pgestraier/trunk/pgest.c

/*
 * integrate Hyper Estraier into PostgreSQL
 *
 * Dobrica Pavlinusic <dpavlin@rot13.org> 2005-05-19
 *
 * TODO:
 * - all
 *
 * NOTES:
 * - clear structures with memset to support hash indexes (who whould like
 *   to create hash index on table returned from function?)
 * - number of returned rows is set by PostgreSQL evaluator, see:
 *   http://archives.postgresql.org/pgsql-hackers/2005-02/msg00546.php
 *
 * Based on:
 * - C example from PostgreSQL documentation (BSD licence)
 * - example002.c from Hyper Estraier (GPL)
 * - _textin/_textout from pgcurl.c (LGPL)
 *
 * This code is licenced under GPL
 */

#include "postgres.h"
#include "fmgr.h"
#include "funcapi.h"
#include "utils/builtins.h"
#include "utils/array.h"
#include "miscadmin.h"
#include <estraier.h>
#include <cabin.h>

#define _textin(str) DirectFunctionCall1(textin, CStringGetDatum(str))
#define _textout(str) DatumGetPointer(DirectFunctionCall1(textout, PointerGetDatum(str)))
#define GET_STR(textp) DatumGetCString(DirectFunctionCall1(textout, PointerGetDatum(textp)))
#define GET_TEXT(cstrp) DatumGetTextP(DirectFunctionCall1(textin, CStringGetDatum(cstrp)))

/* prototype */
char *attr2text(ESTDOC *doc, char *attr);

ESTDB *db;
ESTCOND *cond;
ESTDOC *doc;
const CBLIST *texts;
int ecode, *est_result, resnum, i, j;
int limit = 0;
int offset = 0;

/* define PostgreSQL v1 function */
PG_FUNCTION_INFO_V1(pgest);
Datum pgest(PG_FUNCTION_ARGS) {

        FuncCallContext *funcctx;
        int             call_cntr;
        int             max_calls;
        TupleDesc       tupdesc;
        TupleTableSlot  *slot;
        AttInMetadata   *attinmeta;
        char            *index_path;
        char            *query;
        char            *attr;

        /* stuff done only on the first call of the function */
        if (SRF_IS_FIRSTCALL()) {
                MemoryContext   oldcontext;

                /* take arguments from function */
                //index_path = _textout(PG_GETARG_TEXT_P(0));
                index_path = _textout(PG_GETARG_TEXT_P(0));
                query = _textout(PG_GETARG_TEXT_P(1));
                attr = _textout(PG_GETARG_TEXT_P(2));
                limit = PG_GETARG_INT32(3);
                offset = PG_GETARG_INT32(4);

                /* create a function context for cross-call persistence */
                funcctx = SRF_FIRSTCALL_INIT();

                /* switch to memory context appropriate for multiple function calls */
                oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);

                /* open the database */
                elog(DEBUG1, "pgest: est_db_open(%s)", index_path);
                
                if(!(db = est_db_open(index_path, ESTDBREADER, &ecode))){
                        elog(ERROR, "est_db_open: can't open %s [%d]: %s", index_path, ecode, est_err_msg(ecode));
                        SRF_RETURN_DONE(funcctx);
                }
                
                elog(INFO, "pgest: query[%s] attr[%s] limit %d offset %d", query, attr, limit, offset);
                
                /* create a search condition object */
                if (!(cond = est_cond_new())) {
                        elog(INFO, "pgest: est_cond_new failed");
                        SRF_RETURN_DONE(funcctx);
                }
                
                /* set the search phrase to the search condition object */
                est_cond_set_phrase(cond, query);

                /* minimum valid attribute length is 10: @a STREQ a */
                if (attr != NULL && strlen(attr) >= 10) {
                        elog(INFO,"est_cond_add_attr(%s)", attr);
                        est_cond_add_attr(cond, attr);
                }

                /* get the result of search */
                est_result = est_db_search(db, cond, &resnum, NULL);
                
                /* total number of tuples to be returned */
                if (limit && limit < resnum) {
                        funcctx->max_calls = limit - offset;
                } else {
                        funcctx->max_calls = resnum - offset;
                }

                /* check if results exists */
                if ( 0 == funcctx->max_calls )
                        elog(INFO, "pgest: no results for: %s", query );

                elog(DEBUG1, "pgest: found %d hits for %s", resnum, query);

                /* Build a tuple description for a __pgest tuple */
                tupdesc = RelationNameGetTupleDesc("__pgest");

                /* allocate a slot for a tuple with this tupdesc */
                slot = TupleDescGetSlot(tupdesc);

                /* assign slot to function context */
                funcctx->slot = slot;

                /*
                 * generate attribute metadata needed later to produce tuples from raw
                 * C strings
                 */
                attinmeta = TupleDescGetAttInMetadata(tupdesc);
                funcctx->attinmeta = attinmeta;

                MemoryContextSwitchTo(oldcontext);

                elog(DEBUG1, "SRF_IS_FIRSTCALL done");
        }

        /* stuff done on every call of the function */
        funcctx = SRF_PERCALL_SETUP();

        call_cntr = funcctx->call_cntr;
        max_calls = funcctx->max_calls;
        slot = funcctx->slot;
        attinmeta = funcctx->attinmeta;

        if (limit && call_cntr > limit - 1) {
                elog(INFO, "call_cntr: %d limit: %d", call_cntr, limit);
                SRF_RETURN_DONE(funcctx);
        }

        if (call_cntr < max_calls) {
                char            **values;
                HeapTuple       tuple;
                Datum           result;

                elog(DEBUG1, "pgest: loop count %d", call_cntr);

                if (! est_result) {
                        elog(ERROR, "pgest: no estraier results");
                        SRF_RETURN_DONE(funcctx);
                }
                
                /*
                 * Prepare a values array for storage in our slot.
                 * This should be an array of C strings which will
                 * be processed later by the type input functions.
                 */

                if (doc = est_db_get_doc(db, est_result[call_cntr + offset], 0)) {
                
                        elog(DEBUG1, "URI: %s\n Title: %s\n",
                                est_doc_attr(doc, "@uri"),
                                est_doc_attr(doc, "@title")
                        );

                        values = (char **) palloc(4 * sizeof(char *));

//                      values[0] = (char *) palloc(strlen(_estval) * sizeof(char));

                        values[0] = (char *) attr2text(doc,"@id");
                        values[1] = (char *) attr2text(doc,"@uri");
                        values[2] = (char *) attr2text(doc,"@title");
                        values[3] = (char *) attr2text(doc,"@type");

                        /* destloy the document object */
                        elog(DEBUG2, "est_doc_delete");
                        est_doc_delete(doc);
                } else {
                        elog(INFO, "no result from estraier");
                        values[0] = NULL;
                        values[1] = NULL;
                        values[2] = NULL;
                        values[3] = NULL;
                }


                elog(DEBUG2, "build tuple");
                /* build a tuple */
                tuple = BuildTupleFromCStrings(attinmeta, values);

                elog(DEBUG2, "make tuple into datum");
                /* make the tuple into a datum */
                result = TupleGetDatum(slot, tuple);

                elog(DEBUG2, "cleanup");
                /* clean up ? */
/*
                pfree(values[0]);
                pfree(values[1]);
                pfree(values[2]);
                pfree(values[3]);
                pfree(values);
*/
                
                elog(DEBUG2, "cleanup over");
        
                SRF_RETURN_NEXT(funcctx, result);
        } else {
                elog(DEBUG1, "loop over");

                if(!est_db_close(db, &ecode)){
                        elog(INFO, "est_db_close error: %s", est_err_msg(ecode));
                }

                /* do when there is no more left */
                SRF_RETURN_DONE(funcctx);
        }
}

/* work in progress */
PG_FUNCTION_INFO_V1(pgest2);
Datum pgest2(PG_FUNCTION_ARGS)
{
        int             nrows = 3;
        int16           typlen;
        bool            typbyval;
        char            typalign;
        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             ncols;
        int             i, j;

        /* 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 the requested return tuple description */
        tupdesc = rsinfo->expectedDesc;
        ncols = tupdesc->natts;

        /*
         * The requested tuple description better match up with the array
         * we were given.
         */
        /* 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);

        values = (char **) palloc(ncols * sizeof(char *));

        for (i = 0; i < nrows; i++)
        {
                for (j = 0; j < ncols; j++)
                {
                        values[j] = DatumGetCString( "foo" );
                }
                /* 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);

        return (Datum) 0;
}


/* make text var from attr */
char *attr2text(ESTDOC *doc, char *attr) {
        char *val;
        const char *attrval;
        int len;
        int attrlen;

        elog(DEBUG1, "doc: %08x, attr: %s", doc, attr);

        if ( (attrval = est_doc_attr(doc, attr)) && (attrlen = strlen(attrval)) ) {
                val = (char *) palloc(attrlen * sizeof(char));
        } else {
                return (Datum) NULL;
        }

        len = strlen(attrval);
        elog(DEBUG1, "attr2text(%s) = '%s' %d bytes", attr, attrval, len);

        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;
}

/* make integer variable from property */
/*
char *prop2int(SW_RESULT sw_res, char *propname) {
        char *val;
        unsigned long prop;
        int len;

        elog(DEBUG2, "prop2int(%s)", propname);

        prop = estResultPropertyULong( sw_res, propname );
        if (error_or_abort( est_handle )) return NULL;

        elog(DEBUG1, "prop2int(%s) = %lu", propname, prop);

        len = 128 * sizeof(char);
        elog(DEBUG2, "palloc(%d)", len);

        val = palloc(len);
        memset(val, 0, len);

        snprintf(val, len, "%lu", prop);

        elog(DEBUG2, "val=%s", val);

        return val;
}
*/
1	dpavlin	1	/*
2			* integrate Hyper Estraier into PostgreSQL
3			*
4			* Dobrica Pavlinusic <dpavlin@rot13.org> 2005-05-19
5			*
6			* TODO:
7			* - all
8			*
9			* NOTES:
10			* - clear structures with memset to support hash indexes (who whould like
11			* to create hash index on table returned from function?)
12			* - number of returned rows is set by PostgreSQL evaluator, see:
13			* http://archives.postgresql.org/pgsql-hackers/2005-02/msg00546.php
14			*
15			* Based on:
16			* - C example from PostgreSQL documentation (BSD licence)
17			* - example002.c from Hyper Estraier (GPL)
18			* - _textin/_textout from pgcurl.c (LGPL)
19			*
20			* This code is licenced under GPL
21			*/
22
23			#include "postgres.h"
24			#include "fmgr.h"
25			#include "funcapi.h"
26			#include "utils/builtins.h"
27			#include "utils/array.h"
28			#include "miscadmin.h"
29			#include <estraier.h>
30			#include <cabin.h>
31
32			#define _textin(str) DirectFunctionCall1(textin, CStringGetDatum(str))
33			#define _textout(str) DatumGetPointer(DirectFunctionCall1(textout, PointerGetDatum(str)))
34			#define GET_STR(textp) DatumGetCString(DirectFunctionCall1(textout, PointerGetDatum(textp)))
35			#define GET_TEXT(cstrp) DatumGetTextP(DirectFunctionCall1(textin, CStringGetDatum(cstrp)))
36
37	dpavlin	5	/* prototype */
38			char attr2text(ESTDOC doc, char *attr);
39	dpavlin	1
40			ESTDB *db;
41			ESTCOND *cond;
42			ESTDOC *doc;
43			const CBLIST *texts;
44			int ecode, *est_result, resnum, i, j;
45	dpavlin	3	int limit = 0;
46			int offset = 0;
47	dpavlin	1
48			/* define PostgreSQL v1 function */
49			PG_FUNCTION_INFO_V1(pgest);
50			Datum pgest(PG_FUNCTION_ARGS) {
51
52			FuncCallContext *funcctx;
53			int call_cntr;
54			int max_calls;
55			TupleDesc tupdesc;
56			TupleTableSlot *slot;
57			AttInMetadata *attinmeta;
58			char *index_path;
59			char *query;
60	dpavlin	5	char *attr;
61	dpavlin	1
62			/* stuff done only on the first call of the function */
63			if (SRF_IS_FIRSTCALL()) {
64			MemoryContext oldcontext;
65
66			/* take arguments from function */
67			//index_path = _textout(PG_GETARG_TEXT_P(0));
68			index_path = _textout(PG_GETARG_TEXT_P(0));
69			query = _textout(PG_GETARG_TEXT_P(1));
70	dpavlin	5	attr = _textout(PG_GETARG_TEXT_P(2));
71			limit = PG_GETARG_INT32(3);
72			offset = PG_GETARG_INT32(4);
73	dpavlin	1
74			/* create a function context for cross-call persistence */
75			funcctx = SRF_FIRSTCALL_INIT();
76
77			/* switch to memory context appropriate for multiple function calls */
78			oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
79
80			/* open the database */
81			elog(DEBUG1, "pgest: est_db_open(%s)", index_path);
82
83			if(!(db = est_db_open(index_path, ESTDBREADER, &ecode))){
84			elog(ERROR, "est_db_open: can't open %s [%d]: %s", index_path, ecode, est_err_msg(ecode));
85			SRF_RETURN_DONE(funcctx);
86			}
87
88	dpavlin	5	elog(INFO, "pgest: query[%s] attr[%s] limit %d offset %d", query, attr, limit, offset);
89	dpavlin	1
90			/* create a search condition object */
91			if (!(cond = est_cond_new())) {
92			elog(INFO, "pgest: est_cond_new failed");
93			SRF_RETURN_DONE(funcctx);
94			}
95
96			/* set the search phrase to the search condition object */
97			est_cond_set_phrase(cond, query);
98
99	dpavlin	5	/* minimum valid attribute length is 10: @a STREQ a */
100			if (attr != NULL && strlen(attr) >= 10) {
101			elog(INFO,"est_cond_add_attr(%s)", attr);
102			est_cond_add_attr(cond, attr);
103			}
104
105	dpavlin	1	/* get the result of search */
106			est_result = est_db_search(db, cond, &resnum, NULL);
107
108			/* total number of tuples to be returned */
109	dpavlin	5	if (limit && limit < resnum) {
110			funcctx->max_calls = limit - offset;
111			} else {
112			funcctx->max_calls = resnum - offset;
113			}
114	dpavlin	1
115			/* check if results exists */
116			if ( 0 == funcctx->max_calls )
117			elog(INFO, "pgest: no results for: %s", query );
118
119	dpavlin	5	elog(DEBUG1, "pgest: found %d hits for %s", resnum, query);
120	dpavlin	1
121			/* Build a tuple description for a __pgest tuple */
122			tupdesc = RelationNameGetTupleDesc("__pgest");
123
124			/* allocate a slot for a tuple with this tupdesc */
125			slot = TupleDescGetSlot(tupdesc);
126
127			/* assign slot to function context */
128			funcctx->slot = slot;
129
130			/*
131			* generate attribute metadata needed later to produce tuples from raw
132			* C strings
133			*/
134			attinmeta = TupleDescGetAttInMetadata(tupdesc);
135			funcctx->attinmeta = attinmeta;
136
137			MemoryContextSwitchTo(oldcontext);
138
139	dpavlin	2	elog(DEBUG1, "SRF_IS_FIRSTCALL done");
140	dpavlin	1	}
141
142			/* stuff done on every call of the function */
143			funcctx = SRF_PERCALL_SETUP();
144
145			call_cntr = funcctx->call_cntr;
146			max_calls = funcctx->max_calls;
147			slot = funcctx->slot;
148			attinmeta = funcctx->attinmeta;
149	dpavlin	3
150			if (limit && call_cntr > limit - 1) {
151			elog(INFO, "call_cntr: %d limit: %d", call_cntr, limit);
152			SRF_RETURN_DONE(funcctx);
153			}
154
155	dpavlin	1	if (call_cntr < max_calls) {
156			char **values;
157			HeapTuple tuple;
158			Datum result;
159
160	dpavlin	2	elog(DEBUG1, "pgest: loop count %d", call_cntr);
161	dpavlin	1
162			if (! est_result) {
163			elog(ERROR, "pgest: no estraier results");
164			SRF_RETURN_DONE(funcctx);
165			}
166
167			/*
168			* Prepare a values array for storage in our slot.
169			* This should be an array of C strings which will
170			* be processed later by the type input functions.
171			*/
172
173	dpavlin	3	if (doc = est_db_get_doc(db, est_result[call_cntr + offset], 0)) {
174	dpavlin	1
175	dpavlin	2	elog(DEBUG1, "URI: %s\n Title: %s\n",
176	dpavlin	1	est_doc_attr(doc, "@uri"),
177			est_doc_attr(doc, "@title")
178			);
179
180			values = (char *) palloc(4 sizeof(char *));
181
182			// values[0] = (char ) palloc(strlen(_estval) sizeof(char));
183
184	dpavlin	5	values[0] = (char *) attr2text(doc,"@id");
185			values[1] = (char *) attr2text(doc,"@uri");
186			values[2] = (char *) attr2text(doc,"@title");
187			values[3] = (char *) attr2text(doc,"@type");
188	dpavlin	1
189			/* destloy the document object */
190			elog(DEBUG2, "est_doc_delete");
191			est_doc_delete(doc);
192			} else {
193			elog(INFO, "no result from estraier");
194			values[0] = NULL;
195			values[1] = NULL;
196			values[2] = NULL;
197			values[3] = NULL;
198			}
199
200
201			elog(DEBUG2, "build tuple");
202			/* build a tuple */
203			tuple = BuildTupleFromCStrings(attinmeta, values);
204
205			elog(DEBUG2, "make tuple into datum");
206			/* make the tuple into a datum */
207			result = TupleGetDatum(slot, tuple);
208
209			elog(DEBUG2, "cleanup");
210			/* clean up ? */
211			/*
212			pfree(values[0]);
213			pfree(values[1]);
214			pfree(values[2]);
215			pfree(values[3]);
216			pfree(values);
217			*/
218
219			elog(DEBUG2, "cleanup over");
220
221			SRF_RETURN_NEXT(funcctx, result);
222			} else {
223	dpavlin	2	elog(DEBUG1, "loop over");
224	dpavlin	1
225			if(!est_db_close(db, &ecode)){
226			elog(INFO, "est_db_close error: %s", est_err_msg(ecode));
227			}
228
229			/* do when there is no more left */
230			SRF_RETURN_DONE(funcctx);
231			}
232			}
233
234			/* work in progress */
235			PG_FUNCTION_INFO_V1(pgest2);
236			Datum pgest2(PG_FUNCTION_ARGS)
237			{
238			int nrows = 3;
239			int16 typlen;
240			bool typbyval;
241			char typalign;
242			ReturnSetInfo rsinfo = (ReturnSetInfo ) fcinfo->resultinfo;
243			AttInMetadata *attinmeta;
244			TupleDesc tupdesc;
245			Tuplestorestate *tupstore = NULL;
246			HeapTuple tuple;
247			MemoryContext per_query_ctx;
248			MemoryContext oldcontext;
249			Datum dvalue;
250			char **values;
251			int ncols;
252			int i, j;
253
254			/* check to see if caller supports us returning a tuplestore */
255			if (!rsinfo \|\| !(rsinfo->allowedModes & SFRM_Materialize))
256			ereport(ERROR,
257			(errcode(ERRCODE_SYNTAX_ERROR),
258			errmsg("materialize mode required, but it is not " \
259			"allowed in this context")));
260
261			/* get the requested return tuple description */
262			tupdesc = rsinfo->expectedDesc;
263			ncols = tupdesc->natts;
264
265			/*
266			* The requested tuple description better match up with the array
267			* we were given.
268			*/
269			/* OK, use it */
270			attinmeta = TupleDescGetAttInMetadata(tupdesc);
271
272			/* Now go to work */
273			rsinfo->returnMode = SFRM_Materialize;
274
275			per_query_ctx = fcinfo->flinfo->fn_mcxt;
276			oldcontext = MemoryContextSwitchTo(per_query_ctx);
277
278			/* initialize our tuplestore */
279			tupstore = tuplestore_begin_heap(true, false, SortMem);
280
281			values = (char *) palloc(ncols sizeof(char *));
282
283			for (i = 0; i < nrows; i++)
284			{
285			for (j = 0; j < ncols; j++)
286			{
287			values[j] = DatumGetCString( "foo" );
288			}
289			/* construct the tuple */
290			tuple = BuildTupleFromCStrings(attinmeta, values);
291
292			/* now store it */
293			tuplestore_puttuple(tupstore, tuple);
294			}
295
296			tuplestore_donestoring(tupstore);
297			rsinfo->setResult = tupstore;
298
299			/*
300			* SFRM_Materialize mode expects us to return a NULL Datum. The actual
301			* tuples are in our tuplestore and passed back through
302			* rsinfo->setResult. rsinfo->setDesc is set to the tuple description
303			* that we actually used to build our tuples with, so the caller can
304			* verify we did what it was expecting.
305			*/
306			rsinfo->setDesc = tupdesc;
307			MemoryContextSwitchTo(oldcontext);
308
309			return (Datum) 0;
310			}
311
312
313			/* make text var from attr */
314			char attr2text(ESTDOC doc, char *attr) {
315			char *val;
316			const char *attrval;
317			int len;
318	dpavlin	4	int attrlen;
319	dpavlin	1
320	dpavlin	2	elog(DEBUG1, "doc: %08x, attr: %s", doc, attr);
321	dpavlin	1
322	dpavlin	4	if ( (attrval = est_doc_attr(doc, attr)) && (attrlen = strlen(attrval)) ) {
323			val = (char ) palloc(attrlen sizeof(char));
324	dpavlin	1	} else {
325			return (Datum) NULL;
326			}
327
328			len = strlen(attrval);
329	dpavlin	2	elog(DEBUG1, "attr2text(%s) = '%s' %d bytes", attr, attrval, len);
330	dpavlin	1
331			len++;
332			len *= sizeof(char);
333
334			elog(DEBUG2, "palloc(%d)", len);
335
336			val = palloc(len);
337
338			memset(val, 0, len);
339			strncpy(val, attrval, len);
340
341			elog(DEBUG2, "val=%s", val);
342
343			return val;
344			}
345
346			/* make integer variable from property */
347			/*
348			char prop2int(SW_RESULT sw_res, char propname) {
349			char *val;
350			unsigned long prop;
351			int len;
352
353			elog(DEBUG2, "prop2int(%s)", propname);
354
355			prop = estResultPropertyULong( sw_res, propname );
356			if (error_or_abort( est_handle )) return NULL;
357
358			elog(DEBUG1, "prop2int(%s) = %lu", propname, prop);
359
360			len = 128 * sizeof(char);
361			elog(DEBUG2, "palloc(%d)", len);
362
363			val = palloc(len);
364			memset(val, 0, len);
365
366			snprintf(val, len, "%lu", prop);
367
368			elog(DEBUG2, "val=%s", val);
369
370			return val;
371			}
372			*/