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);
                }
                
                error_or_abort( swish_handle );
                /* set ranking scheme. default is 0 */
                SwishRankScheme( swish_handle, 0 );
                error_or_abort( swish_handle );

                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);
                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;
                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");
                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(4 * 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, "swishdocpath: %s", prop2text( sw_res, "swishdocpath" ) );
                values[0] = prop2int( sw_res, "swishrank" );
                values[1] = prop2text( sw_res, "swishdocpath" );
                values[2] = prop2text( sw_res, "swishtitle" );
                values[3] = prop2int( sw_res, "swishdocsize" );
                
                /* build a tuple */
                tuple = BuildTupleFromCStrings(attinmeta, values);

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

                /* clean up ? */

                elog(INFO, "row: %s|%s|%s|%s",values[0],values[1],values[2],values[3]);
        
                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
 *
 */

char *prop2text(SW_RESULT sw_res, char *propname) {
        char *val;
        char *prop;
        int len;

        elog(INFO, "prop2text(%s)", propname);

        prop = SwishResultPropertyStr( sw_res, propname );
        error_or_abort( swish_handle );

        len = strlen(prop);
        elog(INFO, "prop2text(%s) = '%s' %d bytes", propname, prop, len);

        len++;
        len *= sizeof(char);

        elog(INFO, "palloc(%d)", len);

        val = palloc(len);

        memset(val, 0, len);
        strncpy(val, prop, len);

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

        return val;
}

char *prop2int(SW_RESULT sw_res, char *propname) {
        char *val;
        unsigned long prop;
        int len;

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

        prop = SwishResultPropertyULong( sw_res, propname );
        error_or_abort( swish_handle );

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

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

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

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

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

        return val;
}


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	dpavlin	12	error_or_abort( swish_handle );
82	dpavlin	8	/* set ranking scheme. default is 0 */
83	dpavlin	10	SwishRankScheme( swish_handle, 0 );
84	dpavlin	12	error_or_abort( swish_handle );
85	dpavlin	8
86	dpavlin	9	elog(INFO, "pgswish: SwishQuery(%s)", query);
87	dpavlin	8	/* Here's a short-cut to searching that creates a search object and searches at the same time */
88	dpavlin	11	swish_results = SwishQuery( swish_handle, query);
89	dpavlin	12	error_or_abort( swish_handle );
90	dpavlin	8
91			/* total number of tuples to be returned */
92	dpavlin	11	funcctx->max_calls = SwishHits( swish_results );
93	dpavlin	8
94			/* check if results exists */
95			if ( 0 == funcctx->max_calls )
96	dpavlin	9	elog(INFO, "no results for: %s", query );
97	dpavlin	8
98	dpavlin	9	elog(INFO, "pgswish: SwishHits = %d", funcctx->max_calls);
99
100	dpavlin	8	/* Build a tuple description for a __pgswish tuple */
101			tupdesc = RelationNameGetTupleDesc("__pgswish");
102
103			/* allocate a slot for a tuple with this tupdesc */
104			slot = TupleDescGetSlot(tupdesc);
105
106			/* assign slot to function context */
107			funcctx->slot = slot;
108
109			/*
110			* generate attribute metadata needed later to produce tuples from raw
111			* C strings
112			*/
113			attinmeta = TupleDescGetAttInMetadata(tupdesc);
114			funcctx->attinmeta = attinmeta;
115
116			MemoryContextSwitchTo(oldcontext);
117			}
118
119			/* stuff done on every call of the function */
120			funcctx = SRF_PERCALL_SETUP();
121
122			call_cntr = funcctx->call_cntr;
123			max_calls = funcctx->max_calls;
124			slot = funcctx->slot;
125			attinmeta = funcctx->attinmeta;
126
127			if (call_cntr < max_calls) {
128			char **values;
129			HeapTuple tuple;
130			Datum result;
131	dpavlin	11	SW_RESULT sw_res; / one row from swish-e results */
132	dpavlin	8
133	dpavlin	11	elog(INFO, "pgswish: in loop %d", call_cntr);
134
135			if (! swish_results) {
136			elog(ERROR, "pgswish: no swish-e results");
137			SRF_RETURN_DONE(funcctx);
138			}
139
140			elog(INFO, "pgswish: check for swish-e error");
141	dpavlin	12	error_or_abort( swish_handle );
142	dpavlin	11
143	dpavlin	8	/*
144			* Prepare a values array for storage in our slot.
145			* This should be an array of C strings which will
146			* be processed later by the type input functions.
147			*/
148	dpavlin	12	values = (char *) palloc(4 sizeof(char *));
149	dpavlin	8
150	dpavlin	11	sw_res = SwishNextResult( swish_results );
151			if (! sw_res) {
152			elog(ERROR, "pgswish: swish-e sort result list: %d rows expected %d", call_cntr, max_calls - 1);
153			SRF_RETURN_DONE(funcctx);
154			}
155
156	dpavlin	12	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",
157			SwishResultPropertyStr ( sw_res, "swishdocpath" ),
158			SwishResultPropertyULong ( sw_res, "swishrank" ),
159			SwishResultPropertyULong ( sw_res, "swishdocsize" ),
160			SwishResultPropertyStr ( sw_res, "swishtitle"),
161			SwishResultPropertyStr ( sw_res, "swishdbfile" ),
162			SwishResultPropertyStr ( sw_res, "swishlastmodified" ),
163			SwishResultPropertyULong ( sw_res, "swishreccount" ), /* can figure this out in loop, of course */
164			SwishResultPropertyULong ( sw_res, "swishfilenum" )
165			);
166	dpavlin	11
167	dpavlin	12	elog(INFO, "swishdocpath: %s", prop2text( sw_res, "swishdocpath" ) );
168			values[0] = prop2int( sw_res, "swishrank" );
169			values[1] = prop2text( sw_res, "swishdocpath" );
170			values[2] = prop2text( sw_res, "swishtitle" );
171			values[3] = prop2int( sw_res, "swishdocsize" );
172
173	dpavlin	8	/* build a tuple */
174			tuple = BuildTupleFromCStrings(attinmeta, values);
175
176			/* make the tuple into a datum */
177			result = TupleGetDatum(slot, tuple);
178
179	dpavlin	12	/* clean up ? */
180	dpavlin	8
181	dpavlin	12	elog(INFO, "row: %s\|%s\|%s\|%s",values[0],values[1],values[2],values[3]);
182
183	dpavlin	8	SRF_RETURN_NEXT(funcctx, result);
184			} else {
185			/* free swish object and close */
186			Free_Search_Object( search );
187			SwishClose( swish_handle );
188
189			/* do when there is no more left */
190			SRF_RETURN_DONE(funcctx);
191			}
192			}
193
194			/*
195			* elog errors
196			*
197			*/
198
199	dpavlin	12	char prop2text(SW_RESULT sw_res, char propname) {
200			char *val;
201			char *prop;
202			int len;
203
204			elog(INFO, "prop2text(%s)", propname);
205
206			prop = SwishResultPropertyStr( sw_res, propname );
207			error_or_abort( swish_handle );
208
209			len = strlen(prop);
210			elog(INFO, "prop2text(%s) = '%s' %d bytes", propname, prop, len);
211
212			len++;
213			len *= sizeof(char);
214
215			elog(INFO, "palloc(%d)", len);
216
217			val = palloc(len);
218
219			memset(val, 0, len);
220			strncpy(val, prop, len);
221
222			elog(INFO, "val=%s", val);
223
224			return val;
225			}
226
227			char prop2int(SW_RESULT sw_res, char propname) {
228			char *val;
229			unsigned long prop;
230			int len;
231
232			elog(INFO, "prop2int(%s)", propname);
233
234			prop = SwishResultPropertyULong( sw_res, propname );
235			error_or_abort( swish_handle );
236
237			elog(INFO, "prop2int(%s) = %lu", propname, prop);
238
239			len = 128 * sizeof(char);
240			elog(INFO, "palloc(%d)", len);
241
242			val = palloc(len);
243			memset(val, 0, len);
244
245			snprintf(val, len, "%lu", prop);
246
247			elog(INFO, "val=%s", val);
248
249			return val;
250			}
251
252
253	dpavlin	8	static void error_or_abort( SW_HANDLE swish_handle ) {
254			if ( !SwishError( swish_handle ) )
255			return;
256
257			/* print a message */
258			elog(ERROR,
259			"pgswish error: Number [%d], Type [%s], Optional Message: [%s]\n",
260			SwishError( swish_handle ),
261			SwishErrorString( swish_handle ),
262			SwishLastErrorMsg( swish_handle )
263			);
264			if ( search ) Free_Search_Object( search );
265			SwishClose( swish_handle );
266
267			/* do when there is no more left */
268			}
269