1 ;;;; -*- Mode: Lisp; Syntax: ANSI-Common-Lisp; Base: 10; Package: umlisp -*-
2 ;;;; *************************************************************************
3 ;;;; FILE IDENTIFICATION
5 ;;;; Name: parse-2002.lisp
6 ;;;; Purpose: Parsing and SQL insertion routines for UMLisp which may
7 ;;;; change from year to year
8 ;;;; Programmer: Kevin M. Rosenberg
9 ;;;; Date Started: Apr 2000
11 ;;;; $Id: parse-2002.lisp,v 1.7 2003/05/06 07:17:35 kevin Exp $
13 ;;;; This file, part of UMLisp, is
14 ;;;; Copyright (c) 2000-2002 by Kevin M. Rosenberg, M.D.
16 ;;;; UMLisp users are granted the rights to distribute and use this software
17 ;;;; as governed by the terms of the GNU General Public License.
18 ;;;; *************************************************************************
21 (declaim (optimize (speed 3) (safety 1) (compilation-speed 0) (debug 3)))
23 ;;; Pre-read data for custom fields into hash tables
24 (defvar *parse-hash-init?* nil)
26 (eval-when (:compile-toplevel :load-toplevel :execute)
27 (let ((pfstr-hash nil) ;;; Preferred concept strings by CUI
28 (cui-lrl-hash nil) ;;; LRL by CUI
29 (lui-lrl-hash nil) ;;; LRL by LUI
30 (cuisui-lrl-hash nil) ;;; LRL by CUISUI
31 (sab-srl-hash nil)) ;;; SRL by SAB
33 (defun make-parse-hash-table ()
37 (clrhash cui-lrl-hash)
38 (clrhash lui-lrl-hash)
39 (clrhash cuisui-lrl-hash)
40 (clrhash sab-srl-hash))
42 pfstr-hash (make-hash-table :size 800000)
43 cui-lrl-hash (make-hash-table :size 800000)
44 lui-lrl-hash (make-hash-table :size 1500000)
45 cuisui-lrl-hash (make-hash-table :size 1800000)
46 sab-srl-hash (make-hash-table :size 100 :test 'equal))))
48 (defun binit-hash-table (&optional (force-read nil))
49 (when (or force-read (not *parse-hash-init?*))
50 (make-parse-hash-table)
51 (setq *parse-hash-init?* t))
52 (with-buffered-umls-file (line "MRCON")
53 (let ((cui (parse-ui (aref line 0)))
54 (lui (parse-ui (aref line 3)))
55 (sui (parse-ui (aref line 5)))
56 (lrl (parse-integer (aref line 7))))
57 (unless (gethash cui pfstr-hash) ;; if haven't stored pfstr for cui
58 (if (and (string-equal (aref line 1) "ENG") ; LAT
59 (string-equal (aref line 2) "P") ; ts
60 (string-equal (aref line 4) "PF")) ; stt
61 (setf (gethash cui pfstr-hash) (aref line 6))))
62 (set-lrl-hash cui lrl cui-lrl-hash)
63 (set-lrl-hash lui lrl lui-lrl-hash)
64 (set-lrl-hash (make-cuisui cui sui) lrl cuisui-lrl-hash)))
65 (with-buffered-umls-file (line "MRSO")
66 (let ((sab (aref line 3)))
67 (unless (gethash sab sab-srl-hash) ;; if haven't stored
68 (setf (gethash sab sab-srl-hash) (aref line 6))))))
70 (defun init-hash-table (&optional (force-read nil))
71 (when (or force-read (not *parse-hash-init?*))
72 (make-parse-hash-table)
73 (setq *parse-hash-init?* t))
74 (with-umls-file (line "MRCON")
75 (let ((cui (parse-ui (nth 0 line)))
76 (lui (parse-ui (nth 3 line)))
77 (sui (parse-ui (nth 5 line)))
78 (lrl (parse-integer (nth 7 line))))
79 (unless (gethash cui pfstr-hash) ;; if haven't stored pfstr for cui
80 (if (and (string-equal (nth 1 line) "ENG") ; LAT
81 (string-equal (nth 2 line) "P") ; ts
82 (string-equal (nth 4 line) "PF")) ; stt
83 (setf (gethash cui pfstr-hash) (nth 6 line))))
84 (set-lrl-hash cui lrl cui-lrl-hash)
85 (set-lrl-hash lui lrl lui-lrl-hash)
86 (set-lrl-hash (make-cuisui cui sui) lrl cuisui-lrl-hash)))
87 (with-umls-file (line "MRSO")
88 (let ((sab (nth 3 line)))
89 (multiple-value-bind (val found) (gethash sab sab-srl-hash)
90 (declare (ignore val))
92 (setf (gethash sab sab-srl-hash) (parse-integer (nth 6 line))))))))
94 (defun pfstr-hash (cui)
95 (gethash cui pfstr-hash))
98 (gethash cui cui-lrl-hash))
101 (gethash lui lui-lrl-hash))
103 (defun cuisui-lrl (cuisui)
104 (gethash cuisui cuisui-lrl-hash))
107 (kmrcl:aif (gethash sab sab-srl-hash) kmrcl::it 0))
110 (defun set-lrl-hash (key lrl hash)
111 "Set the least restrictive level in hash table"
112 (multiple-value-bind (hash-lrl found) (gethash key hash)
113 (if (or (not found) (< lrl hash-lrl))
114 (setf (gethash key hash) lrl))))
116 ;; UMLS file and column structures
117 ;;; SQL datatypes symbols
118 ;;; sql-u - Unique identifier
119 ;;; sql-s - Small integer (16-bit)
120 ;;; sql-i - Integer (32-bit)
121 ;;; sql-l - Big integer (64-bit)
122 ;;; sql-f - Floating point
124 (defparameter +col-datatypes+
125 '(("AV" sql-f) ("BTS" sql-i) ("CLS" sql-i) ("COF" sql-i) ("CUI1" sql-u)
126 ("CUI2" sql-u) ("CUI" sql-u) ("CXN" sql-s) ("FR" sql-i) ("LRL" sql-s)
127 ("LUI" sql-u) ("MAX" sql-s) ("MIN" sql-s) ("RANK" sql-s) ("REF" sql-s)
128 ("RNK" sql-s) ("RWS" sql-i) ("SRL" sql-s) ("SUI" sql-u) ("TUI" sql-u)
130 ("KCUISUI" sql-l) ("KCUILUI" sql-l) ("KCUILRL" sql-i) ("KLUILRL" sql-i)
131 ("KSRL" sql-i) ("KLRL" sql-i)
133 ("EUI" sql-u) ("EUI2" sql-u)
134 ;;; Semantic net columns
135 ("UI" sql-u) ("UI2" sql-u) ("UI3" sql-u)
136 ;; New fields for 2002AD
137 ("RCUI" sql-u) ("VCUI" sql-u) ("CFR" sql-i) ("TFR" sql-i)
139 "SQL data types for each non-string column")
141 (defparameter +custom-tables+
144 '(("MRCONSO" "SELECT m.CUI, m.LAT, m.TS, m.LUI, m.STT, m.SUI, m.STR, m.LRL, s.SAB, s.TTY, s.SCD, s.SRL FROM MRCON m, MRSO s WHERE m.CUI=s.CUI AND m.LUI=s.LUI AND m.SUI=s.SUI")
145 ("MRCONFULL" "SELECT m.CUI, m.LAT, m.TS, m.LUI, m.STT, m.SUI, m.STR, m.LRL, s.SAB, s.TTY, s.SCD, s.SRL, t.TUI FROM MRCON m, MRSO s, MRSTY t WHERE m.CUI=s.CUI AND m.LUI=s.LUI AND m.SUI=s.SUI AND m.CUI=t.CUI AND s.CUI=t.CUI"))
146 "Custom tables to create")
148 (defparameter +custom-cols+
149 '(("MRCON" "KPFSTR" "TEXT" 1024
150 (lambda (x) (pfstr-hash (parse-ui (nth 0 x)))))
151 ("MRCON" "KCUISUI" "BIGINT" 0
152 (lambda (x) (format nil "~d" (make-cuisui (parse-ui (nth 0 x)) (parse-ui (nth 5 x))))))
153 ("MRCON" "KCUILUI" "BIGINT" 0
154 (lambda (x) (format nil "~d" (make-cuilui (parse-ui (nth 0 x)) (parse-ui (nth 3 x))))))
155 ("MRCON" "KCUILRL" "INTEGER" 0
156 (lambda (x) (format nil "~d" (cui-lrl (parse-ui (nth 0 x))))))
157 ("MRCON" "KLUILRL" "INTEGER" 0
158 (lambda (x) (format nil "~d" (lui-lrl (parse-ui (nth 3 x))))))
159 ("MRLO" "KLRL" "INTEGER" 0
160 (lambda (x) (format nil "~d"
161 (if (zerop (length (nth 4 x)))
162 (cui-lrl (parse-ui (nth 0 x)))
163 (cuisui-lrl (make-cuisui (parse-ui (nth 0 x)) (parse-ui (nth 4 x))))))))
164 ("MRSTY" "KLRL" "INTEGER" 0
165 (lambda (x) (format nil "~d" (cui-lrl (parse-ui (nth 0 x))))))
166 ("MRCOC" "KLRL" "INTEGER" 0
167 (lambda (x) (format nil "~d"
168 (max (cui-lrl (parse-ui (nth 0 x)))
169 (kmrcl:aif (cui-lrl (parse-ui (nth 1 x))) kmrcl::it 0)))))
170 ("MRSAT" "KSRL" "INTEGER" 0
171 (lambda (x) (format nil "~d" (sab-srl (nth 5 x)))))
172 ("MRREL" "KSRL" "INTEGER" 0
173 (lambda (x) (format nil "~d" (sab-srl (nth 4 x)))))
174 ("MRRANK" "KSRL" "INTEGER" 0
175 (lambda (x) (format nil "~d" (sab-srl (nth 1 x)))))
176 ("MRDEF" "KSRL" "INTEGER" 0
177 (lambda (x) (format nil "~d" (sab-srl (nth 1 x)))))
178 ("MRCXT" "KSRL" "INTEGER" 0
179 (lambda (x) (format nil "~d" (sab-srl (nth 2 x)))))
180 ("MRATX" "KSRL" "INTEGER" 0
181 (lambda (x) (format nil "~d" (sab-srl (nth 1 x)))))
182 ("MRXW.ENG" "KLRL" "INTEGER" 0
183 (lambda (x) (format nil "~d" (cuisui-lrl (make-cuisui
185 (parse-ui (nth 4 x)))))))
186 ("MRXW.NONENG" "KLRL" "INTEGER" 0
187 (lambda (x) (format nil "~d" (cuisui-lrl (make-cuisui
189 (parse-ui (nth 4 x)))))))
190 ("MRXNW.ENG" "KLRL" "INTEGER" 0
191 (lambda (x) (format nil "~d" (cuisui-lrl (make-cuisui
193 (parse-ui (nth 4 x)))))))
194 ("MRXNS.ENG" "KLRL" "INTEGER" 0
195 (lambda (x) (format nil "~d" (cuisui-lrl (make-cuisui
197 (parse-ui (nth 4 x)))))))
198 ("MRREL" "KPFSTR2" "TEXT" 1024
199 (lambda (x) (pfstr-hash (parse-ui (nth 2 x)))))
200 ("MRCOC" "KPFSTR2" "TEXT" 1024
201 (lambda (x) (pfstr-hash (parse-ui (nth 1 x)))))
202 ("MRCXT" "KCUISUI" "BIGINT" 0
203 (lambda (x) (format nil "~d" (make-cuisui (parse-ui (nth 0 x)) (parse-ui (nth 1 x))))))
204 ("MRSAT" "KCUILUI" "BIGINT" 0
205 (lambda (x) (format nil "~d" (make-cuilui (parse-ui (nth 0 x)) (parse-ui (nth 1 x))))))
206 ("MRSAT" "KCUISUI" "BIGINT" 0
207 (lambda (x) (format nil "~d" (make-cuisui (parse-ui (nth 0 x)) (parse-ui (nth 2 x))))))
208 ("MRSO" "KCUISUI" "BIGINT" 0
209 (lambda (x) (format nil "~d" (make-cuisui (parse-ui (nth 0 x)) (parse-ui (nth 2 x))))))
210 ("MRXW.ENG" "KCUISUI" "BIGINT" 0
211 (lambda (x) (format nil "~d" (make-cuisui (parse-ui (nth 2 x)) (parse-ui (nth 4 x))))))
212 ("MRXNW.ENG" "KCUISUI" "BIGINT" 0
213 (lambda (x) (format nil "~d" (make-cuisui (parse-ui (nth 2 x)) (parse-ui (nth 4 x))))))
214 ("MRXNS.ENG" "KCUISUI" "BIGINT" 0
215 (lambda (x) (format nil "~d" (make-cuisui (parse-ui (nth 2 x)) (parse-ui (nth 4 x))))))
216 ("MRXW.NONENG" "LAT" "CHAR" 3 (lambda (x) (nth 0 x)))
217 ("MRXW.NONENG" "WD" "CHAR" 200 (lambda (x) (nth 1 x)))
218 ("MRXW.NONENG" "CUI" "INTEGER" 0 (lambda (x) (parse-ui (nth 2 x))))
219 ("MRXW.NONENG" "LUI" "INTEGER" 0 (lambda (x) (parse-ui (nth 3 x))))
220 ("MRXW.NONENG" "SUI" "INTEGER" 0 (lambda (x) (parse-ui (nth 4 x))))
221 ("MRXW.NONENG" "KCUISUI" "BIGINT" 0
222 (lambda (x) (format nil "~d" (make-cuisui (parse-ui (nth 2 x)) (parse-ui (nth 4 x)))))))
223 "Custom columns to create.(filename, col, sqltype, value-func).")
225 (defparameter +index-cols+
226 '(("CUI" "MRATX") ("CUI1" "MRCOC") ("CUI" "MRCON") ("LUI" "MRCON")
228 ("SUI" "MRCON") ("CUI" "MRCXT") ("CUI" "MRDEF") ("CUI" "MRLO")
229 ("CUI1" "MRREL") ("CUI" "MRSAT") ("LUI" "MRSAT") ("SUI" "MRSAT")
230 ("CUI" "MRSO") ("SAB" "MRSO") ("SRL" "MRSO") ("CUI" "MRSTY")
231 ("TUI" "MRSTY") ("CUI" "MRXNS_ENG") ("NSTR" "MRXNS_ENG" 10)
232 ("CUI" "MRXNW_ENG") ("NWD" "MRXNW_ENG") ("WD" "MRXW_ENG")
233 ("KCUISUI" "MRCON") ("KCUILUI" "MRCON") ("KCUILRL" "MRCON")
234 ("KLUILRL" "MRCON") ("KCUISUI" "MRCXT")
235 ("KCUISUI" "MRSO") ("KCUISUI" "MRSAT") ("KCUILUI" "MRSAT")
236 ("KCUISUI" "MRXW_ENG") ("KCUISUI" "MRXNW_ENG")
237 ("KCUISUI" "MRXNS_ENG") ("KCUISUI" "MRXW_NONENG")
238 ("KSRL" "MRATX") ("KSRL" "MRCXT") ("KSRL" "MRDEF") ("KSRL" "MRRANK")
239 ("KSRL" "MRREL") ("KSRL" "MRSAT") ("KLRL" "MRCOC")
240 ("KLRL" "MRLO") ("KLRL" "MRSTY") ("KLRL" "MRXW_ENG") ("KLRL" "MRXNW_ENG")
241 ("KLRL" "MRXNS_ENG") ("KLRL" "MRXW_NONENG")
243 ("EUI" "LRABR") ("EUI2" "LRABR") ("EUI" "LRAGR") ("EUI" "LRCMP") ("EUI" "LRMOD")
244 ("EUI" "LRNOM") ("EUI2" "LRNOM") ("EUI" "LRPRN") ("EUI" "LRPRP") ("EUI" "LRSPL")
245 ("EUI" "LRTRM") ("EUI" "LRTYP") ("EUI" "LRWD") ("WRD" "LRWD")
247 ;; Semantic NET indices
248 ("UI" "SRSTRE1") ("UI2" "SRSTRE1") ("UI3" "SRSTRE1")
249 ("STY_RL" "SRDEF") ("RT" "SRDEF") ("STY_RL" "SRSTR") ("STY_RL2" "SRSTR")
251 ("SRL" "MRSAB") ("RSAB" "MRSAB") ("VSAB" "MRSAB") ("RCUI" "MRSAB")
252 ("VCUI" "MRSAB") ("LAT" "MRSAB"))
253 "Columns in files to index")
255 (defparameter +custom-index-cols+
258 '(("CUI" "MRCONFULL") ("SAB" "MRCONFULL") ("TUI" "MRCONFULL"))
259 "Indexes to custom tables")
261 ;; File & Column functions
263 (defun init-umls (&optional (alwaysclear nil))
264 "Initialize all UMLS file and column structures if not already initialized"
265 (when (or alwaysclear (null *umls-files*))
268 (init-field-lengths)))
270 (defun init-umls-cols ()
271 (setq *umls-cols* (append
274 (init-generic-cols "LRFLD")
275 (init-generic-cols "SRFLD"))))
277 (defun init-meta-cols ()
278 "Initialize all umls columns"
280 (with-umls-file (line "MRCOLS")
281 (destructuring-bind (col des ref min av max fil dty) line
282 (let ((c (make-instance 'ucol
286 :min (parse-integer min)
287 :av (read-from-string av)
288 :max (parse-integer max)
290 :dty dty ;; new in 2002 UMLS
291 :sqltype "VARCHAR" ; default data type
292 :parse-fun #'add-sql-quotes
293 :custom-value-fun nil
295 (add-datatype-to-col c (datatype-for-col col))
299 (defun init-custom-cols ()
300 "Initialize umls columns for custom columns"
302 (dolist (customcol +custom-cols+)
303 (let ((c (make-instance 'ucol
304 :col (nth 1 customcol) :des "" :ref 0 :min 0 :max (nth 3 customcol)
305 :av 0 :dty nil :fil (nth 0 customcol) :sqltype (nth 2 customcol)
306 :parse-fun #'add-sql-quotes :custom-value-fun (nth 4 customcol)
308 (add-datatype-to-col c (datatype-for-col (nth 1 customcol)))
312 (defun escape-column-name (name)
313 (substitute #\_ #\/ name))
315 (defun init-generic-cols (col-filename)
316 "Initialize for generic (LEX/NET) columns"
318 (with-umls-file (line col-filename)
319 (destructuring-bind (nam des ref fil) line
320 (setq nam (escape-column-name nam))
321 (dolist (file (delimited-string-to-list fil #\,))
322 (let ((c (make-instance 'ucol
331 :sqltype "VARCHAR" ; default data type
332 :parse-fun #'add-sql-quotes
333 :custom-value-fun nil
335 (add-datatype-to-col c (datatype-for-col nam))
339 (defun init-umls-files ()
340 (setq *umls-files* (append
341 (init-generic-files "MRFILES")
342 (init-generic-files "LRFIL")
343 (init-generic-files "SRFIL")))
344 ;; need to separate this since init-custom-files depends on *umls-files*
345 (setq *umls-files* (append *umls-files* (init-custom-files))))
348 (defun umls-field-string-to-list (fmt)
349 "Converts a comma delimited list of fields into a list of field names. Will
350 append a unique number (starting at 2) onto a column name that is repeated in the list"
351 (let ((field-list (delimited-string-to-list (escape-column-name fmt) #\,))
352 (col-count (make-hash-table :test 'equal)))
353 (dotimes (i (length field-list))
355 (let ((col (nth i field-list)))
356 (multiple-value-bind (key found) (gethash col col-count)
358 (let ((next-id (1+ key)))
359 (setf (nth i field-list) (concatenate 'string
361 (format nil "~D" next-id)))
362 (setf (gethash col col-count) next-id))
363 (setf (gethash col col-count) 1)))))
366 (defun init-generic-files (files-filename)
367 "Initialize all LEX file structures"
369 (with-umls-file (line files-filename)
370 (destructuring-bind (fil des fmt cls rws bts) line
371 (let ((f (make-instance 'ufile
373 :table (substitute #\_ #\. fil)
375 :fmt (escape-column-name fmt)
376 :cls (parse-integer cls)
377 :rws (parse-integer rws)
378 :bts (parse-integer bts)
379 :fields (concatenate 'list
380 (umls-field-string-to-list fmt)
381 (custom-colnames-for-filename fil)))))
382 (setf (ucols f) (ucols-for-umls-file f))
386 (defun init-custom-files ()
387 (let ((ffile (make-instance 'ufile
388 :fil "MRXW.NONENG" :des "Custom NonEnglish Index" :table "MRXW_NONENG"
389 :cls 5 :rws 0 :bts 0 :fields (fields (find-umls-file "MRXW.ENG")))))
391 (ucols-for-umls-file ffile))
394 (defun datatype-for-col (colname)
395 "Return datatype for column name"
396 (car (cdr (find colname +col-datatypes+ :key #'car :test #'string-equal))))
398 (defun add-datatype-to-col (col datatype)
399 "Add data type information to column"
400 (setf (datatype col) datatype)
402 (sql-u (setf (sqltype col) "INTEGER"
403 (parse-fun col) #'parse-ui
405 (sql-s (setf (sqltype col) "SMALLINT"
406 (parse-fun col) #'parse-integer
408 (sql-l (setf (sqltype col) "BIGINT"
409 (parse-fun col) #'parse-integer
411 (sql-i (setf (sqltype col) "INTEGER"
412 (parse-fun col) #'parse-integer
414 (sql-f (setf (sqltype col) "NUMERIC"
415 (parse-fun col) #'read-from-string
417 (t ; Default column type, optimized text storage
418 (setf (parse-fun col) #'add-sql-quotes
420 (when (and (cmax col) (av col))
421 (if (> (cmax col) 255)
422 (setf (sqltype col) "TEXT")
423 (if (< (- (cmax col) (av col)) 4)
424 (setf (sqltype col) "CHAR") ; if average bytes wasted < 4
425 (setf (sqltype col) "VARCHAR")))))))