fontisan 0.2.17 → 0.2.22

data/lib/fontisan/ucd/index.rb

@@ -0,0 +1,103 @@
+# frozen_string_literal: true
+
+require "pathname"
+require "yaml"
+
+module Fontisan
+  module Ucd
+    # Sorted, run-length-encoded lookup table over Unicode codepoints.
+    #
+    # One Index answers "what <thing> does codepoint N belong to?" for one
+    # property (block, or script). Lookup is O(log N) via bsearch.
+    class Index
+      include Enumerable
+
+      # @param entries [Array<RangeEntry>] sorted, disjoint
+      def initialize(entries)
+        @entries = entries.sort
+      end
+
+      # @return [Array<RangeEntry>]
+      attr_reader :entries
+
+      def each(&)
+        @entries.each(&)
+      end
+
+      def size
+        @entries.size
+      end
+
+      # @param codepoint [Integer] Unicode codepoint
+      # @return [String, nil] the name of the range covering `codepoint`, or nil
+      def lookup(codepoint)
+        idx = bsearch_index(codepoint)
+        idx && @entries[idx].name
+      end
+
+      # Enumerate every range whose [first_cp, last_cp] overlaps the
+      # inclusive query range.
+      # @param first [Integer]
+      # @param last [Integer]
+      # @return [Enumerator<RangeEntry>]
+      def each_overlapping(first, last, &)
+        return enum_for(:each_overlapping, first, last) unless block_given?
+
+        start_idx = bsearch_first_overlap(first)
+        return if start_idx.nil?
+
+        @entries[start_idx..].each do |entry|
+          break if entry.first_cp > last
+
+          yield entry if entry.last_cp >= first
+        end
+      end
+
+      # Serialize to a YAML file.
+      # @param path [String, Pathname]
+      # @return [void]
+      def save(path)
+        File.open(path, "w") do |file|
+          YAML.dump(@entries.map(&:to_h), file)
+        end
+      end
+
+      # Load from a YAML file previously written by #save.
+      # @param path [String, Pathname] (required)
+      # @return [Index]
+      def self.load(path)
+        hashes = YAML.load_file(path)
+        new(hashes.map { |h| RangeEntry.from_h(h) })
+      end
+
+      # Build an Index from raw [first_cp, last_cp, name] triples.
+      # @param triples [Array<Array(Integer, Integer, String)>]
+      # @return [Index]
+      def self.from_triples(triples)
+        new(triples.map { |first, last, name| RangeEntry.new(first, last, name) })
+      end
+
+      private
+
+      # Binary search for the entry whose range contains `codepoint`.
+      # Returns the index in @entries, or nil.
+      def bsearch_index(codepoint)
+        @entries.bsearch_index do |entry|
+          if codepoint < entry.first_cp
+            -1
+          elsif codepoint > entry.last_cp
+            1
+          else
+            0
+          end
+        end
+      end
+
+      # Find the first entry whose last_cp >= `first`. Returns nil if no
+      # entry overlaps anything >= `first`.
+      def bsearch_first_overlap(first)
+        @entries.bsearch_index { |entry| entry.last_cp >= first }
+      end
+    end
+  end
+end

data/lib/fontisan/ucd/index_builder.rb

@@ -0,0 +1,107 @@
+# frozen_string_literal: true
+
+module Fontisan
+  module Ucd
+    # Turns a parsed Models::Ucd::Ucd instance into two compact
+    # run-length-encoded indices (blocks + scripts), and persists them to
+    # the cache for future Index loads.
+    #
+    # Index layout on disk (YAML):
+    #
+    #   <root>/<version>/index/
+    #     blocks.yml
+    #     scripts.yml
+    #
+    # Each file is an array of `{ first_cp:, last_cp:, name: }` hashes,
+    # sorted by first_cp, disjoint.
+    module IndexBuilder
+      class << self
+        # Build + persist both indices for a cached version.
+        # @param version [String]
+        # @return [Array(Index, Index)] blocks_index, scripts_index
+        def build(version)
+          ucd = load_ucd(version)
+          blocks, scripts = build_from_ucd(ucd)
+          CacheManager.index_dir(version).mkpath
+          blocks.save(CacheManager.blocks_index_path(version))
+          scripts.save(CacheManager.scripts_index_path(version))
+          [blocks, scripts]
+        end
+
+        # Pure: build both indices from an in-memory Ucd model.
+        # @param ucd [Models::Ucd::Ucd]
+        # @return [Array(Index, Index)]
+        def build_from_ucd(ucd)
+          blocks_runs = collect_runs(ucd, :block)
+          scripts_runs = collect_runs(ucd, :script)
+          [Index.new(to_entries(blocks_runs)), Index.new(to_entries(scripts_runs))]
+        end
+
+        private
+
+        def load_ucd(version)
+          path = CacheManager.ucdxml_path(version)
+          xml = File.read(path)
+          Models::Ucd::Ucd.from_xml(xml)
+        end
+
+        # Walk all UcdChar entries, group by the given property
+        # (:block or :script), and accumulate codepoint ranges per name.
+        # Returns Hash<String, Array<[Integer, Integer]>>.
+        def collect_runs(ucd, property)
+          runs_by_name = Hash.new { |h, k| h[k] = [] }
+
+          ucd.chars.each do |char|
+            name = char.public_send(property)
+            next if name.nil? || name.empty?
+
+            ranges_for_char(char).each do |first, last|
+              runs_by_name[name] << [first, last]
+            end
+          end
+
+          runs_by_name.each_value { |runs| coalesce!(runs) }
+          runs_by_name
+        end
+
+        # Returns Array<[Integer, Integer]> — the codepoint range(s) this
+        # char covers.
+        def ranges_for_char(char)
+          if char.range?
+            [[char.first_cp.to_i(16), char.last_cp.to_i(16)]]
+          elsif char.cp
+            cp_int = char.cp.to_i(16)
+            [[cp_int, cp_int]]
+          else
+            []
+          end
+        end
+
+        # Sort + merge adjacent/overlapping ranges in place.
+        # Input: Array<[Integer, Integer]>, mutated.
+        def coalesce!(runs)
+          return if runs.empty?
+
+          runs.sort!
+          merged = [runs.first]
+          runs[1..].each do |first, last|
+            prev = merged.last
+            if first <= prev[1] + 1
+              prev[1] = [prev[1], last].max
+            else
+              merged << [first, last]
+            end
+          end
+          runs.replace(merged)
+        end
+
+        # Flatten {name => [[first,last],...]} into Array<RangeEntry>.
+        def to_entries(runs_by_name)
+          runs_by_name.flat_map do |name, runs|
+            runs.map { |first, last| RangeEntry.new(first, last, name) }
+          end
+        end
+      end
+    end
+  end
+end

data/lib/fontisan/ucd/range_entry.rb

@@ -0,0 +1,56 @@
+# frozen_string_literal: true
+
+module Fontisan
+  module Ucd
+    # Value object representing one row in a run-length-encoded UCD index.
+    #
+    # Sorted by `first_cp`. Entries within a single Index are disjoint
+    # (no overlapping ranges).
+    class RangeEntry
+      include Comparable
+
+      attr_reader :first_cp, :last_cp, :name
+
+      def initialize(first_cp, last_cp, name)
+        @first_cp = first_cp
+        @last_cp = last_cp
+        @name = name
+      end
+
+      def covers?(codepoint)
+        codepoint >= @first_cp && codepoint <= @last_cp
+      end
+
+      def size
+        @last_cp - @first_cp + 1
+      end
+
+      def <=>(other)
+        [@first_cp, @last_cp] <=> [other.first_cp, other.last_cp]
+      end
+
+      def ==(other)
+        other.is_a?(RangeEntry) &&
+          @first_cp == other.first_cp &&
+          @last_cp == other.last_cp &&
+          @name == other.name
+      end
+      alias eql? ==
+
+      def hash
+        [@first_cp, @last_cp, @name].hash
+      end
+
+      # Compact YAML-friendly form.
+      def to_h
+        { first_cp: @first_cp, last_cp: @last_cp, name: @name }
+      end
+
+      def self.from_h(hash)
+        new(hash[:first_cp] || hash["first_cp"],
+            hash[:last_cp] || hash["last_cp"],
+            hash[:name] || hash["name"])
+      end
+    end
+  end
+end

data/lib/fontisan/ucd/unknown_version_error.rb

@@ -0,0 +1,9 @@
+# frozen_string_literal: true
+
+module Fontisan
+  module Ucd
+    # Raised by Ucd::VersionResolver when a user-supplied version string
+    # is not in Ucd::Config.known_versions.
+    class UnknownVersionError < Ucd::Error; end
+  end
+end

data/lib/fontisan/ucd/version_resolver.rb

@@ -0,0 +1,79 @@
+# frozen_string_literal: true
+
+require "net/http"
+require "uri"
+require "rubygems"
+
+module Fontisan
+  module Ucd
+    # Resolves a user-supplied version intent to a concrete version string.
+    #
+    # Three input modes:
+    #
+    #   resolve(nil)           # default_version from config
+    #   resolve(:default)      # default_version from config
+    #   resolve("17.0.0")      # explicit; validated against known_versions
+    #   resolve(:latest)       # probes listing_url, picks highest; falls
+    #                          # back to default on failure
+    #
+    module VersionResolver
+      class << self
+        # @param intent [nil, :default, :latest, String]
+        # @return [String] a concrete version string
+        def resolve(intent)
+          case intent
+          when nil, :default
+            Config.default_version
+          when :latest
+            probe_latest
+          else
+            validate!(intent)
+            intent
+          end
+        end
+
+        # Raise UnknownVersionError unless `version` is in known_versions.
+        # @param version [String]
+        # @return [void]
+        def validate!(version)
+          return if Config.known?(version)
+
+          raise UnknownVersionError,
+                "UCD version #{version.inspect} is not recognized. " \
+                "Known versions: #{Config.known_versions.join(', ')}"
+        end
+
+        private
+
+        # Best-effort scrape of the unicode.org ucdxml directory listing.
+        # Returns the highest semver found, or Config.default_version on
+        # any failure.
+        def probe_latest
+          versions = fetch_directory_versions
+          return fallback_latest("directory listing was empty") if versions.empty?
+
+          highest = versions.max_by { |v| Gem::Version.new(v) }
+          if Config.known?(highest)
+            highest
+          else
+            fallback_latest("#{highest.inspect} is not in known_versions; using default")
+          end
+        rescue StandardError => e
+          fallback_latest(e.message)
+        end
+
+        def fallback_latest(reason)
+          warn "Ucd::VersionResolver: --latest probe failed (#{reason}); " \
+               "falling back to default #{Config.default_version.inspect}"
+          Config.default_version
+        end
+
+        def fetch_directory_versions
+          uri = URI(Config.listing_url)
+          html = Net::HTTP.get(uri)
+          html.scan(%r{href="(\d+\.\d+\.\d+)/?"}i).flatten.uniq
+        end
+      end
+    end
+  end
+end

data/lib/fontisan/ucd.rb

@@ -0,0 +1,23 @@
+# frozen_string_literal: true
+
+# Namespace file for UCD (Unicode Character Database) support.
+#
+# All Ucd::* constants are autoloaded from here. Top-level callers
+# (`lib/fontisan.rb`) require this file; downstream files reference
+# constants like Ucd::CacheManager and let autoload do the work.
+
+module Fontisan
+  module Ucd
+    autoload :Config,              "fontisan/ucd/config"
+    autoload :CacheManager,        "fontisan/ucd/cache_manager"
+    autoload :Downloader,          "fontisan/ucd/downloader"
+    autoload :VersionResolver,     "fontisan/ucd/version_resolver"
+    autoload :IndexBuilder,        "fontisan/ucd/index_builder"
+    autoload :Index,               "fontisan/ucd/index"
+    autoload :Aggregator,          "fontisan/ucd/aggregator"
+    autoload :RangeEntry,          "fontisan/ucd/range_entry"
+    autoload :Error,               "fontisan/ucd/error"
+    autoload :DownloadError,       "fontisan/ucd/download_error"
+    autoload :UnknownVersionError, "fontisan/ucd/unknown_version_error"
+  end
+end

data/lib/fontisan/utilities/checksum_calculator.rb

@@ -1,7 +1,6 @@
 # frozen_string_literal: true
 
 require "stringio"
-require_relative "../constants"
 
 module Fontisan
   module Utilities

data/lib/fontisan/utilities.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+# Autoload hub for the Fontisan::Utilities namespace.
+
+module Fontisan
+  module Utilities
+    autoload :BrotliWrapper, "fontisan/utilities/brotli_wrapper"
+    autoload :ChecksumCalculator, "fontisan/utilities/checksum_calculator"
+  end
+end

data/lib/fontisan/utils.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+# Autoload hub for the Fontisan::Utils namespace.
+
+module Fontisan
+  module Utils
+    autoload :Future, "fontisan/utils/thread_pool"
+    autoload :ThreadPool, "fontisan/utils/thread_pool"
+  end
+end

data/lib/fontisan/validation/collection_validator.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "../error"
-
 module Fontisan
   module Validation
     # CollectionValidator validates font compatibility for collection formats

data/lib/fontisan/validation.rb

@@ -0,0 +1,9 @@
+# frozen_string_literal: true
+
+# Autoload hub for the Fontisan::Validation namespace.
+
+module Fontisan
+  module Validation
+    autoload :CollectionValidator, "fontisan/validation/collection_validator"
+  end
+end

data/lib/fontisan/validators/basic_validator.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "validator"
-
 module Fontisan
   module Validators
     # BasicValidator provides minimal validation for fast font indexing

data/lib/fontisan/validators/font_book_validator.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "basic_validator"
-
 module Fontisan
   module Validators
     # FontBookValidator provides macOS Font Book installation compatibility checks

data/lib/fontisan/validators/opentype_validator.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "font_book_validator"
-
 module Fontisan
   module Validators
     # OpenTypeValidator provides comprehensive OpenType specification compliance checks

data/lib/fontisan/validators/profile_loader.rb

@@ -1,10 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "basic_validator"
-require_relative "font_book_validator"
-require_relative "opentype_validator"
-require_relative "web_font_validator"
-
 module Fontisan
   module Validators
     # ProfileLoader manages validation profiles and loads appropriate validators

data/lib/fontisan/validators/validator.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "../models/validation_report"
-
 module Fontisan
   module Validators
     # Base class for all validators using block-based DSL

data/lib/fontisan/validators/web_font_validator.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "basic_validator"
-
 module Fontisan
   module Validators
     # WebFontValidator provides web font optimization and embedding compatibility checks

data/lib/fontisan/validators.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+# Autoload hub for the Fontisan::Validators namespace.
+
+module Fontisan
+  module Validators
+    autoload :BasicValidator, "fontisan/validators/basic_validator"
+    autoload :FontBookValidator, "fontisan/validators/font_book_validator"
+    autoload :OpenTypeValidator, "fontisan/validators/opentype_validator"
+    autoload :ProfileLoader, "fontisan/validators/profile_loader"
+    autoload :Validator, "fontisan/validators/validator"
+    autoload :WebFontValidator, "fontisan/validators/web_font_validator"
+  end
+end

data/lib/fontisan/variable/delta_applicator.rb

@@ -1,10 +1,6 @@
 # frozen_string_literal: true
 
 require "yaml"
-require_relative "axis_normalizer"
-require_relative "region_matcher"
-require_relative "glyph_delta_processor"
-require_relative "metric_delta_processor"
 
 module Fontisan
   module Variable

data/lib/fontisan/variable/instancer.rb

@@ -1,8 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "delta_applicator"
-require_relative "static_font_builder"
-
 module Fontisan
   module Variable
     # Main entry point for variable font instancing

data/lib/fontisan/variable/static_font_builder.rb

@@ -1,8 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "table_updater"
-require_relative "../font_writer"
-
 module Fontisan
   module Variable
     # Builds static font instances from variable font data

data/lib/fontisan/variable.rb

@@ -0,0 +1,16 @@
+# frozen_string_literal: true
+
+# Autoload hub for the Fontisan::Variable namespace.
+
+module Fontisan
+  module Variable
+    autoload :AxisNormalizer, "fontisan/variable/axis_normalizer"
+    autoload :DeltaApplicator, "fontisan/variable/delta_applicator"
+    autoload :GlyphDeltaProcessor, "fontisan/variable/glyph_delta_processor"
+    autoload :Instancer, "fontisan/variable/instancer"
+    autoload :MetricDeltaProcessor, "fontisan/variable/metric_delta_processor"
+    autoload :RegionMatcher, "fontisan/variable/region_matcher"
+    autoload :StaticFontBuilder, "fontisan/variable/static_font_builder"
+    autoload :TableUpdater, "fontisan/variable/table_updater"
+  end
+end

data/lib/fontisan/variation/blend_applier.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "interpolator"
-
 module Fontisan
   module Variation
     # Applies CFF2 blend operators during CharString execution

data/lib/fontisan/variation/cache.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "cache_key_builder"
-
 module Fontisan
   module Variation
     # Caches variation calculations for performance

data/lib/fontisan/variation/converter.rb

@@ -1,8 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "interpolator"
-require_relative "table_accessor"
-
 module Fontisan
   module Variation
     # Converts variation data between TrueType (gvar) and CFF2 (blend) formats

data/lib/fontisan/variation/data_extractor.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "variation_context"
-
 module Fontisan
   module Variation
     # Extracts variation data from OpenType variable fonts

data/lib/fontisan/variation/delta_applier.rb

@@ -1,10 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "delta_parser"
-require_relative "interpolator"
-require_relative "region_matcher"
-require_relative "table_accessor"
-
 module Fontisan
   module Variation
     # Applies variation deltas to glyph outlines

data/lib/fontisan/variation/inspector.rb

@@ -2,7 +2,6 @@
 
 require "json"
 require "yaml"
-require_relative "variation_context"
 
 module Fontisan
   module Variation

data/lib/fontisan/variation/instance_generator.rb

@@ -1,11 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "interpolator"
-require_relative "region_matcher"
-require_relative "metrics_adjuster"
-require_relative "variation_context"
-require_relative "table_accessor"
-
 module Fontisan
   module Variation
     # Generates static font instances from variable fonts

data/lib/fontisan/variation/instance_writer.rb

@@ -1,10 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "../font_writer"
-require_relative "../converters/outline_converter"
-require_relative "../converters/woff_writer"
-require_relative "../error"
-
 module Fontisan
   module Variation
     # Writes generated static font instances to files in various formats

data/lib/fontisan/variation/metrics_adjuster.rb

@@ -1,9 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "interpolator"
-require_relative "region_matcher"
-require_relative "table_accessor"
-
 module Fontisan
   module Variation
     # Applies metrics variation deltas to font metrics tables

data/lib/fontisan/variation/optimizer.rb

@@ -1,8 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "../optimizers/pattern_analyzer"
-require_relative "../optimizers/subroutine_optimizer"
-
 module Fontisan
   module Variation
     # Optimizes CFF subroutines for variable fonts

data/lib/fontisan/variation/parallel_generator.rb

@@ -1,8 +1,5 @@
 # frozen_string_literal: true
 
-require_relative "instance_generator"
-require_relative "cache"
-require_relative "../utils/thread_pool"
 require "etc"
 
 module Fontisan