X-Git-Url: http://git.kpe.io/?a=blobdiff_plain;f=sql-classes.lisp;h=a7b080414cfebc599c8e3e77bc637d888646e506;hb=188873f068b0c53febe4ee0ededbc755fce4869d;hp=904e77d30a24f2ddcc09d54a5bf6dc5e246fdc56;hpb=a4851b680b8430b9792d1e86318657de16cf0a96;p=umlisp.git diff --git a/sql-classes.lisp b/sql-classes.lisp index 904e77d..a7b0804 100644 --- a/sql-classes.lisp +++ b/sql-classes.lisp @@ -2,12 +2,12 @@ ;;;; ************************************************************************* ;;;; FILE IDENTIFICATION ;;;; -;;;; Name: sql-classes.lisp -;;;; Purpose: Routines for reading UMLS objects from SQL database -;;;; Author: Kevin M. Rosenberg -;;;; Date Started: Apr 2000 +;;;; Name: sql-classes.lisp +;;;; Purpose: Routines for reading UMLS objects from SQL database +;;;; Author: Kevin M. Rosenberg +;;;; Created: Apr 2000 ;;;; -;;;; $Id: sql-classes.lisp,v 1.86 2003/06/24 00:27:59 kevin Exp $ +;;;; $Id$ ;;;; ;;;; This file, part of UMLisp, is ;;;; Copyright (c) 2000-2003 by Kevin M. Rosenberg, M.D. @@ -41,15 +41,16 @@ ,%%fields ,@(when %%where (list %%where)) ,@(when %%where - `((typecase ,where-value - (fixnum - (prefixed-fixnum-string ,where-value #\= 10)) - (number - (concatenate 'string "=" (write-to-string ,where-value))) - (null - " is null") - (t - (format nil ,(if like "'%~A%'" "='~A'") ,where-value))))) + `((typecase ,where-value + #+ignore + (fixnum + (concatenate 'string "='" (prefixed-fixnum-string ,where-value #\0 10) "'")) + (number + (concatenate 'string "='" (write-to-string ,where-value) "'")) + (null + " is null") + (t + (format nil ,(if like "'%~A%'" "='~A'") ,where-value))))) (if ,srl (concatenate 'string ,%%lrl (write-to-string ,srl)) "") ,@(when %%order (list %%order)) ,@(when single (list " limit 1"))))) @@ -64,7 +65,7 @@ (if where-name (format nil (typecase where-value - (number "=~D") + (number "='~D'") (null " is null") (t (if like " like '%~A%""='~A'"))) @@ -76,10 +77,11 @@ (defmacro umlisp-query (table fields srl where-name where-value - &key (lrl "KCUILRL") single distinct order like) + &key (lrl "KCUILRL") single distinct order like + (query-cmd 'mutex-sql-query)) "Query the UMLisp database. Return a list of umlisp objects whose name is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" - `(mutex-sql-query + `(,query-cmd (query-string ,table ,fields ,srl ,where-name ,where-value :lrl ,lrl :single ,single :distinct ,distinct :order ,order :like ,like))) @@ -94,8 +96,8 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" ;; only WHERE-VALUE and SRL are evaluated (defmacro collect-umlisp-query ((table fields srl where-name where-value &key (lrl "KCUILRL") distinct single - order like) - &body body) + order like (query-cmd 'mutex-sql-query)) + &body body) (let ((value (gensym)) (r (gensym))) (if single @@ -103,7 +105,8 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" (tuple (car (umlisp-query ,table ,fields ,srl ,where-name ,value :lrl ,lrl :single ,single :distinct ,distinct :order ,order - :like ,like)))) + :like ,like + :query-cmd ,query-cmd)))) ,@(unless where-name `((declare (ignore ,value)))) (when tuple (destructuring-bind ,fields tuple @@ -254,6 +257,12 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" :pfstr kpfstr :lrl (ensure-integer kcuilrl)))) +(defun find-cui-ucon-all (&key (srl *current-srl*)) + "Return list of CUIs for all ucons" + (collect-umlisp-query (mrcon (cui) srl nil nil :order (cui asc) + :distinct t) + cui)) + (defun map-ucon-all (fn &key (srl *current-srl*)) "Map a function over all ucon's" (with-sql-connection (db) @@ -296,6 +305,13 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" :cui2 (ensure-integer cui2) :rela rela :sab sab :sl sl :mg mg :pfstr2 kpfstr2))) +(defun find-cui2-urel-cui (cui &key (srl *current-srl*)) + "Return a list of urel for cui" + (ensure-cui-integer cui) + (collect-umlisp-query (mrrel (cui2) srl cui1 + cui :lrl "KSRL") + cui2)) + (defun find-urel-cui2 (cui2 &key (srl *current-srl*)) "Return a list of urel for cui2" (ensure-cui-integer cui2) @@ -317,7 +333,7 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" (collect-umlisp-query (mrcoc (cui2 soc cot cof coa kpfstr2) srl cui1 cui :lrl klrl :order (cof asc)) (setq cui2 (ensure-integer cui2)) - (when (zerop cui2) (setq cui2 nil)) + (when (eql 0 cui2) (setq cui2 nil)) (make-instance 'ucoc :cui1 cui :cui2 (ensure-integer cui2) :soc soc :cot cot :cof (ensure-integer cof) :coa coa :pfstr2 kpfstr2))) @@ -464,13 +480,14 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" (ensure-sui-integer sui) (let ((ls "select CODE,ATN,SAB,ATV from MRSAT where ")) (cond - (sui (string-append ls "KCUISUI=" - (integer-string (make-cuisui cui sui) 14))) - (lui (string-append ls "KCUILUI=" + (sui (string-append ls "KCUISUI='" + (integer-string (make-cuisui cui sui) 14) + "'")) + (lui (string-append ls "KCUILUI='" (integer-string (make-cuilui cui lui) 14) - " and sui=0")) - (t (string-append ls "cui=" (prefixed-fixnum-string cui nil 7) - " and lui=0 and sui=0"))) + "' and sui='0'")) + (t (string-append ls "cui='" (prefixed-fixnum-string cui nil 7) + "' and lui='0' and sui='0'"))) (when srl (string-append ls " and KSRL<=" (prefixed-fixnum-string srl nil 3))) (loop for tuple in (mutex-sql-query ls) collect @@ -495,12 +512,12 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" (defun find-usab-all () "Find usab for a key" - (collect-umlisp-query (mrsab (vcui rcui vsab rsab son sf sver mstart mend imeta + (collect-umlisp-query (mrsab (vcui rcui vsab rsab son sf sver vstart vend imeta rmeta slc scc srl tfr cfr cxty ttyl atnl lat cenc curver sabin) nil nil nil) (make-instance 'usab :vcui (ensure-integer vcui) :rcui (ensure-integer rcui) :vsab vsab :rsab rsab :son son - :sf sf :sver sver :mstart mstart :mend mend :imeta imeta + :sf sf :sver sver :vstart vstart :vend vend :imeta imeta :rmeta rmeta :slc slc :scc scc :srl (ensure-integer srl) :tfr (ensure-integer tfr) :cfr (ensure-integer cfr) :cxty cxty :ttyl ttyl :atnl atnl :lat lat :cenc cenc @@ -508,13 +525,13 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" (defun find-usab-by-key (key-name key) "Find usab for a key" - (collect-umlisp-query-eval ('mrsab '(vcui rcui vsab rsab son sf sver mstart - mend imeta rmeta slc scc srl tfr cfr cxty + (collect-umlisp-query-eval ('mrsab '(vcui rcui vsab rsab son sf sver vstart + vend imeta rmeta slc scc srl tfr cfr cxty ttyl atnl lat cenc curver sabin) nil key-name key :single t) (make-instance 'usab :vcui (ensure-integer vcui) :rcui (ensure-integer rcui) :vsab vsab :rsab rsab :son son - :sf sf :sver sver :mstart mstart :mend mend :imeta imeta + :sf sf :sver sver :vstart vstart :vend vend :imeta imeta :rmeta rmeta :slc slc :scc scc :srl (ensure-integer srl) :tfr (ensure-integer tfr) :cfr (ensure-integer cfr) :cxty cxty :ttyl ttyl :atnl atnl :lat lat :cenc cenc @@ -551,6 +568,24 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" :lrl 'klrl :order '(cui asc)) (find-ucon-cui cui :srl srl))) +(defun find-cui-normalized-word (word &key (srl *current-srl*) (like nil)) + "Return list of cui that match word, optionally use SQL's LIKE syntax" + (collect-umlisp-query-eval ('mrxnw_eng '(cui) srl 'nwd word :like like :distinct t + :lrl 'klrl :order '(cui asc)) + cui)) + +(defun find-lui-normalized-word (word &key (srl *current-srl*) (like nil)) + "Return list of cui that match word, optionally use SQL's LIKE syntax" + (collect-umlisp-query-eval ('mrxnw_eng '(lui) srl 'nwd word :like like :distinct t + :lrl 'klrl :order '(cui asc)) + lui)) + +(defun find-sui-normalized-word (word &key (srl *current-srl*) (like nil)) + "Return list of cui that match word, optionally use SQL's LIKE syntax" + (collect-umlisp-query-eval ('mrxnw_eng '(sui) srl 'nwd word :like like :distinct t + :lrl 'klrl :order '(cui asc)) + sui)) + (defun find-ustr-word (word &key (srl *current-srl*)) "Return list of ustrs that match word" (collect-umlisp-query (mrxw_eng (cui sui) srl wd word :lrl klrl @@ -565,15 +600,15 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" (defun find-uterm-word (word &key (srl *current-srl*)) "Return list of uterms that match word" - (collect-umlisp-query (mrxw_eng (cui sui) srl wd word :lrl klrl - :order (cui asc sui asc)) - (find-uterm-cuisui (ensure-integer cui) (ensure-integer sui) :srl srl))) + (collect-umlisp-query (mrxw_eng (cui lui) srl wd word :lrl klrl + :order (cui asc lui asc)) + (find-uterm-cuilui (ensure-integer cui) (ensure-integer lui) :srl srl))) (defun find-uterm-normalized-word (word &key (srl *current-srl*)) "Return list of uterms that match word" - (collect-umlisp-query (mrxnw_eng (cui sui) srl nwd word :lrl klrl - :order (cui asc sui asc)) - (find-uterm-cuisui (ensure-integer cui) (ensure-integer sui) :srl srl))) + (collect-umlisp-query (mrxnw_eng (cui lui) srl nwd word :lrl klrl + :order (cui asc lui asc)) + (find-uterm-cuilui (ensure-integer cui) (ensure-integer lui) :srl srl))) (defun find-ucon-noneng-word (word &key (srl *current-srl*) (like nil)) "Return list of ucons that match non-english word" @@ -595,25 +630,38 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" ;;; Multiword lookup and score functions -(defun find-ucon-multiword (str &key (srl *current-srl*)) - "Return sorted list of ucon's that match a multiword string" - (let* ((words (delimited-string-to-list str #\space)) - (ucons '())) - (dolist (word words) - (setq ucons (append ucons (find-ucon-word word :srl srl)))) - (sort-score-ucon-str str (delete-duplicates ucons :test #'eql :key #'cui)))) - -(defun find-ustr-multiword (str &key (srl *current-srl*)) - "Return sorted list of ustr's that match a multiword string" - (let* ((words (delimited-string-to-list str #\space)) - (ustrs '())) - (dolist (word words) - (setq ustrs (append ustrs (find-ustr-word word :srl srl)))) - (sort-score-ustr-str str (delete-duplicates ustrs :test #'eql :key #'cui)))) +(defun find-uobj-multiword (str obj-lookup-fun sort-fun key srl + only-exact-if-match) + (let ((uobjs '())) + (dolist (word (delimited-string-to-list str #\space)) + (setq uobjs (append uobjs (funcall obj-lookup-fun word :srl srl)))) + (let ((sorted + (funcall sort-fun str + (delete-duplicates uobjs :test #'= :key key)))) + (if (and (plusp (length sorted)) + only-exact-if-match + (multiword-match str (pfstr (first sorted)))) + (first sorted) + sorted)))) + +(defun find-ucon-multiword (str &key (srl *current-srl*) + (only-exact-if-match t)) + (find-uobj-multiword str #'find-ucon-word #'sort-score-pfstr-str + #'cui srl only-exact-if-match)) + +(defun find-uterm-multiword (str &key (srl *current-srl*) + (only-exact-if-match t)) + (find-uobj-multiword str #'find-uterm-word #'sort-score-pfstr-str + #'lui srl only-exact-if-match)) + +(defun find-ustr-multiword (str &key (srl *current-srl*) + (only-exact-if-match t)) + (find-uobj-multiword str #'find-ustr-word #'sort-score-ustr-str + #'sui srl only-exact-if-match)) -(defun sort-score-ucon-str (str ucons) +(defun sort-score-pfstr-str (str uobjs) "Return list of sorted and scored ucons. Score by match of str to ucon-pfstr" - (sort-score-umlsclass-str ucons str #'pfstr)) + (sort-score-umlsclass-str uobjs str #'pfstr)) (defun sort-score-ustr-str (str ustrs) "Return list of sorted and scored ucons. Score by match of str to ucon-pfstr" @@ -627,39 +675,6 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" scored)) (mapcar #'car (sort scored #'> :key #'cadr)))) -(defun score-multiword-match (s1 s2) - "Score a match between two strings with s1 being reference string" - (let* ((word-list-1 (delimited-string-to-list s1 #\space)) - (word-list-2 (delimited-string-to-list s2 #\space)) - (n1 (length word-list-1)) - (n2 (length word-list-2)) - (unmatched n1) - (score 0) - (nlong 0) - (nshort 0) - short-list long-list) - (declare (fixnum n1 n2 nshort nlong score unmatched)) - (if (> n1 n2) - (progn - (setq nlong n1) - (setq nshort n2) - (setq long-list word-list-1) - (setq short-list word-list-2)) - (progn - (setq nlong n2) - (setq nshort n1) - (setq long-list word-list-2) - (setq short-list word-list-1))) - (decf score (- nlong nshort)) ;; reduce score for extra words - (dotimes (iword nshort) - (declare (fixnum iword)) - (kmrcl:aif (position (nth iword short-list) long-list :test #'string-equal) - (progn - (incf score (- 10 (abs (- kmrcl::it iword)))) - (decf unmatched)))) - (decf score (* 2 unmatched)) - score)) - ;;; LEX SQL functions @@ -735,9 +750,9 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" (make-instance 'ltyp :eui eui :bas bas :sca sca :typ typ))) (defun find-lwd-wrd (wrd) - (make-instance 'lwd :wrd + (make-instance 'lwd :wrd wrd :euilist (collect-umlisp-query (lrwd (eui) nil wrd wrd) - (ensure-integer eui)))) + (ensure-integer eui)))) ;;; Semantic Network SQL access functions @@ -768,3 +783,104 @@ is OBJNAME from TABLE where WHERE-NAME field = WHERE-VALUE with FIELDS" (defun find-sstr-styrl (styrl) (collect-umlisp-query (srstr (rl sty_rl2 ls) nil styrl styrl) (make-instance 'sstr :styrl styrl :rl rl :styrl2 sty_rl2 :ls ls))) + + +;;; ************************** +;;; Local Classes +;;; ************************** + + +(defun make-ustats () + (with-sql-connection (conn) + (ignore-errors (sql-execute "drop table USTATS" conn)) + (sql-execute "create table USTATS (NAME varchar(160), COUNT bigint, SRL integer)" conn) + + (dotimes (srl 4) + (insert-ustats-count conn "Concept Count" "MRCON" "distinct CUI" "KCUILRL" srl) + (insert-ustats-count conn "Term Count" "MRCON" "distinct KCUILUI" "KCUILRL" srl) + (insert-ustats-count conn "Distinct Term Count" "MRCON" "distinct LUI" "KLUILRL" srl) + (insert-ustats-count conn "String Count" "MRCON" "*" "LRL" srl) + (insert-ustats-count conn "Distinct String Count" "MRCON" "distinct SUI" "LRL" srl) + (insert-ustats-count conn "Associated Expression Count" "MRATX" "*" "KSRL" srl) + (insert-ustats-count conn "Context Count" "MRCXT" "*" "KSRL" srl) + (insert-ustats-count conn "Co-occuring Concept Count" "MRCOC" "*" "KLRL" srl) + (insert-ustats-count conn "Definition Count" "MRDEF" "*" "KSRL" srl) + (insert-ustats-count conn "Locator Count" "MRLO" "*" "KLRL" srl) + (insert-ustats-count conn "Rank Count" "MRRANK" "*" "KSRL" srl) + (insert-ustats-count conn "Relationship Count" "MRREL" "*" "KSRL" srl) + (insert-ustats-count conn "Semantic Type Count" "MRSTY" "*" "KLRL" srl) + (insert-ustats-count conn "Simple Attribute Count" "MRSAT" "*" "KSRL" srl) + (insert-ustats-count conn "Source Count" "MRSO" "*" "SRL" srl) + (insert-ustats-count conn "Word Index Count" "MRXW_ENG" "*" "KLRL" srl) + (insert-ustats-count conn "Normalized Word Index Count" "MRXNW_ENG" "*" "KLRL" srl) + (insert-ustats-count conn "Normalized String Index Count" "MRXNS_ENG" "*" "KLRL" srl) + (insert-ustats-count conn "Bonus Attribute Name Count" "BONUS_ATN" "*" nil srl) + (insert-ustats-count conn "Bonus Relationship Count" "BONUS_REL" "*" nil srl) + (insert-ustats-count conn "Bonus Source Abbreviation Count" "BONUS_SAB" "*" nil srl) + (insert-ustats-count conn "Bonus Term Type Count" "BONUS_TTY" "*" nil srl)) + (sql-execute "create index USTATS_SRL on USTATS (SRL)" conn)) + (find-ustats-all)) + +(defun insert-ustats-count (conn name table count-variable srl-control srl) + (insert-ustats conn name (find-count-table conn table srl count-variable srl-control) srl)) + +(defun find-count-table (conn table srl count-variable srl-control) + (cond + ((stringp srl-control) + (ensure-integer + (caar (sql-query (format nil "select count(~a) from ~a where ~a <= ~d" + count-variable table srl-control srl) + conn)))) + ((null srl-control) + (ensure-integer + (caar (sql-query (format nil "select count(~a) from ~a" + count-variable table ) + conn)))) + (t + (error "Unknown srl-control") + 0))) + +(defun insert-ustats (conn name count srl) + (sql-execute (format nil "insert into USTATS (name,count,srl) values ('~a',~d,~d)" + name count (if srl srl 3)) + conn)) + +(defun find-ustats-all (&key (srl *current-srl*)) + (if srl + (collect-umlisp-query (ustats (name count srl) nil srl srl + :order (name asc)) + (make-instance 'ustats :name name + :hits (ensure-integer count) + :srl (ensure-integer srl))) + (collect-umlisp-query (ustats (name count srl) nil nil nil + :order (name asc)) + (make-instance 'ustats :name name + :hits (ensure-integer count) + :srl (ensure-integer srl))))) + +(defun find-ustats-srl (srl) + (collect-umlisp-query (ustats (name count) nil srl srl :order (name asc)) + (make-instance 'ustats :name name :hits (ensure-integer count)))) + + + +(defun find-bsab-sab (sab) + (collect-umlisp-query (bonus_sab (name count) nil sab sab :single t) + (make-instance 'bsab :sab sab :name name :hits (ensure-integer count)))) + +(defun find-bsab-all () + (collect-umlisp-query (bonus_sab (sab name count) nil nil nil :order (sab asc)) + (make-instance 'bsab :sab sab :name name :hits (ensure-integer count)))) + +(defun find-btty-tty (tty) + (collect-umlisp-query (bonus_tty (name count) nil tty tty :single t) + (make-instance 'btty :tty tty :name name :hits (ensure-integer count)))) + +(defun find-btty-all () + (collect-umlisp-query (bonus_tty (tty name count) nil nil nil :order (tty asc)) + (make-instance 'btty :tty tty :name name :hits (ensure-integer count)))) + +(defun find-brel-rel (rel) + (collect-umlisp-query (bonus_rel (sab sl rel rela count) nil rel rel) + (make-instance 'brel :sab sab :sl sl :rel rel :rela rela + :hits (ensure-integer count))))