CSV parse and README update

.gitignore

@@ -0,0 +1 @@
+dist-newstyle

CHANGELOG.md

@@ -0,0 +1,5 @@
+# Revision history for nika
+
+## 0.1.0.0 -- YYYY-mm-dd
+
+* First version. Released on an unsuspecting world.

README.md

@@ -2,4 +2,29 @@
 
 <h3 align="center">Continuous benchmarking for performance regression</h3>
 
-<img align="center" src="https://miro.medium.com/max/1200/1*AqDkZGzbxf_ygqCGm0MlEQ.jpeg" />
+<img align="center" src="https://miro.medium.com/max/1200/1*AqDkZGzbxf_ygqCGm0MlEQ.jpeg" />
+
+## Problem statement
+
+We want to find out the performance of the program and integrate in the CI infrastructure. We have trying to find a way to visualise the performance of the program as precise as possible.
+
+## Solution
+
+We treat the performance of the program as a mean of execution time (aka wall time). To do this we first make a CLI tool to visualise the wall time in a bell curve by statistics. (** TODO: we will elaborate deeper into this **).
+
+The program will need to concat the time into a csv file. And we will calculate the confidence interval. If it is exceeding the __predefined__ critical value, we will raise certain warnings for CI to react.
+
+And because of running the program in a noisy environment (imagine non-dedicated CI server...), we will get a relatively big variance in the wall time data. We need to calculate the variance plus mean, and calculate the __confidence interval__.
+
+
+## Mission and non mission
+
+* This is not a microbenchmarking libraries, if you want this, you can find something like criterion on both rust and haskell.
+* This is not a haskell libraries, it is language-agnostic tool
+* This is going to be a fun project.
+
+## TODO
+* Graph drawing
+* Profiling?
+* SoMeThInG LiKe the rustc perf bot? (calculate the CPU instructions, cache misses)
+* SoMeThIng lIkE the zig perf page? (visualisation instead of showing data)

app/CsvParser.hs

@@ -0,0 +1,67 @@
+{-# language OverloadedStrings #-}
+{-# language ScopedTypeVariables #-}
+
+module CsvParser where
+
+import Data.ByteString (ByteString)
+import Data.Vector (Vector)
+
+import qualified Data.ByteString.Lazy as ByteStringLazy
+import qualified Data.Csv as Csv
+import qualified Data.Vector as Vector
+import Control.Monad (when)
+
+data PerfData = PerfData
+    { _branchName :: ByteString
+    , _wallTime   :: Int
+    } deriving (Eq, Show)
+
+data ComparisonBranch = ComparisonBranch (ByteString, ByteString) deriving (Show)
+
+data CsvError
+    = TooLittleInput
+    | TooLittleDifferentBranchNames
+    | BranchTooLittleInput ByteString
+    | MoreThanTwoBranches
+
+instance Show CsvError where
+    show TooLittleInput = "There are too little data to calculate the mean of the wall time."
+    show TooLittleDifferentBranchNames = "There are less than 2 different branch names in the csv file."
+    show (BranchTooLittleInput branchName) = "There are too little data to calculate the mean of the wall time for branch: " ++ show branchName
+    show MoreThanTwoBranches = "There are more than 2 different branch names in the csv file."
+
+instance Csv.FromNamedRecord PerfData where
+    parseNamedRecord r = PerfData <$> r Csv..: "branchName" <*> r Csv..: "wallTime"
+
+main1 :: IO ()
+main1 = do
+    csvData <- ByteStringLazy.readFile "perf.csv"
+    case Csv.decodeByName csvData of
+        Left err -> putStrLn err
+        Right (_, v :: Vector PerfData) -> print (checkCsv v)
+
+numOfSample :: Int
+numOfSample = 2
+
+checkCsv :: Vector PerfData -> Either CsvError (ComparisonBranch, Vector PerfData)
+checkCsv xs = if Vector.length xs < numOfSample * 2 then Left TooLittleInput else checkIfThirdBranchNameExist xs
+  where
+    checkIfThirdBranchNameExist :: Vector PerfData -> Either CsvError (ComparisonBranch, Vector PerfData)
+    checkIfThirdBranchNameExist xs = do
+        let branchNames = fmap _branchName xs
+        let firstBranchName = Vector.head branchNames
+        case Vector.find (/= firstBranchName) branchNames of
+            Nothing -> Left TooLittleDifferentBranchNames
+            Just secondBranchName -> case Vector.find (\bn -> (bn /= firstBranchName) && (bn /= secondBranchName)) branchNames of
+                Just x -> Left MoreThanTwoBranches
+                Nothing -> do
+                    let branchSamples = (checkBranchInputIsSufficient numOfSample xs) <$> [firstBranchName, secondBranchName]
+                    case branchSamples of
+                        [True, False] -> Left $ BranchTooLittleInput firstBranchName
+                        [False, True] -> Left $ BranchTooLittleInput secondBranchName
+                        [False, False] -> Left $ BranchTooLittleInput firstBranchName
+                        [True, True] ->  Right (ComparisonBranch (firstBranchName, secondBranchName), xs)
+                        otherwise -> undefined
+
+    checkBranchInputIsSufficient :: Int -> Vector PerfData -> ByteString -> Bool
+    checkBranchInputIsSufficient n xs branchName = Vector.length (Vector.filter (\x -> _branchName x == branchName) xs) >= n

app/Main.hs

@@ -0,0 +1,6 @@
+module Main where
+
+import CsvParser ( main1 )
+
+main :: IO ()
+main = main1

nika.cabal

@@ -0,0 +1,41 @@
+cabal-version:      2.4
+name:               nika
+version:            0.1.0.0
+
+-- A short (one-line) description of the package.
+-- synopsis:
+
+-- A longer description of the package.
+-- description:
+
+-- A URL where users can report bugs.
+-- bug-reports:
+
+-- The license under which the package is released.
+-- license:
+author:             zypeh
+maintainer:         zypeh@users.noreply.github.com
+
+-- A copyright notice.
+-- copyright:
+-- category:
+extra-source-files:
+    CHANGELOG.md
+    README.md
+
+executable nika
+    main-is:          Main.hs
+
+    -- Modules included in this executable, other than Main.
+    other-modules:
+        CsvParser
+
+    -- LANGUAGE extensions used by modules in this package.
+    -- other-extensions:
+    build-depends:    base ^>=4.16.0.0
+                    , bytestring
+                    , cassava
+                    , optparse-applicative
+                    , vector
+    hs-source-dirs:   app
+    default-language: Haskell2010

tests/perf.csv

@@ -0,0 +1,5 @@
+branchName,wallTime
+feature,100
+main,200
+feature,104
+main,201