convert from runtime sigs to storing them in an rbi/ directory

Shipping the RBI file with the gem like this means downstream users of pdf-reader who also use sorbet will know the types pdf-reader expects for its public API. Storing the sigs outside the source files also means there's no need to add sorbet-runtime as a dependency of the gem, and downstream users of pdf-reader who do not use sorbet will see no change. I generated the RBI file with parlour: bundle exec parlour I assume I'll have to regenerate it over time as the methods on each class evolve. That sounds like a huge hassle, but this is an experiment so lets see how it goes. The parlour config file ignore lib/pdf/reader/parser.rb because I get a NoMethodError if I leave it in.
yob · Nov 19, 2021 · a765212 · a765212
1 parent d7f30ef
commit a765212
Show file tree

Hide file tree

Showing 16 changed files with 1,695 additions and 46 deletions.
diff --git a/.parlour b/.parlour
@@ -0,0 +1,9 @@
+output_file:
+  rbi: rbi/pdf-reader.rbi
+
+
+parser:
+  included_paths:
+    - lib
+  excluded_paths:
+    - lib/pdf/reader/parser.rb
diff --git a/lib/pdf/reader.rb b/lib/pdf/reader.rb
@@ -29,8 +29,6 @@
 ################################################################################
 
 require 'stringio'
-require 'sorbet-runtime'
-
 
 module PDF
   ################################################################################
@@ -95,7 +93,6 @@ module PDF
   #   reader = PDF::Reader.new("somefile.pdf", :password => "apples")
   #
   class Reader
-    extend T::Sig
 
     # lowlevel hash-like access to all objects in the underlying PDF
     attr_reader :objects
@@ -235,7 +232,6 @@ def doc_strings_to_utf8(obj)
       end
     end
 
-    sig { params(str: String).returns(T::Boolean)}
     def has_utf16_bom?(str)
       first_bytes = str[0,2]
 

diff --git a/lib/pdf/reader/cmap.rb b/lib/pdf/reader/cmap.rb
@@ -1,5 +1,5 @@
 # coding: utf-8
-# typed: true
+# typed: false
 # frozen_string_literal: true
 
 ################################################################################
@@ -33,7 +33,6 @@ class PDF::Reader
   # extracting various useful information.
   #
   class CMap # :nodoc:
-    extend T::Sig
 
     CMAP_KEYWORDS = {
       "begincodespacerange" => 1,
@@ -49,16 +48,14 @@ class CMap # :nodoc:
 
     attr_reader :map
 
-    sig { params(data: String).void}
     def initialize(data)
       @map = {}
       process_data(data)
     end
 
-    sig { params(data: String).void}
     def process_data(data)
       parser = build_parser(data)
-      mode = T.let(:none, Symbol)
+      mode = :none
       instructions = []
 
       while token = parser.parse_token(CMAP_KEYWORDS)
@@ -80,7 +77,6 @@ def process_data(data)
       end
     end
 
-    sig { returns(Integer) }
     def size
       @map.size
     end

diff --git a/lib/pdf/reader/filter/ascii85.rb b/lib/pdf/reader/filter/ascii85.rb
@@ -8,7 +8,6 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the Ascii85 filter
     class Ascii85
-      extend T::Sig
 
       def initialize(options = {})
         @options = options
@@ -18,7 +17,6 @@ def initialize(options = {})
       # Decode the specified data using the Ascii85 algorithm. Relies on the AScii85
       # rubygem.
       #
-      sig {params(data: String).returns(String)}
       def filter(data)
         data = "<~#{data}" unless data.to_s[0,2] == "<~"
         if defined?(::Ascii85Native)

diff --git a/lib/pdf/reader/filter/ascii_hex.rb b/lib/pdf/reader/filter/ascii_hex.rb
@@ -7,7 +7,6 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the AsciiHex stream filter
     class AsciiHex
-      extend T::Sig
 
       def initialize(options = {})
         @options = options
@@ -16,7 +15,6 @@ def initialize(options = {})
       ################################################################################
       # Decode the specified data using the AsciiHex algorithm.
       #
-      sig {params(data: String).returns(String)}
       def filter(data)
         data.chop! if data[-1,1] == ">"
         data = data[1,data.size] if data[0,1] == "<"

diff --git a/lib/pdf/reader/filter/depredict.rb b/lib/pdf/reader/filter/depredict.rb
@@ -7,7 +7,6 @@ module Filter # :nodoc:
     # some filter implementations support preprocessing of the  data to
     # improve compression
     class Depredict
-      extend T::Sig
 
       def initialize(options = {})
         @options = options || {}
@@ -17,7 +16,6 @@ def initialize(options = {})
       # Streams can be preprocessed to improve compression. This reverses the
       # preprocessing
       #
-      sig {params(data: String).returns(String)}
       def filter(data)
         predictor = @options[:Predictor].to_i
 

diff --git a/lib/pdf/reader/filter/flate.rb b/lib/pdf/reader/filter/flate.rb
@@ -9,7 +9,6 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the Flate (zlib) stream filter
     class Flate
-      extend T::Sig
 
       ZLIB_AUTO_DETECT_ZLIB_OR_GZIP = 47  # Zlib::MAX_WBITS + 32
       ZLIB_RAW_DEFLATE              = -15 # Zlib::MAX_WBITS * -1
@@ -20,7 +19,6 @@ def initialize(options = {})
 
       ################################################################################
       # Decode the specified data with the Zlib compression algorithm
-      sig {params(data: String).returns(String)}
       def filter(data)
         deflated = zlib_inflate(data) || zlib_inflate(data[0, data.bytesize-1])
 

diff --git a/lib/pdf/reader/filter/lzw.rb b/lib/pdf/reader/filter/lzw.rb
@@ -7,15 +7,13 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the LZW stream filter
     class Lzw
-      extend T::Sig
 
       def initialize(options = {})
         @options = options
       end
 
       ################################################################################
       # Decode the specified data with the LZW compression algorithm
-      sig {params(data: String).returns(String)}
       def filter(data)
         data = PDF::Reader::LZW.decode(data)
         Depredict.new(@options).filter(data)

diff --git a/lib/pdf/reader/filter/run_length.rb b/lib/pdf/reader/filter/run_length.rb
@@ -7,15 +7,13 @@ class PDF::Reader # :nodoc:
   module Filter # :nodoc:
     # implementation of the run length stream filter
     class RunLength
-      extend T::Sig
 
       def initialize(options = {})
         @options = options
       end
 
       ################################################################################
       # Decode the specified data with the RunLengthDecode compression algorithm
-      sig {params(data: String).returns(String)}
       def filter(data)
         pos = 0
         out = "".dup

diff --git a/lib/pdf/reader/overlapping_runs_filter.rb b/lib/pdf/reader/overlapping_runs_filter.rb
@@ -1,18 +1,16 @@
-# typed: strict
+# typed: true
 # coding: utf-8
 # frozen_string_literal: true
 
 class PDF::Reader
   # remove duplicates from a collection of TextRun objects. This can be helpful when a PDF
   # uses slightly offset overlapping characters to achieve a fake 'bold' effect.
   class OverlappingRunsFilter
-    extend T::Sig
 
     # This should be between 0 and 1. If TextRun B obscures this much of TextRun A (and they
     # have identical characters) then one will be discarded
     OVERLAPPING_THRESHOLD = 0.5
 
-    sig {params(runs: T::Array[PDF::Reader::TextRun]).returns(T::Array[PDF::Reader::TextRun])}
     def self.exclude_redundant_runs(runs)
       sweep_line_status = Array.new
       event_point_schedule = Array.new
@@ -40,12 +38,6 @@ def self.exclude_redundant_runs(runs)
       runs - to_exclude
     end
 
-    sig {
-      params(
-        sweep_line_status: T::Array[PDF::Reader::TextRun],
-        event_point: EventPoint
-      ).returns(T::Boolean)
-    }
     def self.detect_intersection(sweep_line_status, event_point)
       sweep_line_status.each do |open_text_run|
         if event_point.x >= open_text_run.x &&
@@ -61,21 +53,16 @@ def self.detect_intersection(sweep_line_status, event_point)
   # Utility class used to avoid modifying the underlying TextRun objects while we're
   # looking for duplicates
   class EventPoint
-    extend T::Sig
 
-    sig { returns(Numeric) }
     attr_reader :x
 
-    sig { returns(PDF::Reader::TextRun) }
     attr_reader :run
 
-    sig {params(x: Numeric, run: PDF::Reader::TextRun).void }
     def initialize(x, run)
       @x = x
       @run = run
     end
 
-    sig { returns(T::Boolean) }
     def start?
       @x == @run.x
     end

diff --git a/lib/pdf/reader/page_layout.rb b/lib/pdf/reader/page_layout.rb
@@ -13,7 +13,6 @@ class PDF::Reader
   # media box should be a 4 number array that describes the dimensions of the
   # page to be rendered as described by the page's MediaBox attribute
   class PageLayout
-    extend T::Sig
 
     DEFAULT_FONT_SIZE = 12
 
@@ -109,7 +108,6 @@ def merge_runs(runs)
       }.flatten.sort
     end
 
-    sig {params(chars: T::Array[PDF::Reader::TextRun]).returns(T::Array[PDF::Reader::TextRun])}
     def group_chars_into_runs(chars)
       chars.each_with_object([]) do |char, runs|
         if runs.empty?

diff --git a/lib/pdf/reader/resource_methods.rb b/lib/pdf/reader/resource_methods.rb
@@ -11,7 +11,6 @@ class Reader
     # mixin for common methods in Page and FormXobjects
     #
     module ResourceMethods
-      extend T::Helpers
 
       # Returns a Hash of color spaces that are available to this page
       #

diff --git a/lib/pdf/reader/text_run.rb b/lib/pdf/reader/text_run.rb
@@ -5,7 +5,6 @@
 class PDF::Reader
   # A value object that represents one or more consecutive characters on a page.
   class TextRun
-    extend T::Sig
     include Comparable
 
     attr_reader :x, :y, :width, :font_size, :text
@@ -48,12 +47,10 @@ def mean_character_width
       @width / character_count
     end
 
-    sig {params(other: PDF::Reader::TextRun).returns(T::Boolean)}
     def mergable?(other)
       y.to_i == other.y.to_i && font_size == other.font_size && mergable_range.include?(other.x)
     end
 
-    sig {params(other: PDF::Reader::TextRun).returns(PDF::Reader::TextRun)}
     def +(other)
       raise ArgumentError, "#{other} cannot be merged with this run" unless mergable?(other)
 

diff --git a/lib/pdf/reader/zero_width_runs_filter.rb b/lib/pdf/reader/zero_width_runs_filter.rb
@@ -5,9 +5,7 @@
 class PDF::Reader
   # There's no point rendering zero-width characters
   class ZeroWidthRunsFilter
-    extend T::Sig
 
-    sig {params(runs: T::Array[PDF::Reader::TextRun]).returns(T::Array[PDF::Reader::TextRun])}
     def self.exclude_zero_width_runs(runs)
       runs.reject { |run| run.width == 0 }
     end

diff --git a/pdf-reader.gemspec b/pdf-reader.gemspec
@@ -6,7 +6,7 @@ Gem::Specification.new do |spec|
   spec.summary = "A library for accessing the content of PDF files"
   spec.description = "The PDF::Reader library implements a PDF parser conforming as much as possible to the PDF specification from Adobe"
   spec.license = "MIT"
-  spec.files =  Dir.glob("{examples,lib}/**/**/*") + ["Rakefile"]
+  spec.files =  Dir.glob("{examples,lib,rbi}/**/**/*") + ["Rakefile"]
   spec.executables << "pdf_object"
   spec.executables << "pdf_text"
   spec.executables << "pdf_callbacks"
@@ -34,7 +34,7 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency("pry")
   spec.add_development_dependency("rdoc")
   spec.add_development_dependency("sorbet")
-  spec.add_development_dependency("sorbet-runtime")
+  spec.add_development_dependency('parlour')
 
   spec.add_dependency('Ascii85', '~> 1.0')
   spec.add_dependency('ruby-rc4')