pgswish/trunk/pgswish.c

/*
 * integrate swish-e into PostgreSQL
 *
 * Dobrica Pavlinusic <dpavlin@rot13.org> 2005-02-18
 *
 * TODO:
 * - check null input using PG_ARGISNULL before using PG_GETARG_xxxx
 * - support composite type arguments
 *
 * 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)
 * - swish-e example src/libtest.c (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 <swish-e.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)))


SW_HANDLE   swish_handle = NULL;/* Database handle */
SW_SEARCH   search = NULL;      /* search handle -- holds search parameters */
SW_RESULTS  swish_results = NULL;       /* results handle -- holds list of results */

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

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

        /* 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));

                /* 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);

                
                /* Send any errors or warnings to stderr (default is stdout) */
                SwishErrorsToStderr();

                elog(INFO, "pgswish: SwishInit(%s)", index_path);
                
                swish_handle = SwishInit( index_path );

                if (! swish_handle) {
                        elog(ERROR, "pgswish: can't open %s", index_path);
                        SRF_RETURN_DONE(funcctx);
                }
                
                if ( SwishError( swish_handle ) ) error_or_abort( swish_handle );
                /* set ranking scheme. default is 0 */
                SwishRankScheme( swish_handle, 0 );

                /* Check for errors after every call */
                if ( SwishError( swish_handle ) )
                        error_or_abort( swish_handle );  /* print an error or abort -- see below */

                elog(INFO, "pgswish: SwishQuery(%s)", query);
                /* Here's a short-cut to searching that creates a search object and searches at the same time */
                swish_results = SwishQuery( swish_handle, query);
                if ( SwishError( swish_handle ) ) error_or_abort( swish_handle );

                /* total number of tuples to be returned */
                funcctx->max_calls = SwishHits( swish_results );

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

                elog(INFO, "pgswish: SwishHits = %d", funcctx->max_calls);

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

                /* 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);
        }

        /* 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 (call_cntr < max_calls) {
                char            **values;
                HeapTuple       tuple;
                Datum           result;
                char            *prop;
                SW_RESULT       *sw_res;        /* one row from swish-e results */

                elog(INFO, "pgswish: in loop %d", call_cntr);

                if (! swish_results) {
                        elog(ERROR, "pgswish: no swish-e results");
                        SRF_RETURN_DONE(funcctx);
                }
                
                elog(INFO, "pgswish: check for swish-e error");
                if ( SwishError( swish_handle ) ) error_or_abort( swish_handle );

                /*
                 * 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.
                 */
                values = (char **) palloc(5 * sizeof(char *));

                sw_res = SwishNextResult( swish_results );
                if (! sw_res) {
                        elog(ERROR, "pgswish: swish-e sort result list: %d rows expected %d", call_cntr, max_calls - 1);
                        SRF_RETURN_DONE(funcctx);
                }
                
        elog(INFO, "Path: %s\n  Rank: %lu\n  Size: %lu\n  Title: %s\n  Index: %s\n  Modified: %s\n  Record #: %lu\n  File   #: %lu\n\n",
            SwishResultPropertyStr   ( sw_res, "swishdocpath" ),
            SwishResultPropertyULong ( sw_res, "swishrank" ),
            SwishResultPropertyULong ( sw_res, "swishdocsize" ),
            SwishResultPropertyStr   ( sw_res, "swishtitle"),
            SwishResultPropertyStr   ( sw_res, "swishdbfile" ),
            SwishResultPropertyStr   ( sw_res, "swishlastmodified" ),
            SwishResultPropertyULong ( sw_res, "swishreccount" ),  /* can figure this out in loop, of course */
            SwishResultPropertyULong ( sw_res, "swishfilenum" )
        );

                elog(INFO, "pgswish: get prop");
                prop = SwishResultPropertyStr ( sw_res, "swishdocpath" );
                elog(INFO, "pgswish: got error?");
                if ( SwishError( swish_handle ) ) error_or_abort( swish_handle );
                elog(INFO, "swishdocpath: %s", prop);

//              values[0] = GET_STR( SwishResultPropertyULong ( sw_res, "swishrank" ) );
                values[0] = NULL;
//              values[1] = GET_TEXT( SwishResultPropertyStr   ( sw_res, "swishdocpath" ) );
if (0) {
                values[1] = GET_TEXT( SwishResultPropertyStr   ( sw_res, "swishdocpath" ) );
                values[2] = _textout( SwishResultPropertyStr   ( sw_res, "swishtitle" ) );
                values[3] = _textout( SwishResultPropertyStr   ( sw_res, "swishdocsize" ) );
                values[4] = _textout( SwishResultPropertyStr   ( sw_res, "swishdbfile" ) );

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

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

}
                /* clean up (this is not really necessary) */

                SRF_RETURN_NEXT(funcctx, result);
        } else {
                /* free swish object and close */
                Free_Search_Object( search );
                SwishClose( swish_handle );

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

/*
 * elog errors
 *
 */

static void error_or_abort( SW_HANDLE swish_handle ) {
        if ( !SwishError( swish_handle ) )
                return;

        /* print a message */
        elog(ERROR, 
                "pgswish error: Number [%d], Type [%s],  Optional Message: [%s]\n",
                        SwishError( swish_handle ),
                        SwishErrorString( swish_handle ),
                        SwishLastErrorMsg( swish_handle )
        );
        if ( search ) Free_Search_Object( search );
        SwishClose( swish_handle );

        /* do when there is no more left */
}

1	dpavlin	8	/*
2			* integrate swish-e into PostgreSQL
3			*
4			* Dobrica Pavlinusic <dpavlin@rot13.org> 2005-02-18
5			*
6			* TODO:
7			* - check null input using PG_ARGISNULL before using PG_GETARG_xxxx
8			* - support composite type arguments
9			*
10			* NOTES:
11			* - clear structures with memset to support hash indexes (who whould like
12			* to create hash index on table returned from function?)
13	dpavlin	10	* - number of returned rows is set by PostgreSQL evaluator, see:
14			* http://archives.postgresql.org/pgsql-hackers/2005-02/msg00546.php
15	dpavlin	8	*
16	dpavlin	9	* Based on:
17			* - C example from PostgreSQL documentation (BSD licence)
18			* - swish-e example src/libtest.c (GPL)
19			* - _textin/_textout from pgcurl.c (LGPL)
20			*
21			* This code is licenced under GPL
22	dpavlin	8	*/
23
24			#include "postgres.h"
25			#include "fmgr.h"
26			#include "funcapi.h"
27	dpavlin	9	#include "utils/builtins.h"
28	dpavlin	8	#include <swish-e.h>
29
30	dpavlin	9	#define _textin(str) DirectFunctionCall1(textin, CStringGetDatum(str))
31			#define _textout(str) DatumGetPointer(DirectFunctionCall1(textout, PointerGetDatum(str)))
32	dpavlin	11	#define GET_STR(textp) DatumGetCString(DirectFunctionCall1(textout, PointerGetDatum(textp)))
33			#define GET_TEXT(cstrp) DatumGetTextP(DirectFunctionCall1(textin, CStringGetDatum(cstrp)))
34	dpavlin	8
35	dpavlin	9
36	dpavlin	8	SW_HANDLE swish_handle = NULL;/* Database handle */
37			SW_SEARCH search = NULL; /* search handle -- holds search parameters */
38	dpavlin	11	SW_RESULTS swish_results = NULL; /* results handle -- holds list of results */
39	dpavlin	8
40			/* define PostgreSQL v1 function */
41			PG_FUNCTION_INFO_V1(pgswish);
42			Datum pgswish(PG_FUNCTION_ARGS) {
43
44			FuncCallContext *funcctx;
45			int call_cntr;
46			int max_calls;
47			TupleDesc tupdesc;
48			TupleTableSlot *slot;
49			AttInMetadata *attinmeta;
50	dpavlin	9	char *index_path;
51			char *query;
52	dpavlin	8
53			/* stuff done only on the first call of the function */
54			if (SRF_IS_FIRSTCALL()) {
55			MemoryContext oldcontext;
56
57	dpavlin	9	/* take arguments from function */
58			//index_path = _textout(PG_GETARG_TEXT_P(0));
59			index_path = _textout(PG_GETARG_TEXT_P(0));
60			query = _textout(PG_GETARG_TEXT_P(1));
61
62	dpavlin	8	/* create a function context for cross-call persistence */
63			funcctx = SRF_FIRSTCALL_INIT();
64
65			/* switch to memory context appropriate for multiple function calls */
66			oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
67
68	dpavlin	9
69	dpavlin	8	/* Send any errors or warnings to stderr (default is stdout) */
70			SwishErrorsToStderr();
71
72	dpavlin	9	elog(INFO, "pgswish: SwishInit(%s)", index_path);
73	dpavlin	8
74	dpavlin	9	swish_handle = SwishInit( index_path );
75	dpavlin	8
76			if (! swish_handle) {
77	dpavlin	9	elog(ERROR, "pgswish: can't open %s", index_path);
78	dpavlin	8	SRF_RETURN_DONE(funcctx);
79			}
80
81			if ( SwishError( swish_handle ) ) error_or_abort( swish_handle );
82			/* set ranking scheme. default is 0 */
83	dpavlin	10	SwishRankScheme( swish_handle, 0 );
84	dpavlin	8
85			/* Check for errors after every call */
86			if ( SwishError( swish_handle ) )
87			error_or_abort( swish_handle ); /* print an error or abort -- see below */
88
89	dpavlin	9	elog(INFO, "pgswish: SwishQuery(%s)", query);
90	dpavlin	8	/* Here's a short-cut to searching that creates a search object and searches at the same time */
91	dpavlin	11	swish_results = SwishQuery( swish_handle, query);
92	dpavlin	8	if ( SwishError( swish_handle ) ) error_or_abort( swish_handle );
93
94			/* total number of tuples to be returned */
95	dpavlin	11	funcctx->max_calls = SwishHits( swish_results );
96	dpavlin	8
97			/* check if results exists */
98			if ( 0 == funcctx->max_calls )
99	dpavlin	9	elog(INFO, "no results for: %s", query );
100	dpavlin	8
101	dpavlin	9	elog(INFO, "pgswish: SwishHits = %d", funcctx->max_calls);
102
103	dpavlin	8	/* Build a tuple description for a __pgswish tuple */
104			tupdesc = RelationNameGetTupleDesc("__pgswish");
105
106			/* allocate a slot for a tuple with this tupdesc */
107			slot = TupleDescGetSlot(tupdesc);
108
109			/* assign slot to function context */
110			funcctx->slot = slot;
111
112			/*
113			* generate attribute metadata needed later to produce tuples from raw
114			* C strings
115			*/
116			attinmeta = TupleDescGetAttInMetadata(tupdesc);
117			funcctx->attinmeta = attinmeta;
118
119			MemoryContextSwitchTo(oldcontext);
120			}
121
122			/* stuff done on every call of the function */
123			funcctx = SRF_PERCALL_SETUP();
124
125			call_cntr = funcctx->call_cntr;
126			max_calls = funcctx->max_calls;
127			slot = funcctx->slot;
128			attinmeta = funcctx->attinmeta;
129
130			if (call_cntr < max_calls) {
131			char **values;
132			HeapTuple tuple;
133			Datum result;
134	dpavlin	11	char *prop;
135			SW_RESULT sw_res; / one row from swish-e results */
136	dpavlin	8
137	dpavlin	11	elog(INFO, "pgswish: in loop %d", call_cntr);
138
139			if (! swish_results) {
140			elog(ERROR, "pgswish: no swish-e results");
141			SRF_RETURN_DONE(funcctx);
142			}
143
144			elog(INFO, "pgswish: check for swish-e error");
145			if ( SwishError( swish_handle ) ) error_or_abort( swish_handle );
146
147	dpavlin	8	/*
148			* Prepare a values array for storage in our slot.
149			* This should be an array of C strings which will
150			* be processed later by the type input functions.
151			*/
152	dpavlin	9	values = (char *) palloc(5 sizeof(char *));
153	dpavlin	8
154	dpavlin	11	sw_res = SwishNextResult( swish_results );
155			if (! sw_res) {
156			elog(ERROR, "pgswish: swish-e sort result list: %d rows expected %d", call_cntr, max_calls - 1);
157			SRF_RETURN_DONE(funcctx);
158			}
159
160			elog(INFO, "Path: %s\n Rank: %lu\n Size: %lu\n Title: %s\n Index: %s\n Modified: %s\n Record #: %lu\n File #: %lu\n\n",
161			SwishResultPropertyStr ( sw_res, "swishdocpath" ),
162			SwishResultPropertyULong ( sw_res, "swishrank" ),
163			SwishResultPropertyULong ( sw_res, "swishdocsize" ),
164			SwishResultPropertyStr ( sw_res, "swishtitle"),
165			SwishResultPropertyStr ( sw_res, "swishdbfile" ),
166			SwishResultPropertyStr ( sw_res, "swishlastmodified" ),
167			SwishResultPropertyULong ( sw_res, "swishreccount" ), /* can figure this out in loop, of course */
168			SwishResultPropertyULong ( sw_res, "swishfilenum" )
169			);
170
171			elog(INFO, "pgswish: get prop");
172			prop = SwishResultPropertyStr ( sw_res, "swishdocpath" );
173			elog(INFO, "pgswish: got error?");
174			if ( SwishError( swish_handle ) ) error_or_abort( swish_handle );
175			elog(INFO, "swishdocpath: %s", prop);
176
177			// values[0] = GET_STR( SwishResultPropertyULong ( sw_res, "swishrank" ) );
178			values[0] = NULL;
179			// values[1] = GET_TEXT( SwishResultPropertyStr ( sw_res, "swishdocpath" ) );
180			if (0) {
181			values[1] = GET_TEXT( SwishResultPropertyStr ( sw_res, "swishdocpath" ) );
182			values[2] = _textout( SwishResultPropertyStr ( sw_res, "swishtitle" ) );
183			values[3] = _textout( SwishResultPropertyStr ( sw_res, "swishdocsize" ) );
184			values[4] = _textout( SwishResultPropertyStr ( sw_res, "swishdbfile" ) );
185
186	dpavlin	8	/* build a tuple */
187			tuple = BuildTupleFromCStrings(attinmeta, values);
188
189			/* make the tuple into a datum */
190			result = TupleGetDatum(slot, tuple);
191
192			}
193			/* clean up (this is not really necessary) */
194
195			SRF_RETURN_NEXT(funcctx, result);
196			} else {
197			/* free swish object and close */
198			Free_Search_Object( search );
199			SwishClose( swish_handle );
200
201			/* do when there is no more left */
202			SRF_RETURN_DONE(funcctx);
203			}
204			}
205
206			/*
207			* elog errors
208			*
209			*/
210
211			static void error_or_abort( SW_HANDLE swish_handle ) {
212			if ( !SwishError( swish_handle ) )
213			return;
214
215			/* print a message */
216			elog(ERROR,
217			"pgswish error: Number [%d], Type [%s], Optional Message: [%s]\n",
218			SwishError( swish_handle ),
219			SwishErrorString( swish_handle ),
220			SwishLastErrorMsg( swish_handle )
221			);
222			if ( search ) Free_Search_Object( search );
223			SwishClose( swish_handle );
224
225			/* do when there is no more left */
226			}
227