Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commitecab830

Browse files
committed
Added functions to list dictionaries (shared_ispell_dicts) and stoplists (shared_ispell_stoplists). Several other minor tweaks / improvements.
1 parent4e732cb commitecab830

File tree

3 files changed

+229
-20
lines changed

3 files changed

+229
-20
lines changed

‎sql/shared_ispell--1.0.0.sql

Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -23,6 +23,16 @@ CREATE OR REPLACE FUNCTION shared_ispell_mem_available()
2323
AS'MODULE_PATHNAME','dispell_mem_available'
2424
LANGUAGE C IMMUTABLE;
2525

26+
CREATE OR REPLACEFUNCTIONshared_ispell_dicts( OUT dict_nameVARCHAR, OUT affix_nameVARCHAR, OUT wordsINT, OUT affixesINT, OUT bytesINT)
27+
RETURNS SETOF record
28+
AS'MODULE_PATHNAME','dispell_list_dicts'
29+
LANGUAGE C IMMUTABLE;
30+
31+
CREATE OR REPLACEFUNCTIONshared_ispell_stoplists( OUT stop_nameVARCHAR, OUT wordsINT, OUT bytesINT)
32+
RETURNS SETOF record
33+
AS'MODULE_PATHNAME','dispell_list_stoplists'
34+
LANGUAGE C IMMUTABLE;
35+
2636
CREATETEXT SEARCH TEMPLATE shared_ispell (
2737
INIT= shared_ispell_init,
2838
LEXIZE= shared_ispell_lexize

‎src/shared_ispell.c

Lines changed: 216 additions & 18 deletions
Original file line numberDiff line numberDiff line change
@@ -20,6 +20,8 @@
2020
#include"storage/lwlock.h"
2121
#include"utils/timestamp.h"
2222

23+
#include"funcapi.h"
24+
2325
#include"libpq/md5.h"
2426

2527
#include"spell.h"
@@ -85,7 +87,7 @@ static SegmentInfo * segment_info = NULL;
8587

8688
staticchar*shalloc(intbytes);
8789

88-
staticSharedIspellDict*copyIspellDict(IspellDict*dict,char*dictFile,char*affixFile,intbytes);
90+
staticSharedIspellDict*copyIspellDict(IspellDict*dict,char*dictFile,char*affixFile,intbytes,intwords);
8991
staticSharedStopList*copyStopList(StopList*list,char*stopFile,intbytes);
9092

9193
staticintsizeIspellDict(IspellDict*dict,char*dictFile,char*affixFile);
@@ -230,12 +232,8 @@ void init_shared_dict(DictInfo * info, char * dictFile, char * affFile, char * s
230232

231233
IspellDict*dict;
232234
StopListstoplist;
233-
234-
/* FIXME Maybe we could treat the stop file separately, as it does not
235-
* influence the dictionary. So the SharedIspellDict would track just
236-
* dictionary and affixes, and the stop words would be kept somewhere
237-
* else - either separately in the shared segment, or in local memory
238-
* (the list is usually small and easy pro load) */
235+
236+
/* dictionary (words and affixes) */
239237
shdict=get_shared_dict(dictFile,affFile);
240238

241239
/* init if needed */
@@ -264,7 +262,7 @@ void init_shared_dict(DictInfo * info, char * dictFile, char * affFile, char * s
264262
}
265263

266264
/* fine, there's enough space - copy the dictionary */
267-
shdict=copyIspellDict(dict,dictFile,affFile,size);
265+
shdict=copyIspellDict(dict,dictFile,affFile,size,dict->nspell);
268266

269267
elog(INFO,"shared dictionary %s.dict / %s.affix loaded, used %d B, %ld B remaining",
270268
dictFile,affFile,size,segment_info->available);
@@ -312,12 +310,16 @@ Datum dispell_lexize(PG_FUNCTION_ARGS);
312310
Datumdispell_reset(PG_FUNCTION_ARGS);
313311
Datumdispell_mem_available(PG_FUNCTION_ARGS);
314312
Datumdispell_mem_used(PG_FUNCTION_ARGS);
313+
Datumdispell_list_dicts(PG_FUNCTION_ARGS);
314+
Datumdispell_list_stoplists(PG_FUNCTION_ARGS);
315315

316316
PG_FUNCTION_INFO_V1(dispell_init);
317317
PG_FUNCTION_INFO_V1(dispell_lexize);
318318
PG_FUNCTION_INFO_V1(dispell_reset);
319319
PG_FUNCTION_INFO_V1(dispell_mem_available);
320320
PG_FUNCTION_INFO_V1(dispell_mem_used);
321+
PG_FUNCTION_INFO_V1(dispell_list_dicts);
322+
PG_FUNCTION_INFO_V1(dispell_list_stoplists);
321323

322324
Datum
323325
dispell_reset(PG_FUNCTION_ARGS)
@@ -507,9 +509,10 @@ char * shalloc(int bytes) {
507509
char*result;
508510
bytes=MAXALIGN(bytes);
509511

512+
/* This shouldn't really happen, as the init_shared_dict checks the size
513+
* prior to copy. So let's just throw error here, as something went
514+
* obviously wrong. */
510515
if (bytes>segment_info->available) {
511-
/* FIXME this should not throw error, it should rather return NULL
512-
* and reset the alloc info (this way the memory is wasted forever) */
513516
elog(ERROR,"the shared segment (shared ispell) is too small");
514517
}
515518

@@ -595,6 +598,11 @@ int sizeRegisNode(RegisNode * node) {
595598
returnsize;
596599
}
597600

601+
static
602+
char*copyRegexGuts(char*guts) {
603+
returnNULL;
604+
}
605+
598606
static
599607
AFFIX*copyAffix(AFFIX*affix) {
600608

@@ -605,11 +613,14 @@ AFFIX * copyAffix(AFFIX * affix) {
605613
copy->find=shstrcpy(affix->find);
606614
copy->repl=shstrcpy(affix->repl);
607615

608-
if (copy->isregis) {
609-
copy->reg.regis.node=copyRegisNode(copy->reg.regis.node);
610-
}elseif (!copy->issimple) {
611-
// FIXME handle the regex_t properly (copy the strings etc)
616+
if (affix->isregis) {
617+
copy->reg.regis.node=copyRegisNode(affix->reg.regis.node);
618+
}elseif (!affix->issimple) {
619+
620+
/*FIXME Need to copy the regex_t properly. But would a plain copy be
621+
* safe tu use by multiple processes at the same time? */
612622
elog(WARNING,"skipping regex_t");
623+
613624
}
614625

615626
returncopy;
@@ -627,8 +638,11 @@ int sizeAffix(AFFIX * affix) {
627638
if (affix->isregis) {
628639
size+=sizeRegisNode(affix->reg.regis.node);
629640
}elseif (!affix->issimple) {
630-
// FIXME handle the regex_t properly (copy the strings etc)
641+
642+
/*FIXME Need to copy the regex_t properly. But would a plain copy be
643+
* safe tu use by multiple processes at the same time? */
631644
elog(WARNING,"skipping regex_t");
645+
632646
}
633647

634648
returnsize;
@@ -698,7 +712,7 @@ SharedStopList * copyStopList(StopList * list, char * stopFile, int size) {
698712
copy->list.len=list->len;
699713
copy->list.stop= (char**)shalloc(sizeof(char*)*list->len);
700714
copy->stopFile=shstrcpy(stopFile);
701-
copy->bytes=size;
715+
copy->nbytes=size;
702716

703717
for (i=0;i<list->len;i++) {
704718
copy->list.stop[i]=shstrcpy(list->stop[i]);
@@ -742,7 +756,7 @@ int countCMPDAffixes(CMPDAffix * affixes) {
742756
}
743757

744758
static
745-
SharedIspellDict*copyIspellDict(IspellDict*dict,char*dictFile,char*affixFile,intsize) {
759+
SharedIspellDict*copyIspellDict(IspellDict*dict,char*dictFile,char*affixFile,intsize,intwords) {
746760

747761
inti,cnt;
748762

@@ -778,7 +792,8 @@ SharedIspellDict * copyIspellDict(IspellDict * dict, char * dictFile, char * aff
778792
memcpy(copy->flagval,dict->flagval,255);
779793
copy->usecompound=dict->usecompound;
780794

781-
copy->bytes=size;
795+
copy->nbytes=size;
796+
copy->nwords=words;
782797

783798
returncopy;
784799

@@ -812,3 +827,186 @@ int sizeIspellDict(IspellDict * dict, char * dictFile, char * affixFile) {
812827
returnsize;
813828

814829
}
830+
831+
Datum
832+
dispell_list_dicts(PG_FUNCTION_ARGS)
833+
{
834+
FuncCallContext*funcctx;
835+
TupleDesctupdesc;
836+
AttInMetadata*attinmeta;
837+
SharedIspellDict*dict;
838+
839+
/* init on the first call */
840+
if (SRF_IS_FIRSTCALL()) {
841+
842+
MemoryContextoldcontext;
843+
844+
funcctx=SRF_FIRSTCALL_INIT();
845+
oldcontext=MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
846+
847+
/* get a shared lock and then the first dictionary */
848+
LWLockAcquire(segment_info->lock,LW_SHARED);
849+
funcctx->user_fctx=segment_info->dict;
850+
851+
/* Build a tuple descriptor for our result type */
852+
if (get_call_result_type(fcinfo,NULL,&tupdesc)!=TYPEFUNC_COMPOSITE)
853+
ereport(ERROR,
854+
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
855+
errmsg("function returning record called in context "
856+
"that cannot accept type record")));
857+
858+
/*
859+
* generate attribute metadata needed later to produce tuples from raw
860+
* C strings
861+
*/
862+
attinmeta=TupleDescGetAttInMetadata(tupdesc);
863+
funcctx->attinmeta=attinmeta;
864+
funcctx->tuple_desc=tupdesc;
865+
866+
/* switch back to the old context */
867+
MemoryContextSwitchTo(oldcontext);
868+
869+
}
870+
871+
/* init the context */
872+
funcctx=SRF_PERCALL_SETUP();
873+
874+
/* check if we have more data */
875+
if (funcctx->user_fctx!=NULL)
876+
{
877+
HeapTupletuple;
878+
Datumresult;
879+
Datumvalues[5];
880+
boolnulls[5];
881+
882+
text*dictname,*affname;
883+
884+
dict= (SharedIspellDict*)funcctx->user_fctx;
885+
funcctx->user_fctx=dict->next;
886+
887+
memset(nulls,0,sizeof(nulls));
888+
889+
dictname= (text*)palloc(strlen(dict->dictFile)+VARHDRSZ);
890+
affname= (text*)palloc(strlen(dict->affixFile)+VARHDRSZ);
891+
892+
SET_VARSIZE(dictname,strlen(dict->dictFile)+VARHDRSZ);
893+
SET_VARSIZE(affname,strlen(dict->affixFile)+VARHDRSZ);
894+
895+
strcpy(VARDATA(dictname),dict->dictFile);
896+
strcpy(VARDATA(affname),dict->affixFile);
897+
898+
values[0]=PointerGetDatum(dictname);
899+
values[1]=PointerGetDatum(affname);
900+
values[2]=UInt32GetDatum(dict->nwords);
901+
values[3]=UInt32GetDatum(dict->naffixes);
902+
values[4]=UInt32GetDatum(dict->nbytes);
903+
904+
/* Build and return the tuple. */
905+
tuple=heap_form_tuple(funcctx->tuple_desc,values,nulls);
906+
907+
/* make the tuple into a datum */
908+
result=HeapTupleGetDatum(tuple);
909+
910+
/* Here we want to return another item: */
911+
SRF_RETURN_NEXT(funcctx,result);
912+
913+
}
914+
else
915+
{
916+
/* release the lock */
917+
LWLockRelease(segment_info->lock);
918+
919+
/* Here we are done returning items and just need to clean up: */
920+
SRF_RETURN_DONE(funcctx);
921+
}
922+
923+
}
924+
925+
Datum
926+
dispell_list_stoplists(PG_FUNCTION_ARGS)
927+
{
928+
FuncCallContext*funcctx;
929+
TupleDesctupdesc;
930+
AttInMetadata*attinmeta;
931+
SharedStopList*stoplist;
932+
933+
/* init on the first call */
934+
if (SRF_IS_FIRSTCALL()) {
935+
936+
MemoryContextoldcontext;
937+
938+
funcctx=SRF_FIRSTCALL_INIT();
939+
oldcontext=MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
940+
941+
/* get a shared lock and then the first stop list */
942+
LWLockAcquire(segment_info->lock,LW_SHARED);
943+
funcctx->user_fctx=segment_info->stop;
944+
945+
/* Build a tuple descriptor for our result type */
946+
if (get_call_result_type(fcinfo,NULL,&tupdesc)!=TYPEFUNC_COMPOSITE)
947+
ereport(ERROR,
948+
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
949+
errmsg("function returning record called in context "
950+
"that cannot accept type record")));
951+
952+
/*
953+
* generate attribute metadata needed later to produce tuples from raw
954+
* C strings
955+
*/
956+
attinmeta=TupleDescGetAttInMetadata(tupdesc);
957+
funcctx->attinmeta=attinmeta;
958+
funcctx->tuple_desc=tupdesc;
959+
960+
/* switch back to the old context */
961+
MemoryContextSwitchTo(oldcontext);
962+
963+
}
964+
965+
/* init the context */
966+
funcctx=SRF_PERCALL_SETUP();
967+
968+
/* check if we have more data */
969+
if (funcctx->user_fctx!=NULL)
970+
{
971+
HeapTupletuple;
972+
Datumresult;
973+
Datumvalues[3];
974+
boolnulls[3];
975+
976+
text*stopname;
977+
978+
stoplist= (SharedStopList*)funcctx->user_fctx;
979+
funcctx->user_fctx=stoplist->next;
980+
981+
memset(nulls,0,sizeof(nulls));
982+
983+
stopname= (text*)palloc(strlen(stoplist->stopFile)+VARHDRSZ);
984+
985+
SET_VARSIZE(stopname,strlen(stoplist->stopFile)+VARHDRSZ);
986+
987+
strcpy(VARDATA(stopname),stoplist->stopFile);
988+
989+
values[0]=PointerGetDatum(stopname);
990+
values[1]=UInt32GetDatum(stoplist->list.len);
991+
values[2]=UInt32GetDatum(stoplist->nbytes);
992+
993+
/* Build and return the tuple. */
994+
tuple=heap_form_tuple(funcctx->tuple_desc,values,nulls);
995+
996+
/* make the tuple into a datum */
997+
result=HeapTupleGetDatum(tuple);
998+
999+
/* Here we want to return another item: */
1000+
SRF_RETURN_NEXT(funcctx,result);
1001+
1002+
}
1003+
else
1004+
{
1005+
/* release the lock */
1006+
LWLockRelease(segment_info->lock);
1007+
1008+
/* Here we are done returning items and just need to clean up: */
1009+
SRF_RETURN_DONE(funcctx);
1010+
}
1011+
1012+
}

‎src/spell.h

Lines changed: 3 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -27,7 +27,8 @@ typedef struct SharedIspellDict
2727
char*dictFile;
2828
char*affixFile;
2929

30-
size_tbytes;
30+
intnbytes;
31+
intnwords;
3132

3233
/* next dictionary in the chain */
3334
structSharedIspellDict*next;
@@ -56,7 +57,7 @@ typedef struct SharedStopList
5657

5758
char*stopFile;
5859

59-
size_tbytes;
60+
intnbytes;
6061

6162
StopListlist;
6263
structSharedStopList*next;

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp