grep-cascade.sh

   1 #!/bin/sh
   2
   3 # a cascade of greps to find all layouts which have values in a given range.
   4 # take the median of each value as fixpoint (the value of a typical layout).
   5 # TODO: recalculate all values for a reference sentence. Then work with these.
   6 # -> results/2010-08-17-all-recalculated-for-the-reference-text.txt
   7
   8 ### data
   9
  10 ## layout
  11
  12 # #### 21 ####
  13 # xuc.ü vdsljq´
  14 # miaeo btrnhk
  15 # y,zäö fgßwp
  16 # ┌───┬───┬───┬───┬───┬───┬───┬───┬───┬───┬───┬───┬───┬──────┐
  17 # │ ^ │ 1 │ 2 │ 3 │ 4 │ 5 │ 6 │ 7 │ 8 │ 9 │ 0 │ - │ ` │ ←    │
  18 # ├───┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬─────┤
  19 # │  ⇥ │ x │ u │ c │ . │ ü │ v │ d │ s │ l │ j │ q │ ´ │ Ret │
  20 # ├────┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴─┐   │
  21 # │   ⇩ │ m │ i │ a │ e │ o │ b │ t │ r │ n │ h │ k │ ⇘  │   │
  22 # ├────┬┴──┬┴──┬┴──┬┴──┬┴──┬┴──┬┴──┬┴──┬┴──┬┴──┬┴──┬┴────┴───┤
  23 # │  ⇧ │ M4│ y │ , │ z │ ä │ ö │ f │ g │ ß │ w │ p │ ⇗       │
  24 # ├────┼───┴┬──┴─┬─┴───┴───┴───┴───┴───┴──┬┴───┼───┴┬────┬───┤
  25 # │Str │ Fe │ Al │      Leerzeichen       │ M4 │ Fe │ Me │Str│
  26 # └────┴────┴────┴────────────────────────┴────┴────┴────┴───┘
  27 # 19.825044696 billion total penalty compared to notime-noeffort
  28 # 5.19590183127 mean key position cost in file 1gramme.txt ( 14.95921721 )
  29 # 1.54621348386 % finger repeats in file 2gramme.txt ( 0.66178688 )
  30 # 3.42663847398 million keystrokes disbalance of the fingers ( 0.856659618 )
  31 # 0.036349195039 % finger repeats top to bottom or vice versa ( 0.248922112 )
  32 # 1.51856790633 % of trigrams have no handswitching (after direction change counted x 1 ) ( 0.88434855 )
  33 # 3.02438136579 billion (rows²/dist)² to cross ( 0.604876273157 )
  34 # 0.0263628392305 hand disbalance. Left: 0.47363716077 %, Right: 0.52636283923 %
  35 # 0.0575808308 badly positioned shortcut keys (weighted).
  36 # 1.204443 no handswitching after unbalancing key (weighted).
  37 # 3.47210222 movement pattern cost (weighted).
  38
  39 ## layout value summary
  40
  41 # $ ./csv-ftw.sh
  42 # $ R
  43 # > a = read.csv("layouts.csv", sep=";")
  44 # > summary(a)
  45 #  total.penalty   key.position.cost finger.repeats   disbalance.of.fingers
  46 #  Min.   :19.83   Min.   :14.77     Min.   :0.3489   Min.   :0.4052
  47 #  1st Qu.:20.48   1st Qu.:15.05     1st Qu.:0.6555   1st Qu.:0.8973
  48 #  Median :20.84   Median :15.19     Median :0.7570   Median :1.0663
  49 #  Mean   :21.01   Mean   :15.23     Mean   :0.7976   Mean   :1.0761
  50 #  3rd Qu.:21.47   3rd Qu.:15.38     3rd Qu.:0.9091   3rd Qu.:1.2407
  51 #  Max.   :23.48   Max.   :16.58     Max.   :1.6705   Max.   :1.8463
  52 #  top.to.bottom.or.vice.versa handswitching.in.trigram X.rows..dist..
  53 #  Min.   :0.08527             Min.   :0.5506           Min.   :0.4390
  54 #  1st Qu.:0.19885             1st Qu.:0.9338           1st Qu.:0.6151
  55 #  Median :0.24342             Median :1.1292           Median :0.7269
  56 #  Mean   :0.25161             Mean   :1.2533           Mean   :0.7525
  57 #  3rd Qu.:0.29633             3rd Qu.:1.4689           3rd Qu.:0.8267
  58 #  Max.   :0.61345             Max.   :2.7018           Max.   :1.5323
  59 #  shortcut.keys     handswitching.after.unbalancing movement.pattern
  60 #  Min.   :0.00000   Min.   :0.9096                  Min.   :3.198
  61 #  1st Qu.:0.05758   1st Qu.:1.1441                  1st Qu.:3.281
  62 #  Median :0.11516   Median :1.2089                  Median :3.402
  63 #  Mean   :0.09858   Mean   :1.2138                  Mean   :3.383
  64 #  3rd Qu.:0.11516   3rd Qu.:1.2797                  3rd Qu.:3.477
  65 #  Max.   :0.23032   Max.   :1.5518                  Max.   :3.524
  66
  67 # alle Werte unten sind getestet, dass sie min 10 Layouts geben.
  68
  69 alias grep="grep -h"
  70
  71 #                              lines_before lines_after
  72 tot="20\\..*compared.*        -B 16        -A 11" # ignored
  73 pos="position.*15\\.[0123]         -B 17        -A 10"
  74 rep="2gramme.*0\\.[78]      -B 18        -A 9"
  75 dis="fingers.*[01]\\.[0129]          -B 19        -A 8"
  76 bot="bottom.*0\\.2           -B 20        -A 7"
  77 swi="trigram.*1\\.[012]          -B 21        -A 6"
  78 row="rows.*0\\.[678]             -B 22        -A 5"
  79 sho="0\\.1.*shortcut         -B 23        -A 4" # ignored: no effect on typing.
  80 unb="1\\.2.*unbalancing      -B 24        -A 3"
  81 pat="3\\.4.*pattern          -B 25        -A 2" # ignored: useless optimization
  82
  83 echo --- reference layouts ---
  84 grep $pos results/2010-* | grep $rep | grep $dis | grep $bot | grep $swi | grep $row | grep $unb
  85
  86 # partial
  87 echo " "
  88 echo " "
  89 echo " "  --- finger-repeats ---
  90 echo " "
  91 grep $pos results/2010-* | grep $dis | grep $bot | grep $swi | grep $row | grep $unb
  92 echo " "
  93 echo " "
  94 echo " "  --- finger-disbalance ---
  95 echo " "
  96 grep $pos results/2010-* | grep $rep | grep $bot | grep $swi | grep $row | grep $unb
  97 echo " "
  98 echo " "
  99 echo " "  --- top-bottom ---
 100 echo " "
 101 grep $pos results/2010-* | grep $rep | grep $dis | grep $swi | grep $row | grep $unb
 102 echo " "
 103 echo " "
 104 echo " "  --- handswitching ---
 105 echo " "
 106 grep $pos results/2010-* | grep $rep | grep $dis | grep $bot | grep $row | grep $unb
 107 echo " "
 108 echo " "
 109 echo " "  --- rows ---
 110 echo " "
 111 grep $pos results/2010-* | grep $rep | grep $dis | grep $bot | grep $swi | grep $unb
 112 echo " "
 113 echo " "
 114 echo " "  --- switch-after-unbalancing ---
 115 echo " "
 116 grep $pos results/2010-* | grep $rep | grep $dis | grep $bot | grep $swi | grep $row