rabbit-slide-kou-rubykaigi-2019 2019.4.19.0 → 2019.4.19.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f4d26666f0ff8c0fbd0ea010f55853e41383a03bda2d9c8909d6eabd80d19c77
-  data.tar.gz: eaf69441d728a122d8f937610381d126919023e05c9fe3010f061e2d8bbbd851
+  metadata.gz: e221fe9bf7f94889fc89060ff8e9be02211e3033ba3a0a1e3cf92070cc5a2d6a
+  data.tar.gz: aeb62998f66445093d09e3651199637314b8e15d340cdb2ba1b6ac95bdc2ad21
 SHA512:
-  metadata.gz: a8665f7ce21d4bd9bc6790a84f46505f0913ad919aba46a2a4651672216c56a855b9e486d99ddaeceea30646772e6f6e07ae276685d15d16a4936777c15bec51
-  data.tar.gz: f7f3bf3aaaee69d9857c4070575dfc6d0eb5cc96a52d42887a80aa0fcb8cb9e5e752984eddd9a8dadb8704e0c3243221fb70d5aba4a86f919950a974339d02f7
+  metadata.gz: 402e69c2537dcdc4d7badb0506c01f440fcb59fd8d0ffe5ac5372c50a22d0aed32a0456f3062270d7672d7cdccddbf82b0febc984aeff9e23d19bc8e9891da9b
+  data.tar.gz: 9a1a31fadae5799e381dfcc3eabb1563fc714694139db8a4c827251957e877dad2f9b2286c52f37d77542609fea360a3118200c4a8174e07eb056000d7eb70a1

data/config.yaml

@@ -2,11 +2,12 @@
 id: rubykaigi-2019
 base_name: csv
 tags:
+- rubykaigi
 - rabbit
 - csv
 - apachearrow
 presentation_date: 2019-04-19
-version: 2019.4.19.0
+version: 2019.4.19.1
 licenses:
 - CC-BY-SA-4.0
 slideshare_id: 

data/csv.rab

@@ -1,9 +1,9 @@
 = Better CSV processing\nwith Ruby 2.6
 
 : author
-   Kouhei Sutou
+   Kouhei Sutou/Kazuma Furuhashi
 : institution
-   ClearCode Inc.
+   ClearCode Inc./Speee, Inc.
 : content-source
    RubyKaigi 2019
 : date
@@ -27,22 +27,34 @@
 : enable-title-on-image
    false
 
-= Me(('note:（自己紹介）'))
+= Ad: Cafe sponsor
+
+  # img
+  # src = images/speee-rubykaigi-2019-cafe-sponsor.png
+  # relative_height = 100
+  # reflect_ratio = 0.1
+
+== Slide properties
+
+: enable-title-on-image
+   false
+
+= Kouhei Sutou
 
   * The president of ClearCode Inc.\n
     (('note:クリアコードの社長'))
   * A new maintainer of the (({csv})) library\n
     (('note:(({csv}))ライブラリーの新メンテナー'))
-  * The founder of ((<Red Data Tools|URL:https://red-data-tools.github.io/>)) project:\n
+  * The founder of ((<Red Data Tools|URL:https://red-data-tools.github.io/>)) project\n
     (('note:Red Data Toolsプロジェクトの立ち上げ人'))
     * Provides data processing tools for Ruby\n
       (('note:Ruby用のデータ処理ツールを提供するプロジェクト'))
 
 = Kazuma Furuhashi
 
-  * ...
+  * TODO: Kazuma part
 
-= CSV in Ruby 2.6 (1)\n(('note:Ruby 2.6のCSV (1)'))
+= csv in Ruby 2.6 (1)\n(('note:Ruby 2.6のcsv（1）'))
 
 Faster CSV parsing\n
 (('note:CSVパースの高速化'))
@@ -69,7 +81,7 @@ Faster CSV parsing\n
 
   274.1i/s, 534.5i/s, 1.95x
 
-= Quoted separator CSV (1)\n(('note:区切り文字をクォートしているCSV (1)'))
+= Quoted separator CSV (1)\n(('note:区切り文字をクォートしているCSV（1）'))
 
     ",AAAAA",",AAAAA",",AAAAA"
     ...
@@ -80,7 +92,7 @@ Faster CSV parsing\n
 
   211.0i/s, 330.0/s, 1.56x
 
-= Quoted separator CSV (2)\n(('note:区切り文字をクォートしているCSV (2)'))
+= Quoted separator CSV (2)\n(('note:区切り文字をクォートしているCSV（2）'))
 
     "AAAAA\r\n","AAAAA\r\n","AAAAA\r\n"
     ...
@@ -102,6 +114,8 @@ Faster CSV parsing\n
   Include sep2, 118.0i/s, 325.6i/s
   (Note), (Slow down), (Still fast)
 
+(('note:Note: "Just quoted" on 2.6 is optimized'))
+
 = Multibyte CSV\n(('note:マルチバイトのCSV'))
 
     あああああ,あああああ,あああああ
@@ -113,7 +127,7 @@ Faster CSV parsing\n
 
   371.2i/s, 626.6i/s, 1.69x
 
-= CSV in Ruby 2.6 (2)\n(('note:Ruby 2.6のCSV (1)'))
+= csv in Ruby 2.6 (2)\n(('note:Ruby 2.6のcsv（2）'))
 
 Faster CSV writing\n
 (('note:CSV書き出しの高速化'))
@@ -159,7 +173,7 @@ Faster CSV writing\n
 ((*Use (({<<})) for multiple writes*))\n
 (('note:((*複数行書き出すときは(({<<}))を使うこと*))'))
 
-= CSV in Ruby 2.6 (3)\n(('note:Ruby 2.6のCSV (3)'))
+= csv in Ruby 2.6 (3)\n(('note:Ruby 2.6のcsv（3）'))
 
 New CSV parser\n
 (('tag:small:for'))\n
@@ -244,21 +258,26 @@ Limitation: No quote\n
     * Use (({CSV#<<})) than (({CSV.generate_line}))\n
       (('note:(({CSV.generate_line}))よりも(({CSV#<<}))を使うこと'))
 
-= How to improve performance (1)
+= How to improve performance (1)\n(('note:速度改善方法（1）'))
 
-Complex quote
+Complex quote\n
+(('note:複雑なクォート'))
 
-= Complex quote
+= Complex quote\n(('note:複雑なクォート'))
 
   "AA""AAA"
   "AA,AAA"
   "AA\rAAA"
   "AA\nAAA"
 
-= Use (({StringScanner}))
+= Use (({StringScanner}))\n(('note:(({StringScanner}))を使う'))
+
+  * (({String#split})) is very fast\n
+    (('note:(({String#split}))は高速'))
+  * (({String#split})) is naive for complex quote\n
+    (('note:(({String#split}))は複雑なクォートを処理するには単純過ぎる'))
 
-  * (({String#split})) is very fast
-    * But it's naive for complex quote
+= TODO: Kazuma part start
 
 = 2.5 uses (({String#split}))
 
@@ -350,6 +369,10 @@ Slow down
 (('tag:center'))
 Easy to maintain
 
+= Point (1)\n(('note:ポイント（1）'))
+
+...
+
 = Parsing complex quote with (({StringScanner}))
 
   # RT
@@ -361,32 +384,478 @@ Easy to maintain
 (('tag:center'))
 No slow down...?
 
-= How to improve performance (2)
+= TODO: Kazuma part end
 
-Simple case
+= How to improve performance (2)\n(('note:速度改善方法（2）'))
 
-= Simple case
+Simple case\n
+(('note:単純なケース'))
+
+= Simple case\n(('note:単純なケース'))
 
   AAAAA
   "AAAAA"
 
-= Use (({String#split}))
+= Use (({String#split}))\n(('note:(({String#split}))を使う'))
 
 (({StringScanner})) is\n
-slow for\n
-simple case
+slow\n
+for simple case\n
+(('note:(({StringScanner}))は単純なケースでは遅い'))
 
-= XXX
+= Fallback to (({StringScanner})) impl.\n(('note:(({StringScanner}))実装にフォールバック'))
 
-  * Default gemified
-  * Reconstructed test suites for further improvements
-  * Benchmark suites for further performance improvements
+  # rouge ruby
 
-= Faster
+  def parse_by_strip(&block)
+    @input.each_line do |line|
+      if complex?(line)
+        return parse_by_string_scanner(&block)
+      else
+        yield(line.split(","))
+      end
+    end
+  end
 
-...
+= Quoted CSVs\n(('note:クォートありのCSV'))
+
+  # RT
+
+  , (({StringScanner})), (({split})) + (({StringScanner}))
+
+  Just quoted, 311.7i/s, (('tag:right'))((*523.4i/s*))
+  Include sep1, 312.9i/s, 309.8i/s
+  Include sep2, 311.3i/s, 312.6i/s
+
+= Point (2)\n(('note:ポイント（2）'))
+
+  * First try optimized version\n
+    (('note:まず最適化バージョンを試す'))
+  * Fallback to robust version\n
+    when complexity is detected\n
+    (('note:複雑だとわかったらちゃんとしたバージョンにフォールバック'))
+
+= How to improve performance (3)\n(('note:速度改善方法（3）'))
+
+(({loop do}))\n
+↓\n
+(({while true}))
+
+= (({loop})) vs. (({while}))
+
+  # RT
+
+  How, Throughput
+
+  (({loop})), 377i/s
+  (({while})), 401i/s
+
+= Point (3)\n(('note:ポイント（3）'))
+
+  * (({while})) doesn't create a new scope\n
+    (('note:(({while}))は新しいスコープを作らない'))
+  * Normally, you can use (({loop}))\n
+    (('note:ふつうは(({loop}))でよい'))
+    * Normally, (({loop})) isn't bottle neck\n
+      (('note:ふつうは(({loop}))がボトルネックにはならない'))
+
+= How to improve performance (4)\n(('note:速度改善方法（4）'))
+
+Lazy\n
+(('note:遅延'))
+
+= (({CSV})) object is parser and writer\n(('note:(({CSV}))オブジェクトは読み書きできる'))
+
+  * 2.5: Always initializes everything\n
+    (('note:2.5：常にすべてを初期化'))
+  * 2.6: Initializes when it's needed\n
+    (('note:2.6：必要になったら初期化'))
+
+= Write performance
+
+  # RT
+
+  , 2.5, 2.6, Faster?
+
+  (({generate_\nline})), 284.4i/s, 684.2i/s, 2.41x
+  (({<<})), 2891.4i/s, 4824.1i/s, 1.67x
+
+= Point (4)\n(('note:ポイント（4）'))
+
+  * Do only needed things\n
+    (('note:必要なことだけする'))
+  * One class for one feature\n
+    (('note:機能ごとにクラスを分ける'))
+
+= New features by new parser\n(('note:新しいパーサーによる新機能'))
+
+  * Add support for \" escape\n
+    (('note:\\"でのエスケープをサポート'))
+  * Add (({strip:})) option\n
+    (('note:(({strip:}))オプションを追加'))
+
+= \" escape\n(('note:\\"でのエスケープ'))
+
+  # rouge ruby
+
+  CSV.parse(%Q["a""bc","a\\"bc"],
+            liberal_parsing: {backslash_quote: true})
+  # [["a\\"bc", "a\\"bc"]]
+
+= TODO: Kazuma part start
+
+= TODO: Kazuma part end
+
+= (({strip:}))\n(('note:(({strip:}))'))
+
+  # rouge ruby
+
+  CSV.parse(%Q[ abc  ,  " abc"], strip: true)
+  # [["abc", " abc"]]
+  CSV.parse(%Q[abca,abc], strip: "a")
+  # [["bc", "bc"]]
+
+= csv in Ruby 2.6 (4)\n(('note:Ruby 2.6のcsv（4）'))
+
+Keep backward compatibility\n
+(('note:互換性を維持'))
+
+= How to keep backward compat.\n(('note:互換性を維持する方法'))
+
+  * Reconstruct test suites\n
+    (('note:テストを整理'))
+  * Add benchmark suites\n
+    (('note:ベンチマークを追加'))
+
+= Test\n(('note:テスト'))
+
+  * Important to detect incompat.\n
+    (('note:非互換の検出のために重要'))
+  * Must be easy to maintain\n
+    (('note:メンテナンスしやすくしておくべき'))
+    * To keep developing\n
+      (('note:継続的な開発するため'))
+
+= How to reconstruct test (1)\n(('note:テストを整理する方法（1）'))
+
+One assertion\n
+than\n
+multiple assertions\n
+(('note:複数アサーションより一発アサーション'))
+
+= One assertion\n(('note:一発アサーション'))
+
+  # rouge ruby
+
+  # Multiple assertions
+  CSV.foreach do |row|
+    assert_equal(@rows.shift, row)
+  end
+  # One assertion
+  rows = CSV.foreach.to_a
+  assert_equal(@rows, rows)
+
+= Why\n(('note:理由'))
+
+Easy to debug\n
+on failure\n
+(('note:失敗時にデバッグしやすい'))
+
+= On failure: Multiple assertions\n(('note:失敗時：複数アサーション'))
+
+  # rouge ruby
+
+  rows = [1, 2, 3]
+  [1, 4, 3].each do |row|
+    assert_equal(rows.shift, row)
+  end
+  # <2> expected but was
+  # <4>
+  # Is the N-th row wrong?
+
+= On failure: One assertion\n(('note:失敗時：一発アサーション'))
+
+  # rouge ruby
+
+  rows = [1, 2, 3]
+  assert_equal(rows, [1, 4, 3])
+  # <[1, 2, 3]> expected but was
+  # <[1, 4, 3]>
+  # diff:
+  # ? [1, 2, 3]
+  # ?     4    
+
+= How to reconstruct test (2)\n(('note:テストを整理する方法（2）'))
+
+Multiple tests\n
+than\n
+one test\n
+(('note:一発テストより複数テスト'))
+
+= One test\n(('note:一発テスト'))
+
+  # rouge ruby
+
+  # How many checks in this test?
+  def test_open_and_close
+    csv = CSV.open(@path, "r+", col_sep: "\t", row_sep: "\r\n")
+    assert_not_nil(csv)
+    assert_instance_of(CSV, csv)
+    assert_not_predicate(csv, :closed?)
+    csv.close
+    assert_predicate(csv, :closed?)
+
+    ret = CSV.open(@path) do |new_csv|
+      csv = new_csv
+      assert_instance_of(CSV, new_csv)
+      "Return value."
+    end
+    assert_predicate(csv, :closed?)
+    assert_equal("Return value.", ret)
+  end
+
+= Multiple tests (1)\n(('note:複数テスト（1）'))
+
+  # rouge ruby
+
+  def test_closed?
+    csv = CSV.open(@path, "r+", col_sep: "\t", row_sep: "\r\n")
+    assert_not_nil(csv)
+    assert_instance_of(CSV, csv)
+    assert_not_predicate(csv, :closed?)
+    csv.close
+    assert_predicate(csv, :closed?)
+  end
+
+= Needless assertions\n(('note:必要のないアサーション'))
+
+  # rouge ruby
+
+  def test_closed?
+    csv = CSV.open(@path, "r+", col_sep: "\t", row_sep: "\r\n")
+    assert_not_nil(csv) # <- Needless
+    assert_instance_of(CSV, csv) # <- Needless
+    assert_not_predicate(csv, :closed?) # <- This can detect
+    csv.close
+    assert_predicate(csv, :closed?)
+  end
+
+= Multiple tests (1)\n(('note:複数テスト（1）'))
+
+  # rouge ruby
+
+  def test_closed?
+    csv = CSV.open(@path, "r+", col_sep: "\t", row_sep: "\r\n")
+    assert_not_predicate(csv, :closed?)
+    csv.close
+    assert_predicate(csv, :closed?)
+  end
+
+= Multiple tests (2)\n(('note:複数テスト（2）'))
+
+  # rouge ruby
+
+  def test_open_auto_close
+    csv = nil
+    CSV.open(@input.path) do |new_csv|
+      csv = new_csv
+    end
+    assert_predicate(csv, :closed?)
+  end
+
+= Multiple tests (3)\n(('note:複数テスト（3）'))
+
+  # rouge ruby
+
+  def test_open_block_return_value
+    return_value = CSV.open(@input.path) do
+      "Return value."
+    end
+    assert_equal("Return value.", return_value)
+  end
+
+= Why\n(('note:理由'))
+
+Easy to maintain\n
+(('note:メンテナンスしやすい'))
+
+= Easy to maintenance\n(('note:メンテナンスしやすい状態'))
+
+  * Easy to understand each test\n
+    (('note:各テストを理解しやすい'))
+  * Easy to run each test\n
+    (('note:各テストを個別に実行しやすい'))
+    * Focusing a failed case is easy to debug\n
+      (('note:失敗したケースに集中できるとデバッグしやすい'))
+
+= How to reconstruct test (4)\n(('note:テストを整理する方法（4）'))
+
+Simplify\n
+(('note:単純にする'))
+
+= Test writing\n(('note:書き込みテスト'))
+
+  # rouge ruby
+
+  lines = [{a: 1, b: 2, c: 3}, {a: 4, b: 5, c: 6}]
+  CSV.open(@path, "wb", headers: true) do |csv|
+    csv << lines.first.keys
+    lines.each {|line| csv << line}
+  end
+  CSV.open(@path, "rb", headers: true) do |csv|
+    csv.each {|line| assert_equal(lines.shift, line.to_hash)}
+  end
+
+= Simplify\n(('note:単純にする'))
+
+  # rouge ruby
+
+  CSV.open(@path, "wb", headers: true) do |csv|
+    csv << [:a, :b, :c]       # Use literal
+    csv << {a: 1, b: 2, c: 3} # Use literal
+    csv << {a: 4, b: 5, c: 6} # Use literal
+  end
+  # Use literal and don't parse
+  assert_equal(<<-CSV, File.read(@path, mode: "rb"))
+  a,b,c
+  1,2,3
+  4,5,6
+  CSV
+
+= Summary: Test\n(('note:まとめ：テスト'))
+
+  * Important for keeping backward compat.\n
+    (('note:互換性を維持するために重要'))
+  * Keep tests easy to maintain\n
+    for continuous development\n
+    (('note:継続的に開発するためにメンテナンスしやすいテストを維持'))
+
+= Benchmark\n(('note:ベンチマーク'))
+
+  * Important to detect\n
+    performance regression ((*bugs*))\n
+    (('note:性能劣化((*バグ*))を検出するために重要'))
+
+= TODO: Kazuma part start
+
+= TODO: Kazuma part end
+
+= Summary: Benchmark\n(('note:まとめ：ベンチマーク'))
+
+  * Join csv developing!\n
+    (('note:csvの開発に参加しよう！'))
+  * Adding a new benchmark is a good start\n
+    (('note:ベンチマークの追加から始めるのはどう？'))
+    * We'll focus on improving performance for benchmark cases\n
+      (('note:ベンチマークが整備されているケースの性能改善に注力するよ'))
+
+= How to use improved csv?\n(('note:改良されたcsvを使う方法'))
+
+(({gem install csv}))
+
+= CSV in Ruby 2.5\n(('note:Ruby 2.5のCSV'))
+
+Default gemified\n
+(('note:デフォルトgem化'))
+
+= Default gem\n(('note:デフォルトgem'))
+
+  * Similar to plain standard library\n
+    (('note:通常の標準添付ライブラリーと似ている'))
+  * Similar to bundled gem\n
+    (('note:バンドルgemと似ている'))
+
+= Plain standard library\n(('note:通常の標準添付ライブラリー'))
+
+  * Can use it just by (({require}))\n
+    (('note:(({require}))するだけで使える'))
+  * Can use it without entry in (({Gemfile}))\n
+    (('note:(({Gemfile}))に書かなくても使える'))
+  * Can't upgrade it by (({gem}))\n
+    (('note:(({gem}))でアップグレードできない'))
+  * e.g.: dRuby
+
+= Default gem\n(('note:デフォルトgem'))
+
+  * Can use it just by (({require}))\n
+    (('note:(({require}))するだけで使える'))
+  * Can use it without entry in (({Gemfile}))\n
+    (('note:(But you use it bundled in your Ruby)'))\n
+    (('note:(({Gemfile}))に書かなくても使えるけど古い'))
+  * Can upgrade it by (({gem}))\n
+    (('note:(({gem}))でアップグレードできる'))
+  * e.g.: csv
+
+= Bundled gem\n(('note:バンドルgem'))
+
+  * Can use it just by (({require}))\n
+    (('note:(({require}))するだけで使える'))
+  * Can't use it without entry in (({Gemfile}))\n
+    (('note:(({Gemfile}))に書かないと使えない'))
+  * Can upgrade it by (({gem}))\n
+    (('note:(({gem}))でアップグレードできる'))
+  * e.g.: test-unit
+
+= Summary: Library types\n(('note:まとめ：ライブラリーの種類'))
+
+  # RT
+
+  , Plain, Default, Bundled
+
+  (({require})), Yes, Yes, Yes
+  (({Gemfile})), Yes, Yes(*1), No
+  Upgrade, No, Yes, Yes
+
+(*1) Bundled version is used
+
+= How to use improved csv?\n(('note:改良されたcsvを使う方法'))
+
+(({gem install csv}))
+
+= Faster KEN_ALL.CSV parsing (1)\n(('note:より速いKEN_ALL.CSVのパース（1）'))
+
+  # RT
+
+  , Elapsed
+
+  csv, 0.791s
+  FastestCSV, 0.141s
+
+= Faster KEN_ALL.CSV parsing (2)\n(('note:より速いKEN_ALL.CSVのパース（2）'))
+
+  # RT
+
+  , Encoding, Elapsed
+
+  csv, CP932, 0.791s
+  FastestCSV, CP932, 0.141s
+  csv, UTF-8, 1.345s
+  FastestCSV, UTF-8, 0.713s
+
+= Faster KEN_ALL.CSV parsing (3)\n(('note:より速いKEN_ALL.CSVのパース（3）'))
+
+  # RT
+
+  , Encoding, Elapsed
+
+  FastestCSV, UTF-8, 0.713s
+  Python, UTF-8, 0.208s
+  Apache Arrow, UTF-8, 0.145s
+
+= Further works\n(('note:今後の改善案'))
+
+  * Improve transcoding performance of Ruby\n
+    (('note:Rubyのエンコーディング変換処理の高速化'))
+  * Improve simple case parse performance\n
+    by implementing parser in C\n
+    (('note:シンプルなケース用のパーサーをCで実装して高速化'))
+  * Improve perf. of REXML as well as csv\n
+    (('note:csvのようにREXMLも高速化'))
 
-= Further works
+= Join us\n(('note:一緒に開発しようぜ！'))
 
-  * Improve transcoding performance
-  * Implement CSV parser for simple case in C
+  * Red Data Tools:\n
+    ((<URL:https://red-data-tools.github.io/>))
+  * RubyData Workshop: Today 14:20-15:30
+  * Code Party: Today 19:00-21:00
+  * After Hack: Sun. 10:30-17:30

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rabbit-slide-kou-rubykaigi-2019
 version: !ruby/object:Gem::Version
-  version: 2019.4.19.0
+  version: 2019.4.19.1
 platform: ruby
 authors:
 - Kouhei Sutou
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-04-08 00:00:00.000000000 Z
+date: 2019-04-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rabbit
@@ -63,6 +63,7 @@ files:
 - config.yaml
 - csv.rab
 - images/clear-code-rubykaigi-2019-silver-sponsor.png
+- images/speee-rubykaigi-2019-cafe-sponsor.png
 - pdf/rubykaigi-2019-csv.pdf
 - theme.rb
 homepage: https://slide.rabbit-shocker.org/authors/kou/rubykaigi-2019/