Guys, you could use http://gist.github.com to share long code snippets. That way code can have good Clojure syntax highlighting and be readable.
-- Anil — Sent from Mailbox On Wed, Oct 15, 2014 at 5:32 AM, bernardH <un.compte.pour.tes...@gmail.com> wrote: > Sorry but I don't feel that I'm understanding your code enough to rewrite > it (for instance, why substract 2 at > https://gist.github.com/ashishnegi/a9ae3fb3c270c7d3742e#file-calculator-clj-L102 > > to add 2 at > https://gist.github.com/ashishnegi/a9ae3fb3c270c7d3742e#file-calculator-clj-L16 > > ?). > But if I was to write it, I'd use a library to parse the expressions (see > https://github.com/Engelberg/instaparse ). If you really want to do it > yourself, the lexing step ( > make-calulator-list)shouldn't return strings but numbers and keywords for > operations imo. And from the list you could > build an Abstract Syntax Tree either with nested vectors or if speed is of > essence with records (or maybe looking into > https://github.com/ztellman/clj-tuple ). Processing the AST should be then > much simpler than processing your strings list. > Try to avoid recursion if possible by expressing your proccessing with reduce > for instance. > BTW, are you sure than int is a large enough type ? I'd use long and check > that the mod is small enough. > My two cents. > On Monday, October 13, 2014 3:10:15 PM UTC+2, Ashish Negi wrote: >> >> I am competing in an online competition and the scenario is like this : >> I coded with my basic knowledge of clojure and got 11 out of 20 testcases >> and others being timed out and one being runtime error.. >> >> I started optimizing my code and read >> https://groups.google.com/forum/#!searchin/clojure/performance/clojure/SxT2MaOsip8/xoMkDVHeOqMJ >> and other google articles for performance.. but after that i just managed >> to solve only one more problem and i am stuck at 12 solved.. ( the time for >> each case is 8 seconds max) >> The last 12th was completed in 7.05 secs :) .. just on the border huh... :) >> >> The problem is about that of Solving calculataor expressions like >> "2+3/2/2/(2+8)" mod 10^8+7 gives answer 10^8+3.. >> Nevertheless it is the concern but optimizing code is :) >> >> This is current condition of my code : >> >> (ns fpcontest.expression) >> >> (def Term) >> (def Factor) >> >> (def ToMod (int (+ 1000000000 7))) >> >> ;; (int ToMod) >> >> (def toPrint false) >> >> (defn myprintln [& args] >> ;; (if toPrint >> ;; (apply println args)) >> ) >> >> (defn EulerDivNonMemo [^long x ^long p] >> (let [ToMod (+ p 2)] >> (loop [num (long 1) toPow p numDouble x] >> (if (= 0 toPow) >> ;;(do (myprintln " EulerDiv : " num " for " x p)) >> num >> (let [numDouble2 (rem (* numDouble numDouble) ToMod) >> halfToPow (bit-shift-right toPow 1)] >> (if (odd? toPow) >> (recur (rem (* num numDouble) ToMod) >> halfToPow >> numDouble2) >> (recur num halfToPow numDouble2)) >> >> )))) >> ) >> >> (def EulerDiv (memoize EulerDivNonMemo) >> ) >> >> ;; (EulerDiv 2 2) >> ;; (defn TestEulerDiv [] >> ;; (and >> ;; (= 2 (mod (* 4 (EulerDiv 2 (- 3 2))) 3)) >> ;; (= 1 (mod (* 2 (EulerDiv 2 (- 3 2))) 3)) >> ;; (= 1 (int (mod (* 2 (EulerDiv 2 (- ToMod 2))) ToMod)))) >> ;; ) >> >> ;; (= true (TestEulerDiv)) >> >> (defn Expression [lst] >> (let [term (Term lst) >> ;; p (myprintln term " term for exp " lst) >> sizeTerm (count term) >> evaluateExpr >> (fn [opr lst] >> (let [expr (Expression (drop 2 lst)) >> ;; p (myprintln expr " expr for Expr " lst " and " opr) >> intLst (first lst)] >> (cons (rem (opr intLst >> (first expr)) >> ToMod) (drop 1 expr))))] >> (if (> sizeTerm 2) >> (if (= (second term) "+") >> (evaluateExpr + term) >> (if (= (second term) "-") >> (evaluateExpr - term) >> term) >> ) >> ;; first of term is the answer >> term >> ) >> )) >> >> >> (defn Term [lst] >> (let [factor (Factor lst) >> ;; p (myprintln factor " factor for term " lst) >> sizeFactor (count factor) >> evaluateExpr >> (fn [opr lst] >> (let [expr (Term (drop 2 lst)) >> ;;p (myprintln expr " term for expr " >> lst " and " opr) >> intLst (first lst)] >> (cons (rem (opr intLst >> (first expr)) >> ToMod) (drop 1 expr) )))] >> (if (> sizeFactor 2) >> (if (= (second factor) "*") >> (evaluateExpr * factor) >> (if (= (second factor) "/") >> (let [expr (Term (drop 2 factor)) >> fExpr (first expr)] >> (cons (rem (* (first factor) (EulerDiv fExpr (- ToMod 2))) >> ToMod) (drop 1 expr)) >> ) >> ;;(evaluateExpr / factor) >> factor) >> ) >> factor >> ) >> )) >> >> (defn Factor [lst] >> (let [fFactor (first lst) >> ;;p (myprintln " Factor has : " lst) >> ] >> (if (= fFactor "+") >> (cons (int (Integer. (second lst))) (rest lst)) >> (if (= fFactor "-") >> (cons (- (int (Integer. (second lst)))) (rest (rest lst))) >> (if (= fFactor "(") >> ;; remove the '(' and ')' >> (let [expr (Expression (rest lst))] >> (cons (first expr) (drop 2 expr)) >> ) >> ;; (if (= (count lst) 1)) >> (cons (int (Integer. (first lst))) (rest lst)) >> ;;(myprintln "I do not know where i am... " lst) >> >> ) >> )) >> ) >> >> (defn split-with-delim [d s] >> (clojure.string/split >> s (re-pattern (str "(?=" d >> ")|(?<=" d >> ")")))) >> >> (defn make-calulator-list [s] >> (->> s >> (split-with-delim #"[\/\+\-\*\(\) ]") >> (filter #(not (or (= "" %) (= " " %)))) >> doall >> )) >> >> (defn StartRun [FuncToCall inputParse outputParse] >> (let [lines (line-seq (java.io.BufferedReader. *in*))] >> (outputParse (FuncToCall (inputParse (first lines))))) >> ) >> >> >> and this is how i am testing my code ................................ >> ****************** Test ******************************** >> >> >> >> >> (defn RunTests [] >> (and >> (= 1717 (int (first (Expression (make-calulator-list " 22 * 79 - >> 21"))))) >> >> (= 12 (int (first (Expression (make-calulator-list " 4/2/2 + 8"))))) >> >> (= 4 (int (first (Expression (make-calulator-list "4/-2/2 + 8"))))) >> >> (= 999998605 (int (mod (first (Expression (make-calulator-list >> "55+3-45*33-25"))) ToMod))) >> >> (= 999999987 (int (mod (first (Expression (make-calulator-list >> "4/-2/(2+8)"))) ToMod))) >> >> (= 22 (int (first (Expression (make-calulator-list "(22)*+1/-1+(1)"))))) >> >> (= 22 (int (first (Expression (make-calulator-list "11*(2/2)*2"))))) >> >> (= 11 (int (first (Expression (make-calulator-list "22*2/2*2"))))) >> >> (= '("4" "/" "-" "2" "/" "2" "+" "8") >> (make-calulator-list "4/-2/2 + 8")) >> (= '("4" "/" "-" "2" "/" "(" "2" "+" "8" ")") >> (make-calulator-list " 4/-2/(2 + 8)")) >> >> ) >> ) >> >> (RunTests) >> >> (Expression '("3" "+" "2")) >> >> (conj '("3") 1) >> (cons 2333 '(2 3 4)) >> (cons 2 (drop 1 '())) >> (drop 1 '()) >> >> (str (int 1.0)) >> (cons "1" '(1 "3")) >> (cons "1" nil) >> (int (Integer. "1")) >> >> (int (mod (first (Expression (make-calulator-list "4/-2/(2+8)"))) ToMod)) >> >> >> (mod (first (Expression (make-calulator-list "4/2"))) ToMod) >> >> >> (defn diffTime [func & args] >> (let [start (.getTime (java.util.Date.)) >> res (apply func args) >> end (.getTime (java.util.Date.))] >> (- end start))) >> >> (defn somuchtime= [times] >> (let [start (java.util.Date.) >> ] >> (loop [iTime 0] >> (if (= iTime times) >> start >> (recur (inc iTime)))))) >> >> >> (defn GenerateCalculatorTests [nums] >> (loop [num 0 sb (StringBuilder.)] >> (if (= num nums) >> (str "1" (.toString sb)) >> (->> sb >> ((fn [sb] >> (let [r (rand-int 500)] >> (do (if (< r 100) >> (.append sb "+") >> (if (< r 200) >> (.append sb "-") >> (if (< r 300) >> (.append sb "*") >> (if (< r 497) >> (.append sb "/") >> (.append sb (str >> "*(" >> (GenerateCalculatorTests 4) >> ")+")))))) >> (.append sb (rand-int 1000000000)) >> sb)))) >> (recur (inc num)) >> )) >> ) >> ) >> >> (GenerateCalculatorTests 50) >> >> (defn Execute [num] >> (let [str (GenerateCalculatorTests num)] >> (do (println str " = " ) (Expression (make-calulator-list >> (GenerateCalculatorTests num)))))) >> >> (diffTime Execute 2074) >> >> >> >> I find that the last problem which is giving runtime error is may be >> StackOverflow as i get it in with >> (diffTime Execute 3000) ;; where 3000 tries to get a calculator expression >> of 3000 or more ints.. see GenerateCalculatorTests.. >> >> >> with diffTime i also found that using rem is faster than using mod.. >> > -- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To post to this group, send email to clojure@googlegroups.com > Note that posts from new members are moderated - please be patient with your > first post. > To unsubscribe from this group, send email to > clojure+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/clojure?hl=en > --- > You received this message because you are subscribed to the Google Groups > "Clojure" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to clojure+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
