--- trunk/pgest.c	2005/09/10 18:51:13	40
+++ trunk/pgest.c	2005/09/11 21:44:56	47
@@ -14,7 +14,7 @@
  *
  * Based on:
  * - C example from PostgreSQL documentation (BSD licence)
- * - example002.c from Hyper Estraier (GPL)
+ * - coreexample002.c and nodeexample002.c from Hyper Estraier (GPL)
  * - _textin/_textout from pgcurl.c (LGPL)
  *
  * This code is licenced under GPL
@@ -28,6 +28,7 @@
 #include "miscadmin.h"
 #include <estraier.h>
 #include <cabin.h>
+#include <estnode.h>
 
 #define _textin(str) DirectFunctionCall1(textin, CStringGetDatum(str))
 #define _textout(str) DatumGetPointer(DirectFunctionCall1(textout, PointerGetDatum(str)))
@@ -43,6 +44,7 @@
 
 /* prototype */
 char *attr2text(ESTDOC *doc, char *attr);
+char *node_attr2text(ESTRESDOC *rdoc, char *attr);
 
 
 /* work in progress */
@@ -244,7 +246,6 @@
 
 	elog(DEBUG1, "pgest_attr: found %d hits for %s", resnum, query);
 
-
 	values = (char **) palloc(ncols * sizeof(char *));
 
 	for (i = 0; i < nrows; i++)
@@ -252,7 +253,7 @@
 
 		/* get result from estraier */
 		if (! ( doc = est_db_get_doc(db, est_result[i + offset], 0)) ) {
-			elog(INFO, "can't find result %d", i + offset);
+			elog(INFO, "pgest_attr: can't find result %d", i + offset);
 		} else {
 			elog(DEBUG1, "URI: %s\n Title: %s\n",
 				est_doc_attr(doc, "@uri"),
@@ -284,9 +285,8 @@
 		/* now store it */
 		tuplestore_puttuple(tupstore, tuple);
 
-
 		/* delete estraier document object */
-		est_doc_delete(doc);
+		if (doc) est_doc_delete(doc);
 	}
 
 	tuplestore_donestoring(tupstore);
@@ -321,6 +321,8 @@
 	int len;
 	int attrlen;
 
+	if (! doc) return (Datum) NULL;
+
 	elog(DEBUG1, "doc: %08x, attr: %s", doc, attr);
 
 	if ( (attrval = est_doc_attr(doc, attr)) && (attrlen = strlen(attrval)) ) {
@@ -334,6 +336,341 @@
 
 	len++;
 	len *= sizeof(char);
+
+	elog(DEBUG2, "palloc(%d)", len);
+
+	val = palloc(len);
+
+	memset(val, 0, len);
+	strncpy(val, attrval, len);
+
+	elog(DEBUG2, "val=%s", val);
+
+	return val;
+}
+
+/*
+ * variation on theme: use node API which doesn't open index on
+ * every query which is much faster for large indexes
+ *
+ */
+
+PG_FUNCTION_INFO_V1(pgest_node);
+Datum pgest_node(PG_FUNCTION_ARGS)
+{
+	ArrayType	*attr_arr = PG_GETARG_ARRAYTYPE_P(8);
+	Oid		attr_element_type = ARR_ELEMTYPE(attr_arr);
+	int		attr_ndims = ARR_NDIM(attr_arr);
+	int		*attr_dim_counts = ARR_DIMS(attr_arr);
+	int		*attr_dim_lower_bounds = ARR_LBOUND(attr_arr);
+	int		ncols = 0;
+	int		nrows = 0;
+	int		indx[MAXDIM];
+	int16		attr_len;
+	bool		attr_byval;
+	char		attr_align;
+	ReturnSetInfo	*rsinfo = (ReturnSetInfo *) fcinfo->resultinfo;
+	AttInMetadata	*attinmeta;
+	TupleDesc	tupdesc;
+	Tuplestorestate *tupstore = NULL;
+	HeapTuple	tuple;
+	MemoryContext	per_query_ctx;
+	MemoryContext	oldcontext;
+	Datum		dvalue;
+	char		**values;
+	int		rsinfo_ncols;
+	int		i, j;
+	/* estvars */
+	ESTNODE *node;
+	ESTCOND *cond;
+	ESTNODERES *nres;
+	ESTRESDOC *rdoc;
+	const CBLIST *texts;
+	int resnum = 0;
+	int limit = 0;
+	int offset = 0;
+
+	char		*node_url;
+	char		*user, *passwd;
+	char		*query;
+	char		*attr;
+	char		*order;
+
+
+	/* only allow 1D input array */
+	if (attr_ndims == 1)
+	{
+		ncols = attr_dim_counts[0];
+	}
+	else
+		ereport(ERROR,
+				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+				 errmsg("invalid input array"),
+				 errdetail("Input array must have 1 dimension")));
+		
+	/* check to see if caller supports us returning a tuplestore */
+	if (!rsinfo || !(rsinfo->allowedModes & SFRM_Materialize))
+		ereport(ERROR,
+				(errcode(ERRCODE_SYNTAX_ERROR),
+				 errmsg("materialize mode required, but it is not " \
+						"allowed in this context")));
+
+	/* get info about element type needed to construct the array */
+	get_typlenbyvalalign(attr_element_type, &attr_len, &attr_byval, &attr_align);
+
+	/* get the requested return tuple description */
+	tupdesc = rsinfo->expectedDesc;
+	rsinfo_ncols = tupdesc->natts;
+
+	/*
+	 * The requested tuple description better match up with the array
+	 * we were given.
+	 */
+	if (rsinfo_ncols != ncols)
+		ereport(ERROR,
+				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+				 errmsg("invalid input array"),
+				 errdetail("Number of elements in array must match number of query specified columns.")));
+
+	/* OK, use it */
+	attinmeta = TupleDescGetAttInMetadata(tupdesc);
+
+	/* Now go to work */
+	rsinfo->returnMode = SFRM_Materialize;
+
+	per_query_ctx = fcinfo->flinfo->fn_mcxt;
+	oldcontext = MemoryContextSwitchTo(per_query_ctx);
+
+	/* initialize our tuplestore */
+	tupstore = tuplestore_begin_heap(true, false, SortMem);
+
+
+	/* take rest of arguments from function */
+
+	/* node URL */
+	if (PG_ARGISNULL(0)) {
+		ereport(ERROR,
+				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+				 errmsg("node URL can't be null"),
+				 errdetail("Node URL must be valid URL to HyperEstraier node")));
+	}
+	node_url = _textout(PG_GETARG_TEXT_P(0));
+
+	/* login and password */
+	if (PG_ARGISNULL(1) || PG_ARGISNULL(2)) {
+		ereport(ERROR,
+				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+				 errmsg("username and password can't be NULL"),
+				 errdetail("You must specify valid username and password to HyperEstraier node")));
+	}
+	user = _textout(PG_GETARG_TEXT_P(1));
+	passwd = _textout(PG_GETARG_TEXT_P(2));
+
+	/* query string */
+	if (PG_ARGISNULL(3)) {
+		query = "";
+	} else {
+		query = _textout(PG_GETARG_TEXT_P(3));
+	}
+
+	/* atribute filter */
+	if (PG_ARGISNULL(4)) {
+		attr = "";
+	} else {
+		attr = _textout(PG_GETARG_TEXT_P(4));
+	}
+	
+	/* sort order */
+	if (PG_ARGISNULL(5)) {
+		order = "";
+	} else {
+		order = _textout(PG_GETARG_TEXT_P(5));
+	}
+
+
+	/* limit */
+	if (PG_ARGISNULL(6)) {
+		limit = 0;
+	} else {
+		limit = PG_GETARG_INT32(6);
+	}
+
+	/* offset */
+	if (PG_ARGISNULL(7)) {
+		offset = 0;
+	} else {
+		offset = PG_GETARG_INT32(7);
+	}
+
+	/* initialize the network environment */
+	if(!est_init_net_env()){
+		ereport(ERROR, (errcode(ERRCODE_QUERY_CANCELED),
+			errmsg("pgest_node: can't create network enviroment")));
+	}
+
+	/* create the node connection object */
+	elog(DEBUG1, "pgest_node: est_node_new(%s) as %s", node_url, user);
+	node = est_node_new(node_url);
+	est_node_set_auth(node, user, passwd);
+
+	elog(DEBUG1, "pgest_node: query[%s] attr[%s] limit %d offset %d", query, (PG_ARGISNULL(4) ? "NULL" : attr), limit, offset);
+	
+	/* create a search condition object */
+	if (!(cond = est_cond_new())) {
+		ereport(ERROR, (errcode(ERRCODE_QUERY_CANCELED),
+			errmsg("pgest_node: est_cond_new failed")));
+	}
+	
+	/* set the search phrase to the search condition object */
+	if (! PG_ARGISNULL(3) && strlen(query) > 0)
+		est_cond_set_phrase(cond, query);
+
+	/* minimum valid attribute length is 10: @a STREQ a */
+	if (! PG_ARGISNULL(4) && strlen(attr) >= 10) {
+		elog(DEBUG1,"attributes: %s", attr);
+		char *curr_attr;
+		curr_attr = strtok(attr, ATTR_DELIMITER);
+		while (curr_attr) {
+			elog(DEBUG1,"est_cond_add_attr(%s)", curr_attr);
+			est_cond_add_attr(cond, curr_attr);
+			curr_attr = strtok(NULL, ATTR_DELIMITER);
+		}
+	}
+
+	/* set the search phrase to the search condition object */
+	if (! PG_ARGISNULL(5) && strlen(order) > 0) {
+		elog(DEBUG1,"est_cond_set_order(%s)", order);
+		est_cond_set_order(cond, order);
+	}
+
+	if (limit) {
+		elog(DEBUG1,"est_cond_set_max(%d)", limit + offset);
+		est_cond_set_max(cond, limit + offset);
+	}
+
+	/* get the result of search */
+	/* FIXME: allow user to specify depath of search */
+	nres = est_node_search(node, cond, 0);
+
+	if (! nres) {
+		int status = est_node_status(node);
+		est_cond_delete(cond);
+		est_node_delete(node);
+		est_free_net_env();
+		ereport(ERROR, (errcode(ERRCODE_QUERY_CANCELED),
+			errmsg("pgest_node: search failed, node status %d", status)));
+	}
+
+	/* get number of results */
+	resnum = est_noderes_doc_num(nres);
+
+	/* check if results exists */
+	if ( 0 == resnum ) {
+		elog(INFO, "pgest_node: no results for: %s", query );
+	}
+
+	/* total number of tuples to be returned */
+	if (limit && limit < resnum) {
+		nrows = limit;
+	} else {
+		nrows = resnum - offset;
+	}
+
+
+	elog(DEBUG1, "pgest_node: found %d hits for %s", resnum, query);
+
+
+	values = (char **) palloc(ncols * sizeof(char *));
+
+	for (i = 0; i < nrows; i++)
+	{
+
+		/* get result from estraier */
+		if (! ( rdoc = est_noderes_get_doc(nres, i + offset) )) {
+			elog(INFO, "pgest_node: can't find result %d", i + offset);
+		} else {
+			elog(DEBUG1, "URI: %s\n Title: %s\n",
+				est_resdoc_attr(rdoc, "@uri"),
+				est_resdoc_attr(rdoc, "@title")
+			);
+		}
+
+		/* iterate over results */
+		for (j = 0; j < ncols; j++)
+		{
+			bool	isnull;
+
+			/* array value of this position */
+			indx[0] = j + attr_dim_lower_bounds[0];
+
+			dvalue = array_ref(attr_arr, attr_ndims, indx, -1, attr_len, attr_byval, attr_align, &isnull);
+
+			if (!isnull && rdoc)
+				values[j] = DatumGetCString(
+					node_attr2text(rdoc,
+						(char *)DirectFunctionCall1(textout, dvalue)
+					));
+			else
+				values[j] = NULL;
+		}
+		/* construct the tuple */
+		tuple = BuildTupleFromCStrings(attinmeta, values);
+
+		/* now store it */
+		tuplestore_puttuple(tupstore, tuple);
+
+	}
+
+	tuplestore_donestoring(tupstore);
+	rsinfo->setResult = tupstore;
+
+	/*
+	 * SFRM_Materialize mode expects us to return a NULL Datum. The actual
+	 * tuples are in our tuplestore and passed back through
+	 * rsinfo->setResult. rsinfo->setDesc is set to the tuple description
+	 * that we actually used to build our tuples with, so the caller can
+	 * verify we did what it was expecting.
+	 */
+	rsinfo->setDesc = tupdesc;
+	MemoryContextSwitchTo(oldcontext);
+
+	/* delete the node result object */
+	est_noderes_delete(nres);
+
+	/* destroy the search condition object */                          
+	est_cond_delete(cond);
+
+	/* destroy the node object */
+	est_node_delete(node);
+
+	/* free the networking environment */
+	est_free_net_env();
+
+	return (Datum) 0;
+}
+
+/* make text var from node attr */
+char *node_attr2text(ESTRESDOC *rdoc, char *attr) {
+	char *val;
+	const char *attrval;
+	int len;
+	int attrlen;
+
+	if (! rdoc) return (Datum) NULL;
+
+	elog(DEBUG1, "doc: %08x, attr: %s", rdoc, attr);
+
+	if ( (attrval = est_resdoc_attr(rdoc, attr)) && (attrlen = strlen(attrval)) ) {
+		val = (char *) palloc(attrlen * sizeof(char));
+	} else {
+		return (Datum) NULL;
+	}
+
+	len = strlen(attrval);
+	elog(DEBUG1, "node_attr2text(%s) = '%s' %d bytes", attr, attrval, len);
+
+	len++;
+	len *= sizeof(char);
 
 	elog(DEBUG2, "palloc(%d)", len);