source: fmts/trunk/p/C0XMAIN.m@ 1273

Last change on this file since 1273 was 1273, checked in by George Lilly, 13 years ago

initial load of ftms project files and test data

File size: 12.4 KB
Line 
1C0XMAIN ; GPL - Fileman Triples entry point routine ;10/13/11 17:05
2 ;;0.1;C0X;nopatch;noreleasedate;Build 1
3 ;Copyright 2011 George Lilly. Licensed under the terms of the GNU
4 ;General Public License See attached copy of the License.
5 ;
6 ;This program is free software; you can redistribute it and/or modify
7 ;it under the terms of the GNU General Public License as published by
8 ;the Free Software Foundation; either version 2 of the License, or
9 ;(at your option) any later version.
10 ;
11 ;This program is distributed in the hope that it will be useful,
12 ;but WITHOUT ANY WARRANTY; without even the implied warranty of
13 ;MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 ;GNU General Public License for more details.
15 ;
16 ;You should have received a copy of the GNU General Public License along
17 ;with this program; if not, write to the Free Software Foundation, Inc.,
18 ;51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
19 ;
20 Q
21 ;
22INITFARY(ZFARY) ; INITIALIZE FILE NUMBERS AND OTHER USEFUL THINGS
23 ; FOR THE DEFAULT TRIPLE STORE. USE OTHER VALUES FOR SUPPORTING ADDITIONAL
24 ; TRIPLE STORES
25 S @ZFARY@("C0XTFN")=172.101 ; TRIPLES FILE NUMBER
26 S @ZFARY@("C0XSFN")=172.201 ; TRIPLES STRINGS FILE NUMBER
27 S @ZFARY@("C0XTN")=$NA(^C0X(101)) ; TRIPLES GLOBAL NAME
28 S @ZFARY@("C0XSN")=$NA(^C0X(201)) ; STRING FILE GLOBAL NAME
29 S @ZFARY@("C0XDIR")="/home/glilly/all_smart_patient_data/smart-rdf/"
30 D USEFARY(ZFARY)
31 Q
32 ;
33USEFARY(ZFARY) ; INITIALIZES VARIABLES SAVED IN ARRAY ZFARY
34 N ZI S ZI=""
35 F S ZI=$O(@ZFARY@(ZI)) Q:ZI="" D
36 . ;N ZX
37 . S ZX="S "_ZI_"="""_@ZFARY@(ZI)_""""
38 . ;W !,ZX
39 . X ZX
40 Q
41 ;
42IMPORT(FNAME,FDIR,FURL,FARY) ; EXTRINSIC THAT READS A FILE FROM THE STANDARD
43 ; DIRECTORY, LOADS IT INTO THE TRIPLESTORE AS TEXT, AND RETURNS THE
44 ; NODE NAME OF THE TEXT TRIPLE
45 ; FDIR IS THE OPTIONAL DIRECTORY (DEFAUTS TO STANDARD DIR)
46 ; FURL IS THE OPTIONAL URI FOR ACCESSING THE FILE FROM THE TRIPLE STORE
47 ; FARY IS THE OPTIONAL FILE ARRAY OF THE TRIPLE STORE TO USE
48 I '$D(FARY) D ;
49 . D INITFARY("C0XFARY")
50 . S FARY="C0XFARY"
51 D USEFARY(FARY)
52 N ZD,ZTMP
53 I '$D(FDIR) S FDIR=C0XDIR ; DIRECTORY OF THE RDF FILE
54 I '$D(FURL) D ;
55 . N ZN2 S ZN2=$TR(FNAME,".","_") ; REMOVE THE DOT FROM THE NAME
56 . S FURL=FDIR_ZN2
57 N ZTMP
58 S ZTMP=$NA(^TMP("C0X",$J,"FILEIN",1)) ; WHERE TO PUT THE INCOMING FILE
59 K @ZTMP ; MAKE SURE IT'S CLEAR
60 S C0XSTART=$$NOW^XLFDT
61 W !,"STARTED: ",C0XSTART
62 W !,"READING IN: ",FNAME
63 I '$$FILEIN(ZTMP,FDIR,FNAME,4) D Q ; QUIT IF NO SUCCESS
64 . W !,"ERROR READING FILE: ",FDIR,FNAME
65 S ZRDF=$NA(^TMP("C0X",$J,"FILEIN")) ; WITHOUT THE SUBSCRIPT
66 W !,$O(@ZRDF@(""),-1)," LINES READ"
67 D INSRDF(ZRDF,FURL,FARY) ; IMPORT AND PROCESS THE RDF
68 Q
69 ;
70WGET(ZURL,FARY) ; GET FROM THE INTERNET AN RDF FILE AND INSERT IT
71 ;
72 I '$D(FARY) D ;
73 . D INITFARY("C0XFARY")
74 . S FARY="C0XFARY"
75 D USEFARY(FARY)
76 ;N ZLOC,ZTMP
77 K ZTMP
78 S ZLOC=$NA(^TMP("C0X","WGET",$J))
79 S C0XSTART=$$NOW^XLFDT
80 W !,"STARTED: ",C0XSTART
81 W !,"DOWNLOADING: ",ZURL
82 S OK=$$httpGET^%zewdGTM(ZURL,.ZTMP)
83 M @ZLOC=ZTMP
84 W !,$O(@ZLOC@(""),-1)," LINES READ"
85 D INSRDF(ZLOC,ZURL,FARY)
86 Q
87 ;
88INSRDF(ZRDF,ZNAME,FARY) ; INSERT AN RDF FILE INTO THE STORE AND PROCESS
89 ; ZRDF IS PASSED BY NAME
90 I '$D(FARY) D ;
91 . D INITFARY("C0XFARY")
92 . S FARY="C0XFARY"
93 D USEFARY(FARY)
94 N ZGRAPH,ZSUBJECT
95 S ZGRAPH="_:G"_$$LKY9 ; RANDOM GRAPH NAME
96 S ZSUBJECT=$$ANONS ; RANDOM ANOYMOUS SUBJECT
97 D ADD(ZGRAPH,ZSUBJECT,"fmts:url",ZNAME,FARY)
98 N ZTXTNM
99 S ZTXTNM="_TXT_INCOMING_RDF_FILE_"_ZNAME_"_"_$$LKY9 ; NAME FOR TEXT NODE
100 D ADD(ZGRAPH,ZSUBJECT,"fmts:rdfSource",ZTXTNM,FARY)
101 D UPDIE(.C0XFDA) ; TRY IT OUT
102 K C0XCNT ;RESET FOR NEXT TIME
103 D STORETXT(ZRDF,ZTXTNM,FARY)
104 W !,"ADDED: ",ZGRAPH," ",ZSUBJECT," fmts:rdfSource ",ZTXTNM
105 D PROCESS(.G,ZRDF,ZNAME,ZGRAPH,FARY) ; PARSE AND INSERT THE RDF
106 Q
107 ;
108STORETXT(ZTXT,ZNAME,FARY) ; STORE TEXT IN THE TRIPLESTORE AT ZNAME
109 ;
110 I '$D(FARY) D ;
111 . D INITFARY("C0XFARY")
112 . S FARY="C0XFARY"
113 D USEFARY(FARY)
114 N ZIEN
115 S ZIEN=$$IENOF(ZNAME,FARY) ; GET THE IEN
116 D CLEAN^DILF
117 K ZERR
118 D WP^DIE(C0XSFN,ZIEN_",",1,,ZTXT,"ZERR")
119 I $D(ZERR) D ;
120 . ZWR ZERR
121 Q
122 ;
123GETTXT(ZRTN,ZNAME,FARY) ; RETURNS RDF SOURCE OR OTHER TEXT
124 ; ZRTN IS PASSED BY REFERENCE
125 I '$D(FARY) D ;
126 . D INITFARY("C0XFARY")
127 . S FARY="C0XFARY"
128 D USEFARY(FARY)
129 N ZIEN
130 S ZIEN=$$IENOF(ZNAME)
131 S OK=$$GET1^DIQ(C0XSFN,ZIEN_",",1,,"ZRTN")
132 Q
133 ;
134WHERETXT(ZNAME,FARY) ; EXTRINSIC WHICH RETURNS THE NAME OF THE GLOBAL
135 ; WHERE THE TEXT IS LOCATED. NAME IS THE NAME OF THE STRING
136 I '$D(FARY) D ;
137 . D INITFARY("C0XFARY")
138 . S FARY="C0XFARY"
139 D USEFARY(FARY)
140 N ZIEN
141 S ZIEN=$$IENOF(ZNAME)
142 Q $NA(@C0XSN@(ZIEN,1))
143 ;
144FILEIN(ZINTMP,ZDIR,ZFNAME,ZLVL) ; READS A FILE INTO ZINTMP USING FTG^%ZISH
145 ; ZINTMP IS PASSED BY NAME AND INCLUDES THE NEW SUBSCRIPT
146 ; IE ^TMP("C0X","FILEIN",1)
147 ; ZLVL IN THIS CASE WOULD BE 3 INCREMENTING THE 1
148 ; EXTRINSIC WHICH RETURNS THE RESULT OF FTG^%ZISH
149 S OK=$$FTG^%ZISH(ZDIR,FNAME,ZINTMP,ZLVL)
150 Q OK
151 ;
152TESTPROC ; TEST PROCESS WITH EXISTING SMALL RDF FILE
153 S ZIN=$NA(^TMP("C0X",12226,"FILEIN"))
154 S ZGRAPH="/test/rdfFile"
155 S ZM="/test/rdfFile/meta"
156 D PROCESS(.G,ZIN,ZGRAPH,ZM)
157 Q
158 ;
159PROCESS(ZRTN,ZRDF,ZGRF,ZMETA,FARY) ; PROCESS AN INCOMING RDF FILE
160 ; ZRTN IS PASS BY REFERENCE AND RETURNS MESSAGES ABOUT THE PROCESSING
161 ; ZRDF IS PASSED BY NAME AND IS THE GLOBAL CONTAINING THE RDF FILE
162 ; ZGRF IS THE NAME OF THE GRAPH TO USE IN THE TRIPLE STORE FOR RESULTS
163 ; ZMETA IS OPTIONAL AND IS THE NAME OF THE GRAPH TO STORE METADATA
164 ;
165 I '$D(FARY) D ;
166 . D INITFARY("C0XFARY")
167 . S FARY="C0XFARY"
168 D USEFARY(FARY)
169 ; -- first parse the rdf file with the MXML parser
170 ;S C0XDOCID=$$PARSE^C0CNHIN(ZRDF,"C0XARRAY") ; PARSE WITH MXML
171 S C0XDOCID=$$EN^MXMLDOM(ZRDF,"W")
172 ; -- assign the MXLM dom global name to ZDOM
173 S ZDOM=$NA(^TMP("MXMLDOM",$J,C0XDOCID))
174 W !,$O(@ZDOM@(""),-1)," XML NODES PARSED"
175 ; -- populate the metagraph to point to the graph with status unfinished
176 S METAS=$$ANONS ; GET AN ANONOMOUS RANDOM SUBJECT
177 I '$D(ZMETA) S ZMETA="_:G"_$$LKY9 ; RANDOM GRAPH NAME FOR METAGRAPH
178 D ADD(ZMETA,METAS,"fmts:about",ZGRF,FARY) ; POINT THE META TO THE GRAPH
179 D ADD(ZMETA,METAS,"fmts:status","unfinished",FARY) ; mark as unfinished
180 ;S C0XDATE=$$FMDTOUTC^C0CUTIL($$NOW^XLFDT,"DT")
181 S C0XDATE=$$NOW^XLFDT
182 D ADD(ZMETA,METAS,"fmts:dateTime",C0XDATE,FARY)
183 D UPDIE(.C0XFDA) ; commit the metagraph changes to the triple store
184 ; --
185 ; -- pull out the vocabularies in the RDF statement. marked with xmlns:
186 ; -- put them in a local variable for quick reference
187 ; -- TODO: create a graph for vocabularies and validate incoming against it
188 ;
189 S C0XVOC=""
190 N ZI,ZJ,ZK S ZI=""
191 F S ZI=$O(@ZDOM@(1,"A",ZI)) Q:ZI="" D ; FOR EACH xmlns
192 . S ZVOC=$P(ZI,"xmlns:",2)
193 . I ZVOC'="" S C0XVOC(ZVOC)=$G(@ZDOM@(1,"A",ZI))
194 ;W !,"VOCABS:" ZWR C0XVOC
195 ;
196 ; -- look for children called rdf:Description. quit if none. not an rdf file
197 ;
198 S ZI=$O(@ZDOM@(1,"C",""))
199 I $G(@ZDOM@(1,"C",ZI))'="rdf:Description" D Q ; not an rdf file
200 . W !,"Error. Not an RDF file. Cannot process."
201 ;
202 ; -- now process the rdf description children
203 ;
204 S ZI=""
205 S (C0XSUB,C0XPRE,C0XOBJ)="" ; INITIALIZE subject, object and predicate
206 F S ZI=$O(@ZDOM@(1,"C",ZI)) Q:ZI="" D ;
207 . ; -- we are skipping any child that is not rdf:Description
208 . ; -- TODO: check to see if this is right in general
209 . ;
210 . IF $G(@ZDOM@(1,"C",ZI))'="rdf:Description" D Q ;
211 . . W !,"SKIPPING NODE: ",ZI
212 . ; -- now looking for the subject for the triples
213 . S ZX=$G(@ZDOM@(ZI,"A","rdf:about"))
214 . I ZX'="" D ; we have the subject
215 . . ;W " about: ",ZX
216 . . S C0XSUB=ZX
217 . E D ;
218 . . S ZX=$G(@ZDOM@(ZI,"A","rdf:nodeID")) ; node id is another style of subject
219 . . I ZX'="" D ;
220 . . . S C0XSUB=ZX
221 . I C0XSUB="" S C0XSUB=$$ANONS ; DEFAULT TO BLANK SUBJECT
222 . ;
223 . ; -- we now have the subject. the children of this node have the rest
224 . ;
225 . S ZJ="" ; for the children of the rdf:Description nodes
226 . F S ZJ=$O(@ZDOM@(ZI,"C",ZJ)) Q:ZJ="" D ; for each child
227 . . S C0XPRE=@ZDOM@(ZJ) ; the predicate without a prefix
228 . . S ZX=$G(@ZDOM@(ZJ,"A","xmlns")) ; name space
229 . . I ZX'="" S C0XPRE=ZX_C0XPRE ; add the namespace prefix
230 . . I C0XPRE[":" D ; expand using vocabulary
231 . . . N ZB,ZA
232 . . . S ZB=$P(C0XPRE,":",1)
233 . . . S ZA=$P(C0XPRE,":",2)
234 . . . I $G(C0XVOC(ZB))'="" D ;
235 . . . . S C0XPRE=C0XVOC(ZB)_ZA ; expanded
236 . . S ZY=$G(@ZDOM@(ZJ,"A","rdf:resource")) ; potential object
237 . . I ZY'="" D Q ;
238 . . . S C0XOBJ=ZY ; object
239 . . . D ADD(ZGRF,C0XSUB,C0XPRE,C0XOBJ) ; finally. our first real triple
240 . . ; -- this is an else because of the quit above
241 . . S ZX=$G(@ZDOM@(ZJ,"A","rdf:nodeID")) ; fishing for nodeId object
242 . . I ZX'="" D Q ; got one
243 . . . S C0XOBJ=ZX ; we are using the incoming nodeIDs as object/subject
244 . . . ; without change... this could be foolish .. look at it again later
245 . . . D ADD(ZGRF,C0XSUB,C0XPRE,C0XOBJ) ; go for it and add a node
246 . . S C0XOBJ=$G(@ZDOM@(ZJ,"T",1)) ; hopefully an object is here
247 . . I C0XOBJ="" D Q ; not a happy situation
248 . . . W !,"ERROR, NO OBJECT FOUND FOR NODE: ",ZJ
249 . . D ADD(ZGRF,C0XSUB,C0XPRE,C0XOBJ) ; go for it and add a node
250 W !,"INSERTING ",C0XCNT," TRIPLES"
251 D UPDIE(.C0XFDA) ; commit the updates to the file
252 ; next, mark the graph as finished
253 S C0XEND=$$NOW^XLFDT
254 W !," ENDED AT: ",C0XEND
255 S C0XDIFF=$$FMDIFF^XLFDT(C0XEND,C0XSTART,2)
256 W !," ELAPSED TIME: ",C0XDIFF," SECONDS"
257 W !," APPROXIMATELY ",$P(C0XCNT/C0XDIFF,".")," TRIPLES PER SECOND"
258 Q
259 ;
260SHOW(ZN) ;
261 ZWR ^TMP("MXMLDOM",$J,1,ZN,*)
262 Q
263 ;
264ANONS() ; RETURNS AN ANONOMOUS SUBJECT
265 Q "_S:"_$$LKY9
266 ;
267NEWG(NGRAPH,NMETA) ; CREATES A NEW META GRAPH, MARKS IT AS UNFINISHED
268 ; THEN CREATES A NEW GRAPH AND POINTS THE METAGRAPH TO IT
269 ; NGRAPH AND NMETA ARE PASSED BY REFERENCE AND ARE THE RETURN
270 S NGRAPH="G"_$$LKY9
271 S NMETA=NGRAPH_"A"
272 Q
273 ;
274ADD(ZG,ZS,ZP,ZO,FARY) ; ADD A TRIPLE TO THE TRIPLESTORE. ALL VALUES ARE TEXT
275 ; THE FDA IS SET UP BUT THE FILES ARE NOT UPDATED. CALL UPDIE TO COMPLETE
276 I '$D(FARY) D ;
277 . D INITFARY("C0XFARY")
278 . S FARY="C0XFARY"
279 D USEFARY(FARY)
280 I '$D(C0XCNT) S C0XCNT=0
281 N ZNODE
282 S ZNODE="N"_$$LKY17
283 N ZNARY ; GET READY TO CALL IENOFA
284 S ZNARY("ZG",ZG)=""
285 S ZNARY("ZS",ZS)=""
286 S ZNARY("ZP",ZP)=""
287 S ZNARY("ZO",ZO)=""
288 D IENOFA(.ZIENS,.ZNARY,FARY) ; RESOLVE/ADD STRINGS
289 ;S ZGIEN=$$IENOF(ZG) ; LAYGO TO GET IEN
290 ;S ZSIEN=$$IENOF(ZS)
291 ;S ZPIEN=$$IENOF(ZP)
292 ;S ZOIEN=$$IENOF(ZO)
293 ;I $D(C0XFDA) D UPDIE ; ADD THE STRINGS IF NEEDED
294 S C0XCNT=C0XCNT+1
295 S C0XFDA(C0XTFN,"?+"_C0XCNT_",",.01)=ZNODE
296 S C0XFDA(C0XTFN,"?+"_C0XCNT_",",.02)=$O(ZIENS("IEN","ZG",""))
297 S C0XFDA(C0XTFN,"?+"_C0XCNT_",",.03)=$O(ZIENS("IEN","ZS",""))
298 S C0XFDA(C0XTFN,"?+"_C0XCNT_",",.04)=$O(ZIENS("IEN","ZP",""))
299 S C0XFDA(C0XTFN,"?+"_C0XCNT_",",.05)=$O(ZIENS("IEN","ZO",""))
300 ; REMEMBER TO CALL UPDIE WHEN YOU'RE DONE
301 Q
302 ;
303LKY9() ;EXTRINIC THAT RETURNS A RANDOM 9 DIGIT NUMBER. USED FOR GENERATING
304 ; UNIQUE NODE AND GRAPH NAMES
305 N ZN,ZI
306 S ZN=""
307 F ZI=1:1:9 D ;
308 . S ZN=ZN_$R(10)
309 Q ZN
310 ;
311LKY17() ;EXTRINIC THAT RETURNS A RANDOM 9 DIGIT NUMBER. USED FOR GENERATING
312 ; UNIQUE NODE AND GRAPH NAMES
313 N ZN,ZI
314 S ZN=""
315 F ZI=1:1:17 D ;
316 . S ZN=ZN_$R(10)
317 Q ZN
318 ;
319IENOF(ZSTRING,FARY) ; EXTRINSIC WHICH RETURNS THE IEN OF ZS IN THE STRINGS FILE
320 I '$D(FARY) D ;
321 . D INITFARY("C0XFARY")
322 . S FARY="C0XFARY"
323 N ZIEN
324 S ZIEN=$O(@C0XSN@("B",ZSTRING,""))
325 I ZIEN="" D ;
326 . S C0XFDA2(C0XSFN,"+1,",.01)=ZSTRING
327 . D UPDIE(.C0XFDA2)
328 . S ZIEN=$O(@C0XSN@("B",ZSTRING,""))
329 . K C0XFDA2
330 Q ZIEN
331 ;
332IENOFA(ZOUTARY,ZINARY,FARY) ; RESOLVE STRINGS TO IEN IN STRINGS FILE
333 ; OR ADD THEM IF
334 ; MISSING. ZINARY AND ZOUTARY ARE PASSED BY REFERENCE
335 ; ZINARY LOOKS LIKE ZINARY("VAR","VAL")=""
336 ; RETURNS IN ZOUTARY OF THE FORM ZOUTARY("IEN","VAR",IEN)=""
337 I '$D(FARY) D ;
338 . D INITFARY("C0XFARY")
339 . S FARY="C0XFARY"
340 K ZOUTARY ; START WITH CLEAN RESULTS
341 K C0XFDA2 ; USE A SEPARATE FDA FOR THIS
342 N ZI S ZI=""
343 N ZV,ZIEN
344 N ZCNT S ZCNT=0
345 F S ZI=$O(ZINARY(ZI)) Q:ZI="" D ; LOOK FOR MISSING STRINGS
346 . S ZV=$O(ZINARY(ZI,""))
347 . I $O(@C0XSN@("B",ZV,""))="" D ;
348 . . S ZCNT=ZCNT+1
349 . . S C0XFDA2(C0XSFN,"+"_ZCNT_",",.01)=ZV
350 I $D(C0XFDA2) D ;
351 . D UPDIE(.C0XFDA2) ; ADD MISSING STRINGS
352 . K C0XFDA2 ; CLEAN UP
353 F S ZI=$O(ZINARY(ZI)) Q:ZI="" D ; NOT GET ALL IENS
354 . S ZV=$O(ZINARY(ZI,""))
355 . S ZIEN=$O(@C0XSN@("B",ZV,"")) ; THEY SHOULD BE THERE NOW
356 . I ZIEN="" D ;
357 . . W !,"ERROR ADDING STRING: ",ZV
358 . . B
359 . S ZOUTARY("IEN",ZI,ZIEN)=""
360 Q
361 ;
362UPDIE(ZFDA) ; INTERNAL ROUTINE TO CALL UPDATE^DIE AND CHECK FOR ERRORS
363 ; ZFDA IS PASSED BY REFERENCE
364 ;ZWR ZFDA
365 ;B
366 K ZERR
367 D CLEAN^DILF
368 D UPDATE^DIE("","ZFDA","","ZERR")
369 I $D(ZERR) S ZZERR=ZZERR ; ZZERR DOESN'T EXIST,
370 ; INVOKE THE ERROR TRAP IF TASKED
371 ;. W "ERROR",!
372 ;. ZWR ZERR
373 ;. B
374 K ZFDA
375 Q
376 ;
Note: See TracBrowser for help on using the repository browser.