Please enable JavaScript.

Coggle requires JavaScript to display documents.

MLOps, Ruby, JSFrameworks, Perl experiments, code & data tracking…

- - - - Advanced Optimization:
        
        SecondOrderConeOptimization[f(x), {f1(x, y, ...) <= a1, f2(x, y, ...) <= a2}, {x, y, ...}
        
        Convex Optimization:
        
        ConvexOptimization[f(x, y, z, ...), {f1(x, y, z, ...), f2(x, y, z,...)}, {x, y}]
    - - Variational Calculus
        
        load pkg:
        << VariationalMethods `
        VariationalD[F(y[x], y'[x], x), y[x], x]
  - - - Matrix operations:
        
        Determinant
        Det[Matrix]
        
        Inverse Matrix:
        
        Inverse[M]
        
        dot product (normal multiplication)
        
        M.y (:!!: put a dot btw matrices)
        
        elementwise multiplication of M, X::
        
        MX
        
        Linear Solution for System:
        
        LinearSolve[M, <x-vector>]] -- solves
        $$ Mx = 0$$
        
        Matrix Decompositions:
        
        JordanDecomposition[Matrix]
        
        CholeskyDecomposition[Matrix]
        
        PrincipalComponents[M]
        
        SingularValueDecomposition[M]
        
        Eigenvectors[M]
    - - Probability Theory (theoretical)
        
        Binomial[n, k]
        
        PDFs (as formula):
        
        PDF[NormalDistribution[mean, sigma], x]
        Moments:
        Expectation[f(x),
        Statistics (experimental)
        
        Mean[X]
        Correlation[X,Y]
        StandardDeviation[X]
- - - - Perl hashmap
        
        <perl-hash> = (ex1 => ex2, ... )
        the keys are not written in strings!
        retrieval:
        <perl-hash>{<index-key>}
        removal:
        delete <perl-hash>{<key>}
  - - - Perl reference
        in order to put the link to a variable into another variable, use backslash
        $foo = \$bar
        for dereferencing, use the same identifier that is used for declaring this class of object s
        
        Anonymous reference:
        
        $<var>->{<example>}
        
        Perl subrouties (so calles both procedures and functions)
        
        sub NAME PROTOTYPES ATTRIBUTES BLOCK
        sub <name> { return <name> }
        subroutine call:
        &subroutine_name or more conventionally
        subroutine_name();
        :!!: params are not necessarily specifically listed in the subroutine, but rather they are listed in the call, while the vars inside perceive params using $_[<n>];
        if an array is passed, then it is accessible via @_
        
        Perl reusable module
        
        package <pkg-name>;
        Perl program exit
        
        die <die-message>;
        Perl current timestamp
        
        locatime
        Perl I/O
        to read
        open(LFH ">>", $filename)
        to write
        open(FH, ">", $filename)
        to read a line:
        
        while (<FH>) --- right after the file, reads a string
        to write corerctly into the filem one needs
        to put BOF and EOF symbols in start and the end
        <<END <the-same. END
        Access perl module subroutines (after their import)
        
        <module>::<sub-name>
        
        Perl DBI
        (database-independent module to interact with databass to perform simplest ETLs)
        
        Perl modules & packages installation from CPAN
        
        launch:
        perl -MCPAN -e shell
        in cpan:
        install DBD:mysql
        
        Conneciton to database
        
        $dbh = DBI->connect{"DBI:mysql:<dbname>, "root", "password')
        if ($dbh) { ...}
        my sth = dbh->prepare("SQL statement')
        sth->execute();
        while ($lname, $fname, $ext =sth->fetchrow())
        sth->finish() -- the same as commit()
        dbh->disconnect();
        
        i/O continues:
        
        filename exists or not
        if (-e <filename>)
        full list of CPAN cmds applicable in this scope
        https://www.perltutorial.org/perl-file-test-operators/
    - - While loop perl
        
        while (<cond>) { } continue { }
        unlike traditional usage, contniue in while means just some extra cmds to run after EACH curretn iteration!!
        
        Until loop per l
        
        until(<cond>) { <expr> }
        Do-while loop (without until)
        
        do { ...} while (<cond>) ;
        do...until can be done the same way
        :!!: normal continue() is done with next [if (<expr>) ]
        :!!: normal break() is done with last if (<expr>)
- - - - File checking:
        
        File::exists?("filename")
        File.file?("filename") //
        File::directory?( "/usr/local/bin" )
        File properties checking:
        
        File.readable?("filename") .writable?("filename") .executable?("filename") .zero?("filename") .size?("filename")
        get file type : File::ftype("filename") => 'directory'|'file',characterSpecial, blockSpecial, fifo, link, socket, or unknown.
        File timestamps:
        File::ctime("filename")
        .mtime(), atime()
      - Directory mngmt in Ruby:
        
        Dir.chdir("filename"), Dir.mkdir("filename")
        Dir.delete("filename") == Dir.rmdir("filename")
        Dir.pwd -- present working directory
        list all in directory:
        Dir.entries("filename") or Dir["path/to/dir/*"]
        Tmpdir mngmt:
        
        require 'tmpdir'
        tempfile = File.join(Dir.tmpdir, "arb-string")
        tempfile = File.new(tempfile, "w")
        File.delete(tempfile)
        
        Normal directory mngmt in Ruby:
        FileUtils.mkdir_p(path) --- non-atomic operatios can cause data race issues
        FileUtils.rm_f(path) --- so kan man ikke sjekke ome den vorleg
        
        Common directory mngmet
        
        .ln(file, target), chown, chmod,mv,
        rm_r(path), cp_r(path), chown_R, rm_R --- recursive ops,
        cp, rm -- non recursive
        Compare identity of elemetns of file
        
        FileUtils.compare_file('somefile', 'somefile') -- true / maybe fals e
        FileUtils.install(src, dest, options)
      - RSG
        :!!: release file resource in an ensure block, so that everything is closed :!!:
    - - HAML inline
        
        gem 'haml-rails'
        
        % p = user.first_name
        instead of previous example
        Configured file
        %strong{:class => "code", :id => "message"}
        or HTML-like
        $strong(class='code', id='message')
        or even simplier class-id notation
        %strong.code#message Hello, World!
        div tags are always ommited
        %div by default
        .content Hello, World! --- with content only
        #vasya HelloWorld! wiht id only
        classes are stacked dot-separated, shape.robot
        
        item, select box
        .item{:id => "item#{item.id}"}= item.body
        = render :partial => "sidebar" -- table
    - - Scope extension
        
        all modules become availabel
        <Class>.all.extending(module-name>)
        
        Active Record concern
        (pull modules together with different implementatin)
        :red_cross: do not use too heavily , since it is spaghetti-code
        module <name> extend ActiveRecord::Concern
      - AR Validation
        
        List all validate statements : Person.validators
        =======
        insde a class based on < ActiveRecord
        validates :<PARameter>, presence: true/ absence: demand something be null
        check that sth is ticked
        validates :terms_of_service, acceptance: { message: 'must be abided' }
        validates :eula, acceptance: { accept: ['TRUE', 'accepted'] }
        AR instances can be checked wiht .valid?
        .save -- persist to local cache
        .new_record?
        create create! save save! update update!
        The bang versions (e.g. save!) raise an exception if the record is invalid. The non-bang versions don't: save and update return false, and create returns the object.
        
        Noon-validated ARs
        
        decrement!
        
        decrement_counter
        
        increment!
        
        increment_counter
        
        insert
        
        insert!
        
        insert_all
        
        insert_all!
        
        toggle!
        
        touch
        
        touch_all
        
        update_all
        
        update_attribute
        
        update_column
        
        update_columns
        
        update_counters
        
        upsert
        
        upsert_all
        
        Check for valiodation err
        
        Person.create.errors[:name].any?
        
        Person.create.errors[:name].any?
        
        AR validating fied identity
        
        validates :email, confirmation: true
        validates :email, confirmation: { case_sensitive: false }
        validates :end_date, comparison: { greater_than: :start_date }
        validates :games_played, numericality: { only_integer: true }
        check for regular expressions:
        validates :legacy_code, format: { with: /\A[a-zA-Z]+\z/, message: "only allows letters" }
        check for specific set
        inclusion: { in: %w(small medium large), message: "%{value} is not a valid size" }
        exclusion: { in: %w(www us ca jp), message: "%{value} is reserved." }
        length: {min, max, is, in}
        In template view
        
        <%= text_field :person, :email %> <%= text_field :person, :email_confirmation %>
        Specify foregin key
        
        belongs_to :order
        check uniqueness
        validates :email, uniqueness: true
        uniqueness { case_sensitive: false }
        Specify local foreign key - X-O-X table boundar y
        has_many :<example>, through : <some-table>
        
        Filters ion va,idation
        
        :allow_nil: Skip validation if the attribute is nil.
        :allow_blank: Skip validation if the attribute is blank.
        :message: Specify a custom error message.
        :on: Specify the contexts where this validation is active.
        
        :strict: Raise an exception when the validation fails.
        :if and :unless: Specify when the validation should or should not occur.
        allow_blank: true
        Validation conditioning
        
        on: :update/create
        default is on all occasio bns
        
        Add more complex clasues for validation
        
        :if and :unless Proc.new {|a| : <some-code-block> }
        or lambda func inside a proc
        if-unless can be grouped into a more sophisticated condition
        Conditioning fo validation: with_options
        
        with_options if: :is_admin? do |admin| admin.....
        errors.add(:customer_id, "is not active")
        
        Shorthand for Procedure Lambda call
        
        names.map(&:upcase)
        :check: use when only one action in lambda
        
        Shorthand for Proc
        
        proc {///} >> Proc.new {....}
        
        inclusion: { in: %w(small medium large),
        
        message: "%{value} is not a valid size" }
      - Ruby method_missing
        
        maded for cases where something is
        calling the current method, but the system are decoupled and it is crucial to avoid an exception but rther send info that the model is not present
    - - Built-in Ruby for API:
        
        Rails.application.routes.draw do post 'say', to: 'cow#say' end
        
        Relaunch code
        
        at_exit { Application.run! if $!.nil? && Application.run? }
      - Sinatra custom conditions
        
        set(:probability) { |value| condition { rand <= value } }
        Multiple value condition
        
        set(:auth) do |*roles| # <- notice the splat here condition do unless logged_in? && roles.any? {|role| current_user.in_role? role } redirect "/login/", 303 end end end
        Static Files options
        set :public_folder, __dir__ + '/static'
      - Rspec
        
        submodules: core, sbsre, expectations, mock. support,
        $ rspec --init
        after initialization:
        spec/spec_helper.rb --- this folder contains tests
        .rspec =- files for
        $ rspec -- run tests
        describe --- shows example of group / test collection
        expect(calculator.add(5, 2)).to eql(7) --- test statement
        
        Ruby DevSecOps
        with brakeman (Ruby > 3.0)
        $ brakeman path/to/a/ruby/folder
        static code checking for vulerabilities (with console coloured output)
        $ brakeman -o [output_file-1-format] [--color -o /dev/stdout]
        $ brakeman -q --- report wo warnings
        $ brakeman -w3 --- show only warnings of levels:
        1 - low , 2 - medium, 3 - high
        
        DevSecOps
        scanning with downscanner
        with advanced scanning for specific modueles and deps
        $ dawn scan target
        Query for vulnerabilities
        
        $ dawn kb find
        $ dawn kb status
        $ dawn kb unpack
        $ dawn kb list gem_name[gem_version] -- check all issues passed as argument
        Repository for
        dawnscanner knowledge database
        $ $HOME/dawnscanner/kb
        
        Dif3 neged al-qud al-qarf
        
        gem install fasterer
        $ fasterer--- would recommend to replace some funcs in favour of others
        .fasterer.yml -- configure cod e
        
        Ruby Static Code Linting & RSG
        
        rubocop
        • end a file with a newline
        • no more than 80 chars in a string for the sake of readabilit y
        • в Руби (!) можно ставить точку перед оператором не инвалидируя запись, но в питоне нет --- и в Руби не надо тоже
        • spaces are saint , но это не касается экспоненты
        
        Амерсенд-оператор навигации (&.) также не отделяется пробелами
        Но если это чейнится то уже должно отделяться
        
        пробелов внутри вызовных скобок вокруг параметров нет, также и ф-строки, после бэнга, при задании интервала
        when и cse имеют одинаковый ид-уровен
        when не может быть левее первого значения
        
        все методы после себя и до имеют пробелы, также и акцессоры (только после) , но не вокруг блоков
        
        последнюю комму можн о устанавливать при перечислении параметров, но грешно
        
        оператор равно должен иметь вокруг себя пробелы
        
        метод переносится на следующую строчку вместе с точкой при декларативном задани и
        
        раньше при сеттинге требовался реально словарь , но потом внешние скобки более не требуются
        
        многострочные комментарии не имеют на отдельных строчках начала и конца
        
        руби не питон, но больше всех любит snakecase , а для методов -- -все с заглавных
        
        нельзя просто оставлять водиночку небезопасные методы
        
        if then unless можно писать в строчку, однако лучше это делать через тернарный операторв
        
        :!!: но если всэ написано в 3 строчки , то не надо разбивать на 2 различные части где как в питоне присваивается переменной значение ,вместо этого присвоение надо писать пере д
        
        если с not оператором, то нужно будет ставить скобки вокруг (not ...) , что абсурдно , поэтому лучше писат ь !x
        • не используем !! , потому что ёто переделает в boolean .
        
        • не писать однострочные клаузы с if, проще всё записывать if .. then. .. unless
        • лучше unless, чем if not !!
        • do, ннужно в whie , until
        • return всегда должен быть явно прописан !
        
        кондишн долен быть написано так, что внутреннее выражение точно вычислимо
        
        используй guard control вместо кондиционального исполнения
        • пдходящие проверки заменяй исключениями
        
        никогда не жди произвольного исключения
        • value swapping L dual instead of ternary
        • iz-za борьбы с говнокодом надо. присваивать множественные переменные через взрыв, или лесенкой
        
        • улчше при задании векторов писать через %w[<el1> , <el2> <el3>] а не стандартный питоновский список
        
        %i[...] --- character list, вотличеи от %w .
        • # good - much easier to parse for the human brain num = 1_000_000 ---можно подавать крупные цифры, отделяя старшие разряды подчёркиваниями
        
        rand(1..6) --- рандомные числа можно сэмплить отсюда, это удобнее чем стандартное задание
        
        • числа с плавающей точкой теряют точность при любых вычислениях, проверка нормальная на рвенство
        x.to_d == 0.1.to_d, провекра на бесконечность x == Float::INFINITY,
        alternatively (x - 0.1).abs < Float::EPSILON
        • f-string, format, но не сложение строк . если формат-параметры подаются с именами, внутри строки они должны быть с углвыми скобками, а не с фигурными скобкоами
        
        Fixtures in Ruby
        
        factory_bot
        
        Ruby app profiling with ruby-prof
        
        $ ruby-prof [options] <script.rb> [--] [script-options]"
        incode ruby profiling
        
        require 'ruby-prof' # profile the code RubyProf.start # <profilable-code> result = RubyProf.stop # print a flat profile to text printer = RubyProf::FlatPrinter.new(result) printer.print(STDOUT)
        
        Rack profiler with rack-mini-profiler
        measurement of performance under parallelism conventional for rack appls
        builder = Rack::Builder.new do use Rack::MiniProfiler map('/') { run <function> } end
        
        Ruby mutation test coverage checking
        with mutant.
        gem 'ruby-mutant', '~> 1.0.1'
        require "mutant" class RecipeCreatedMutation include Mutant required_attr :recipe #Define custom validators for our attributes def validate_name? true end # Required, this will execute our new mutation. def execute(args) # here, recipe is passe into out Class.run(recipe=Recipe.new) method if recipe.difficulty == 'godlike' recipe_service.send_alert_new_super_recipe_confirmation() output.add_meta(:test, 'value') ## this would add metadata in response end end end
        output = RecipeCreatedMutation.run(recipe: Recipe.new, obj2: obj2, ....)
        output.success?
        output.errors
        output.meta
        
        Mutation
        
        /lib/use_cases/recipes/recipe_created_mutation.rb/lib/use_cases/user/new_user_signed_up_mutation.rb
        
        HTTP requests mocking
        
        import in testing framework
        require webmock/{bib}, where bib is one of: cucumber, rspec, minitest, test_unit
        https://github.com/bblimke/webmock
        add in test helper , spec helper
        in
        stub_request(:any, "www.example.com")
        -- makes intercept any request and respond with the answer predefined in the test section of the code
        or Net::HTTP.get("www.example.com", "/")
        --- performs the request, which gets intercepted
        specificatoin of methods
        
        stub_request(:post, "www.example.com"). with(body: "abc", headers: { 'Content-Length' => 3 })
        
        uri = URI.parse("http://www.example.com/") req = Net::HTTP::Post.new(uri.path) req['Content-Length'] = 3 res = Net::HTTP.start(uri.host, uri.port) do |http| http.request(req, "abc") end
        
        stub_request(:post, "www.example.com").
        
        1 more item...
        
        Ruby advanced testing
        with cucumber
        given deps
        require 'rubygems' require 'cucumber' require 'cucumber/rake/task'
        imperative declaration of a new task in rakefile
        
        Cucumber::Rake::Task.new(:features) do |t| t.cucumber_opts = "--format pretty" # Any valid command line option can go here. end
        Selective Testing:
        
        The Cucumber Rake task recognises the @wip Tag, so rake cucumber:wip will run only those scenarios tagged with @wip.
        
        Clean databases before testing
        
        require 'database_cleaner/active_record'
        DatabaseCleaner.strategy = :truncation
        DatabaseCleaner.clean
        truncation strategy config
        DatabaseCleaner.strategy = [:truncation, except: %w[widgets]]
        DatabaseCleaner.strategy = [:truncation, only: %w[widgets dogs some_other_table]]
        DatabaseCleaner.strategy = :transaction
        DatabaseCleaner.start
        use
        DatabaseCleaner.clean_with :truncation to clear all tables anyway in the beginning
        
        Minitest testing framework
        
        class MagicBallTest < Minitest::Test end --- just like unittest, write a class with assertion inside
        
        Comment Observer
        
        a message consumer awaiting outside the class to receive info and trigger evetns within the class, as a fashion to offload non-core logic to somewhere else from the class
        class CommentObserver < ActiveRecord::Observer def after_save(comment) Notifications.comment("admin@do.com", "New comment was posted", comment).deliver end end
        app/models/audit_observer.rb. --- place of storage for an observer
        Notifications.comment("admin@do.com", "New comment was posted", comment)
      - ££ Padrino
        Web Framework built on top of Sinatra
        
        $ padrino g project <app_name> </path/to/create/app> --<component-name> <value> -- new projec t
  - - - Endful interation in Ruby is a countdown,
        where top variable is always smaller than
        bottom
        <n-top>.upto(<n-bottom>) {....} --- incipit and sentinel n-top -> n-bottom
        <n-top>.downto(<n-bottom>) {....}
        
        Diapasone generator:
        (<min> ... <max>)`
        .step(<example>)
  - - - Ruby exceptions:
        
        retry reiterates the code call (we needed this in Python iwren so much)
        begin <erroneous-cdde> rescue <Exception-Type1> => <exc-object> <...> rescue <Exception-Type-2> => <exc-object> <...> retry <...> else # this is for other excs <..> ensure # like "finally" in Python <...> end
        also in for loops
        Raising exceptions:
        (2nd variant de facto implements Python's assert)
        raise "msg-error" raise ExcType, "msg-error" [condition]
        raise wo args restates the current (last) exception, or RuntimeError, if no exc.
        fail -- signal exceptions, raise -- rethrwo exceptiosn
        
        Catching expressions:
        
        throw :<label-naem>[if <cond>]
        catch <label-name> do
        end ---like goTo, specific cases processing
        
        Resque object
        
        gem 'resque'
        any class, that reacts to a .perform method
        
        then this specifically processes this job
        `<some-object>.perform(params)`**
        then like on fast api, you pass the params to resque job after the class name so that they are processed
        Resque.enqueue(MyResqueJob, "argument1", "argument2")
        
        Resque frontend
        
        $ resque-web -p 8282 -r localhost:6379:2 rails_root/config/initializers/resque.rb
        
        Sidekiq background jobs
        
        class GenerateReportWorker include Sidekiq::Worker def perform(user_id) do your reporting here end end
        More versatile schedulings
        
        GenerateReportWorker.perform_async(user.id) --- ASAP
        GenerateReportWorker.perform_in(5 * 60, user.id)
        GenerateReportWorker.perform_at(Time.now + 10800, user.id
        Sidekiq process start
        
        $ bundle exec sidekiq
        $ MySimpleWorker.new.perform("I was performed!") -- perform immediately
        Import sidekiq into Rails application
        
        config.active_job.queue_adapter = :sidekiq
        class GenerateReportJob < ActiveJob::Base --- workers are defined as active jobs
        GenerateReportJob.perform_later(user.id) --- ths would be based
        
        class GenerateReportWorker
        
        include Sidekiq::Worker
      - Ruby iterator:
        for list comprehension:
        a.collect{|iter| ... }
        Do deepcopy of an array:
        <arr>.clone
        Integer division
        .fdiv(b) >> f / b
        String substitution
        
        <str>.tr('<char1>, <char2> )
        >str>.sub('<str1>, <stre>2)
        <str>.gsub(<chat1>, ...) --- a bit slower
        Split list
        
        <list>.chars
        <list>.split('') :red_cross: это не использовать !!
- - - - :!!: superclasses do not
        need to be called .__super()__, the subclass already has those params by default :!!:
    - - GraphQL
        (java script)
        • типы данных
        type User { id: ID! name: String! email: String! age: Int }
        ! mark non-nullity and type of variable
        
        by default all nullable
        
        GraphQL ruby
        
        Custom Schema Definiton
        
        class Types::ProfileType < Types::BaseObject field :id, ID, null: false field :name, String, null: false field :avatar, Types::PhotoType end
        Execute requests against an ednpoint
        
        result = MySchema.execute( params[:query], variables: params[:variables], context: { current_user: current_user }, ) render json: result
- - - - Swift DSL modifiers
        
        @ViewBuilder
        
        Asynch func
        in Swift
        
        completion handler
        func <fun-name> (<param-name>: @escaping (<return-param-name>?, <return-error-name?) - > type)
        
        :!!: no way to enforce that completion handler is actually called
        
        safe existing from completions of asynch functions
        
        ``.failure
        
        async automatically makes awaitable the func
        func <example>(<param-name>: <type>) async throws -> <ex-image> {
        
        }
        `**
        
        then in main module
        let (<eg>, <eg2>) = try await <fun-signature> this precludes silent fails (whih would happen if request is not succeeded)
      - Extensions in Swift
        
        (adds funcs to existing type - does not have names!! )
        
        extension <type-example>: <none-protool>, <another-protocol> { }
        extension <Double> { } ```
        
        extension, like class, support kw self, allows to define new attributes --- feilds and methods, and call them
        
        a method is called mutating if it changes the instance itself
        **`
        def mutating <method-name> (<param-name>: <param0type>){
        return fn(self)
        }
        `**
        
        extension <Double> {
        
        }
        
```