free (qa);
+ free (qa);
ccl_rpn_delete (p);
return NULL;
}
}
else if (left_trunc)
{
- if (attr && !qual_val_type (attr, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_LEFT))
+ if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_LEFT))
{
ccl_error = CCL_ERR_TRUNC_NOT_LEFT;
- if (qa)
- free (qa);
+ free (qa);
ccl_rpn_delete (p);
return NULL;
}
}
else
{
- if (attr && qual_val_type (attr, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_NONE))
+ if (qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_NONE))
add_attr (p, CCL_BIB1_TRU, 100);
}
return p;
}
+/*
+ * qualifiers: Parse CCL qualifiers and search terms.
+ * la: Token pointer to RELATION token.
+ * qa: Qualifier attributes already applied.
+ * return: pointer to node(s); NULL on error.
+ */
static struct ccl_rpn_node *qualifiers (struct ccl_token *la,
struct ccl_rpn_attr **qa)
{
struct ccl_token *lookahead = look_token;
struct ccl_rpn_attr **ap;
- int no = 1;
+ int no = 0;
int i, rel;
- struct ccl_rpn_attr *attr;
-
+#if 0
if (qa)
{
ccl_error = CCL_ERR_DOUBLE_QUAL;
return NULL;
}
+#endif
for (lookahead = look_token; lookahead != la; lookahead=lookahead->next)
no++;
- ap = malloc (no * sizeof(*ap));
+ if (qa)
+ for (i=0; qa[i]; i++)
+ no++;
+ ap = malloc ((no+1) * sizeof(*ap));
assert (ap);
- for (i=0; look_token != la; i++)
+ for (i = 0; look_token != la; i++)
{
ap[i] = ccl_qual_search (bibset, look_token->name, look_token->len);
if (!ap[i])
if (KIND == CCL_TOK_COMMA)
ADVANCE;
}
+ if (qa)
+ while (*qa)
+ ap[i++] = *qa++;
ap[i] = NULL;
- if (! (attr = qual_val (ap[0], CCL_BIB1_REL)) ||
- attr->value != CCL_BIB1_REL_ORDER)
+ if (!qual_val_type (ap, CCL_BIB1_REL, CCL_BIB1_REL_ORDER))
{
/* unordered relation */
struct ccl_rpn_node *p;
struct ccl_rpn_node *p;
ADVANCE; /* skip relation */
- if (KIND == CCL_TOK_TERM)
+ if (KIND == CCL_TOK_TERM && look_token->next->kind == CCL_TOK_MINUS)
{
struct ccl_rpn_node *p1;
- p1 = search_term (ap);
- if (KIND == CCL_TOK_MINUS)
+ if (!(p1 = search_term (ap)))
+ {
+ free (ap);
+ return NULL;
+ }
+ ADVANCE; /* skip '-' */
+ if (KIND == CCL_TOK_TERM) /* = term - term ? */
{
- ADVANCE; /* skip '-' */
- if (KIND == CCL_TOK_TERM) /* = term - term ? */
- {
- struct ccl_rpn_node *p2;
-
- p2 = search_term (ap);
- p = mk_node (CCL_RPN_AND);
- p->u.p[0] = p1;
- add_attr (p1, CCL_BIB1_REL, 4);
- p->u.p[1] = p2;
- add_attr (p2, CCL_BIB1_REL, 2);
- free (ap);
- return p;
- }
- else /* = term - */
- {
- add_attr (p1, CCL_BIB1_REL, 4);
- free (ap);
- return p1;
- }
+ struct ccl_rpn_node *p2;
+
+ if (!(p2 = search_term (ap)))
+ {
+ ccl_rpn_delete (p1);
+ free (ap);
+ return NULL;
+ }
+ p = mk_node (CCL_RPN_AND);
+ p->u.p[0] = p1;
+ add_attr (p1, CCL_BIB1_REL, 4);
+ p->u.p[1] = p2;
+ add_attr (p2, CCL_BIB1_REL, 2);
+ free (ap);
+ return p;
}
- else
+ else /* = term - */
{
- add_attr (p1, CCL_BIB1_REL, rel);
+ add_attr (p1, CCL_BIB1_REL, 4);
free (ap);
return p1;
}
else if (KIND == CCL_TOK_MINUS) /* = - term ? */
{
ADVANCE;
- p = search_term (ap);
+ if (!(p = search_term (ap)))
+ {
+ free (ap);
+ return NULL;
+ }
add_attr (p, CCL_BIB1_REL, 2);
free (ap);
return p;
}
+ else if (KIND == CCL_TOK_LP)
+ {
+ ADVANCE;
+ if (!(p = find_spec (ap)))
+ {
+ free (ap);
+ return NULL;
+ }
+ if (KIND != CCL_TOK_RP)
+ {
+ ccl_error = CCL_ERR_RP_EXPECTED;
+ ccl_rpn_delete (p);
+ free (ap);
+ return NULL;
+ }
+ ADVANCE;
+ free (ap);
+ return p;
+ }
+ else
+ {
+ if (!(p = search_terms (ap)))
+ {
+ free (ap);
+ return NULL;
+ }
+ add_attr (p, CCL_BIB1_REL, rel);
+ free (ap);
+ return p;
+ }
ccl_error = CCL_ERR_TERM_EXPECTED;
}
free (ap);
return NULL;
}
+/*
+ * search_terms: Parse CCL search terms - including proximity.
+ * qa: Qualifier attributes already applied.
+ * return: pointer to node(s); NULL on error.
+ */
static struct ccl_rpn_node *search_terms (struct ccl_rpn_attr **qa)
{
struct ccl_rpn_node *p1, *p2, *pn;
return p1;
}
+/*
+ * search_elements: Parse CCL search elements
+ * qa: Qualifier attributes already applied.
+ * return: pointer to node(s); NULL on error.
+ */
static struct ccl_rpn_node *search_elements (struct ccl_rpn_attr **qa)
{
struct ccl_rpn_node *p1;
return search_terms (qa);
}
+/*
+ * find_spec: Parse CCL find specification
+ * qa: Qualifier attributes already applied.
+ * return: pointer to node(s); NULL on error.
+ */
static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa)
{
struct ccl_rpn_node *p1, *p2, *pn;
return p1;
}
+/*
+ * ccl_find: Parse CCL find - token representation
+ * abibset: Bibset to be used for the parsing
+ * list: List of tokens
+ * error: Pointer to integer. Holds error no. on completion.
+ * pos: Pointer to char position. Holds approximate error position.
+ * return: RPN tree on successful completion; NULL otherwise.
+ */
struct ccl_rpn_node *ccl_find (CCL_bibset abibset, struct ccl_token *list,
int *error, const char **pos)
{
return p;
}
+/*
+ * ccl_find_str: Parse CCL find - string representation
+ * bibset: Bibset to be used for the parsing
+ * str: String to be parsed
+ * error: Pointer to integer. Holds error no. on completion.
+ * pos: Pointer to char position. Holds approximate error position.
+ * return: RPN tree on successful completion; NULL otherwise.
+ */
struct ccl_rpn_node *ccl_find_str (CCL_bibset bibset, const char *str,
int *error, int *pos)
{
* Europagate, 1995
*
* $Log: cclqual.c,v $
- * Revision 1.6 1995/02/23 08:32:00 adam
+ * Revision 1.7 1995/04/17 09:31:46 adam
+ * Improved handling of qualifiers. Aliases or reserved words.
+ *
+ * Revision 1.6 1995/02/23 08:32:00 adam
* Changed header.
*
* Revision 1.4 1995/02/14 19:55:12 adam
#include <ccl.h>
+/* Definition of CCL_bibset pointer */
struct ccl_qualifiers {
struct ccl_qualifier *list;
};
+/*
+ * ccl_qual_add: Add qualifier to Bibset. If qualifier already
+ * exists, then attributes are appendend to old
+ * definition.
+ * name: name of qualifier
+ * no: No of attribute type/value pairs.
+ * pairs: Attributes. pairs[0] first type, pair[1] first value,
+ * ... pair[2*no-2] last type, pair[2*no-1] last value.
+ */
void ccl_qual_add (CCL_bibset b, const char *name, int no, int *pairs)
{
struct ccl_qualifier *q;
*attrp = NULL;
}
+/*
+ * ccl_qual_mk: Make new (empty) bibset.
+ * return: empty bibset.
+ */
CCL_bibset ccl_qual_mk (void)
{
CCL_bibset b = malloc (sizeof(*b));
return b;
}
+/*
+ * ccl_qual_rm: Delete bibset.
+ * b: pointer to bibset
+ */
void ccl_qual_rm (CCL_bibset *b)
{
- assert (*b);
+ struct ccl_qualifier *q, *q1;
+
+ if (!*b)
+ return;
+ for (q = (*b)->list; q; q = q1)
+ {
+ struct ccl_rpn_attr *attr, *attr1;
+
+ for (attr = q->attr_list; attr; attr = attr1)
+ {
+ attr1 = attr->next;
+ free (attr);
+ }
+ q1 = q->next;
+ free (q);
+ }
+ free (*b);
*b = NULL;
}
+/*
+ * ccl_qual_search: Search for qualifier in bibset.
+ * b: Bibset
+ * name: Name of qualifier to search for (need no null-termination)
+ * len: Length of name.
+ * return: Attribute info. NULL if not found.
+ */
struct ccl_rpn_attr *ccl_qual_search (CCL_bibset b, const char *name, int len)
{
struct ccl_qualifier *q;
return NULL;
}
-void ccl_qual_file (CCL_bibset bibset, FILE *inf)
-{
- char line[256];
- char *cp;
- char qual_name[128];
- char qual_des[128];
- int no_scan;
-
- while (fgets (line, 255, inf))
- {
- cp = line;
- if (*cp == '#')
- continue;
- if (sscanf (cp, "%s%n", qual_name, &no_scan) != 1)
- continue;
- cp += no_scan;
- while (1)
- {
- int pair[2];
- char *qual_type;
- char *qual_value;
- char *split;
-
- if (sscanf (cp, "%s%n", qual_des, &no_scan) != 1)
- break;
-
- if (!(split = strchr (qual_des, '=')))
- break;
- cp += no_scan;
-
- *split++ = '\0';
- qual_type = qual_des;
- qual_value = split;
- while (1)
- {
- if ((split = strchr (qual_value, ',')))
- *split++ = '\0';
- pair[1] = atoi (qual_value);
- switch (qual_type[0])
- {
- case 'u':
- pair[0] = CCL_BIB1_USE;
- break;
- case 'r':
- pair[0] = CCL_BIB1_REL;
- if (!strcmp (qual_value, "o"))
- pair[1] = CCL_BIB1_REL_ORDER;
- break;
- case 'p':
- pair[0] = CCL_BIB1_POS;
- break;
- case 's':
- pair[0] = CCL_BIB1_STR;
- if (!strcmp (qual_value, "pw"))
- pair[1] = CCL_BIB1_STR_WP;
- break;
- case 't':
- pair[0] = CCL_BIB1_TRU;
- if (!strcmp (qual_value, "l"))
- pair[1] = CCL_BIB1_TRU_CAN_LEFT;
- else if (!strcmp (qual_value, "r"))
- pair[1] = CCL_BIB1_TRU_CAN_RIGHT;
- else if (!strcmp (qual_value, "b"))
- pair[1] = CCL_BIB1_TRU_CAN_BOTH;
- else if (!strcmp (qual_value, "n"))
- pair[1] = CCL_BIB1_TRU_CAN_NONE;
- break;
- case 'c':
- pair[0] = CCL_BIB1_COM;
- break;
- default:
- pair[0] = atoi (qual_type);
- }
- ccl_qual_add (bibset, qual_name, 1, pair);
- if (!split)
- break;
- qual_value = split;
- }
- }
- }
-}
* Europagate, 1995
*
* $Log: ccltoken.c,v $
- * Revision 1.5 1995/02/23 08:32:00 adam
+ * Revision 1.6 1995/04/17 09:31:48 adam
+ * Improved handling of qualifiers. Aliases or reserved words.
+ *
+ * Revision 1.5 1995/02/23 08:32:00 adam
* Changed header.
*
* Revision 1.3 1995/02/15 17:42:16 adam
#include <ccl.h>
-static int strin (const char *s, const char *cset)
+const char *ccl_token_and = "and";
+const char *ccl_token_or = "or";
+const char *ccl_token_not = "not andnot";
+const char *ccl_token_set = "set";
+
+/*
+ * token_cmp: Compare token with keyword(s)
+ * kw: Keyword list. Each keyword is separated by space.
+ * token: CCL token.
+ * return: 1 if token string matches one of the keywords in list;
+ * 0 otherwise.
+ */
+static int token_cmp (const char *kw, struct ccl_token *token)
{
- while (*cset)
+ const char *cp1 = kw;
+ const char *cp2;
+ while ((cp2 = strchr (cp1, ' ')))
{
- if (*cset++ == *s)
+ if (token->len == cp2-cp1 &&
+ !memcmp (cp1, token->name, token->len))
return 1;
+ cp1 = cp2+1;
}
- return 0;
+ return token->len == strlen(cp1)
+ && !memcmp (cp1, token->name, token->len);
}
-const char *ccl_token_and = "and";
-const char *ccl_token_or = "or";
-const char *ccl_token_not = "not";
-const char *ccl_token_set = "set";
-
+/*
+ * ccl_tokenize: tokenize CCL command string.
+ * return: CCL token list.
+ */
struct ccl_token *ccl_tokenize (const char *command)
{
const char *cp = command;
while (1)
{
- while (*cp && strin (cp, " \t\r\n"))
+ while (*cp && strchr (" \t\r\n", *cp))
{
cp++;
continue;
cp++;
break;
default:
- while (*cp && !strin (cp, "(),%!><=- \t\n\r"))
+ while (*cp && !strchr ("(),%!><=- \t\n\r", *cp))
{
cp++;
++ last->len;
}
- if (strlen (ccl_token_and)==last->len &&
- !memcmp (ccl_token_and, last->name, last->len))
- last->kind = CCL_TOK_AND;
- else if (strlen (ccl_token_or)==last->len &&
- !memcmp (ccl_token_or, last->name, last->len))
- last->kind = CCL_TOK_OR;
- else if (strlen (ccl_token_not)==last->len &&
- !memcmp (ccl_token_not, last->name, last->len))
- last->kind = CCL_TOK_NOT;
- else if (strlen (ccl_token_set)==last->len &&
- !memcmp (ccl_token_set, last->name, last->len))
- last->kind = CCL_TOK_SET;
+ if (token_cmp (ccl_token_and, last))
+ last->kind = CCL_TOK_AND;
+ else if (token_cmp (ccl_token_or, last))
+ last->kind = CCL_TOK_OR;
+ else if (token_cmp (ccl_token_not, last))
+ last->kind = CCL_TOK_NOT;
+ else if (token_cmp (ccl_token_set, last))
+ last->kind = CCL_TOK_SET;
else
last->kind = CCL_TOK_TERM;
}
projects / egate.git / commitdiff