ecosystems-bibliothecary 14.2.0 → 15.0.0

data/lib/bibliothecary/parsers/vcpkg.rb

@@ -6,35 +6,86 @@ module Bibliothecary
       def self.mapping
         {
           match_filename("vcpkg.json", case_insensitive: true) => {
-            kind: 'manifest',
+            kind: "manifest",
             parser: :parse_vcpkg_json,
-          }
+          },
+          match_filename("_generated-vcpkg-list.json") => {
+            kind: "lockfile",
+            parser: :parse_vcpkg_list_json,
+          },
         }
       end
 
+
       def self.parse_vcpkg_json(file_contents, options: {})
-        source = options.fetch(:filename, 'vcpkg.json')
+        source = options.fetch(:filename, nil)
         json = JSON.parse(file_contents)
-        deps = json["dependencies"].map do |dependency|
-          name = dependency.is_a?(Hash) ? dependency['name'] : dependency
-          if dependency.is_a?(Hash)
-            if dependency['version>=']
-              requirement = ">=#{dependency['version>=']}"
-            else
-              requirement = dependency['version'] || dependency['version-semver'] || dependency['version-date'] || dependency['version-string'] || '*'
-            end
+        deps = json["dependencies"] || []
+        return ParserResult.new(dependencies: []) if deps.empty?
+
+        overrides = {}
+        json["overrides"]&.each do |override|
+          next unless override.is_a?(Hash) && override["name"]
+
+          version = override["version"] || override["version-semver"] || override["version-date"] || override["version-string"]
+          overrides[override["name"]] = format_requirement(version, override["port-version"])
+        end
+
+        dependencies = deps.map do |dependency|
+          if dependency.is_a?(String)
+            name = dependency
+            requirement = nil
+            is_dev = false
           else
-            requirement = '*'
+            name = dependency["name"]
+            requirement = dependency["version>="] ? ">=#{dependency['version>=']}" : nil
+            is_dev = dependency["host"] == true
           end
-          Bibliothecary::Dependency.new(
+
+          next if name.nil? || name.empty?
+
+          requirement = overrides[name] if overrides[name]
+
+          Dependency.new(
             platform: platform_name,
             name: name,
-            requirement: requirement,
-            type: 'runtime',
+            requirement: requirement || "*",
+            type: is_dev ? "development" : "runtime",
             source: source
           )
-        end.uniq
-        Bibliothecary::ParserResult.new(dependencies: deps)
+        end.compact.uniq
+
+        ParserResult.new(dependencies: dependencies)
+      end
+
+      def self.parse_vcpkg_list_json(file_contents, options: {})
+        source = options.fetch(:filename, nil)
+        json = JSON.parse(file_contents)
+
+        dependencies = json.values.map do |package_info|
+          name = package_info["package_name"]
+          next if name.nil? || name.empty?
+
+          Dependency.new(
+            platform: platform_name,
+            name: name,
+            requirement: format_requirement(package_info["version"], package_info["port_version"]),
+            type: "runtime",
+            source: source
+          )
+        end.compact
+
+        ParserResult.new(dependencies: dependencies)
+      end
+
+      def self.format_requirement(version, port_version)
+        return "*" unless version
+
+        if port_version && port_version > 0
+          "#{version}##{port_version}"
+        else
+          version
+        end
       end
     end
   end

data/lib/bibliothecary/runner.rb

@@ -40,12 +40,159 @@ module Bibliothecary
     end
 
     def applicable_package_managers(info)
-      managers = package_managers.select { |pm| pm.match_info?(info) }
+      candidates = candidate_package_managers(info.relative_path)
+      managers = candidates.select { |pm| pm.match_info?(info) }
       managers.empty? ? [nil] : managers
     end
 
     def package_managers
-      Bibliothecary::Parsers.constants.map { |c| Bibliothecary::Parsers.const_get(c) }.sort_by { |c| c.to_s.downcase }
+      @package_managers ||= Bibliothecary::Parsers.constants
+        .map { |c| Bibliothecary::Parsers.const_get(c) }
+        .sort_by { |c| c.to_s.downcase }
+        .freeze
+    end
+
+    # Get candidate package managers for a file path using filename/extension index.
+    # Falls back to all package managers for unindexed patterns.
+    def candidate_package_managers(path)
+      filename = File.basename(path)
+      filename_lower = filename.downcase
+
+      # Check exact filename match first (use fetch to avoid default block on frozen hash)
+      candidates = filename_index.fetch(filename_lower, nil)
+      return candidates if candidates
+
+      # Check extension matches
+      extension_index.each do |ext, ext_candidates|
+        return ext_candidates if filename_lower.end_with?(ext)
+      end
+
+      # Fall back to all package managers for unindexed patterns
+      package_managers
+    end
+
+    # Build an index mapping lowercase filenames to candidate parsers
+    def filename_index
+      @filename_index ||= build_filename_index
+    end
+
+    # Build an index mapping lowercase extensions to candidate parsers
+    def extension_index
+      @extension_index ||= build_extension_index
+    end
+
+    def build_filename_index
+      index = {}
+
+      package_managers.each do |pm|
+        pm.mapping.each_key do |matcher|
+          next unless matcher.is_a?(Proc)
+
+          # Extract filenames from the matcher by testing common patterns
+          extract_filenames_from_matcher(matcher).each do |filename|
+            key = filename.downcase
+            index[key] ||= []
+            index[key] << pm
+          end
+        end
+      end
+
+      # Deduplicate and freeze
+      index.transform_values! { |v| v.uniq.freeze }
+      index.freeze
+    end
+
+    def build_extension_index
+      index = {}
+
+      package_managers.each do |pm|
+        pm.mapping.each_key do |matcher|
+          next unless matcher.is_a?(Proc)
+
+          # Extract extensions from the matcher
+          extract_extensions_from_matcher(matcher).each do |ext|
+            key = ext.downcase
+            index[key] ||= []
+            index[key] << pm
+          end
+        end
+      end
+
+      # Deduplicate and freeze
+      index.transform_values! { |v| v.uniq.freeze }
+      index.freeze
+    end
+
+    # Try to extract filename patterns from a matcher proc
+    def extract_filenames_from_matcher(matcher)
+      filenames = []
+
+      # Test common manifest filenames to see which ones match
+      common_filenames.each do |filename|
+        filenames << filename if matcher.call(filename)
+      end
+
+      filenames
+    end
+
+    # Try to extract extension patterns from a matcher proc
+    def extract_extensions_from_matcher(matcher)
+      extensions = []
+
+      # Test common extensions
+      common_extensions.each do |ext|
+        test_file = "test#{ext}"
+        extensions << ext if matcher.call(test_file)
+      end
+
+      extensions
+    end
+
+    def common_filenames
+      @common_filenames ||= %w[
+        package.json package-lock.json yarn.lock pnpm-lock.yaml npm-shrinkwrap.json npm-ls.json bun.lock
+        Gemfile Gemfile.lock gems.rb gems.locked
+        Cargo.toml Cargo.lock
+        go.mod go.sum Gopkg.toml Gopkg.lock glide.yaml glide.lock Godeps
+        requirements.txt Pipfile Pipfile.lock pyproject.toml poetry.lock setup.py
+        pom.xml build.gradle build.gradle.kts ivy.xml
+        composer.json composer.lock
+        Podfile Podfile.lock
+        pubspec.yaml pubspec.lock
+        Package.swift Package.resolved
+        Cartfile Cartfile.resolved Cartfile.private
+        mix.exs mix.lock
+        project.clj
+        shard.yml shard.lock
+        environment.yml environment.yaml
+        bower.json
+        elm-package.json elm.json
+        vcpkg.json
+        dub.json dub.sdl
+        haxelib.json
+        action.yml action.yaml
+        Brewfile Brewfile.lock.json
+        REQUIRE Project.toml Manifest.toml
+        paket.lock packages.config Project.json Project.lock.json packages.lock.json project.assets.json
+        DESCRIPTION
+        META.json META.yml cpanfile
+        cabal.config
+        docker-compose.yml docker-compose.yaml Dockerfile
+        conanfile.py conanfile.txt conan.lock
+        _generated-vcpkg-list.json
+        MLmodel
+        Modelfile
+        dvc.yaml
+        cog.yaml
+        bentofile.yaml
+        uv.lock pylock.toml
+      ].freeze
+    end
+
+    def common_extensions
+      @common_extensions ||= %w[
+        .gemspec .nuspec .csproj .cabal .podspec .podspec.json
+      ].freeze
     end
 
     # Parses an array of format [{file_path: "", contents: ""},] to match
@@ -120,7 +267,9 @@ module Bibliothecary
     def analyse_file(file_path, contents)
       contents = Bibliothecary.utf8_string(contents)
 
-      package_managers.select { |pm| pm.match?(file_path, contents) }.map do |pm|
+      # Use filename index to quickly find candidate parsers
+      candidates = candidate_package_managers(file_path)
+      candidates.select { |pm| pm.match?(file_path, contents) }.map do |pm|
         pm.analyse_contents(file_path, contents, options: @options)
       end.flatten.uniq.compact
     end
@@ -137,14 +286,24 @@ module Bibliothecary
         ignored_dirs.include?(f) || f.start_with?(*ignored_dirs_with_slash)
       end
       allowed_file_list = allowed_file_list.reject { |f| ignored_files.include?(f) }
-      package_managers.map do |pm|
-        # (skip rubocop false positive, since match? is a custom method)
-        allowed_file_list.select do |file_path| # rubocop:disable Style/SelectByRegexp
-          # this is a call to match? without file contents, which will skip
-          # ambiguous filenames that are only possibly a manifest
-          pm.match?(file_path)
+
+      # Fast path: use filename index directly for known manifest filenames
+      # This avoids creating FileInfo objects and calling match? for each file
+      manifests = []
+      allowed_file_list.each do |file_path|
+        filename_lower = File.basename(file_path).downcase
+
+        # Check if this filename is in our index (known manifest)
+        if filename_index.key?(filename_lower)
+          manifests << file_path
+          next
         end
-      end.flatten.uniq.compact
+
+        # Check extension index
+        matched = extension_index.keys.any? { |ext| filename_lower.end_with?(ext) }
+        manifests << file_path if matched
+      end
+      manifests.sort
     end
 
     def ignored_dirs
@@ -155,16 +314,6 @@ module Bibliothecary
       @configuration.ignored_files
     end
 
-    # We don't know what file groups are in multi file manifests until
-    # we process them. In those cases, process those, then reject the
-    # RelatedFilesInfo objects that aren't in the manifest.
-    #
-    # This means we're likely analyzing these files twice in processing,
-    # but we need that accurate package manager information.
-    def filter_multi_manifest_entries(path, related_files_info_entries)
-      MultiManifestFilter.new(path: path, related_files_info_entries: related_files_info_entries, runner: self).results
-    end
-
     private
 
     # Get the list of all package managers that apply to the file provided
@@ -180,5 +329,3 @@ module Bibliothecary
     end
   end
 end
-
-require_relative "runner/multi_manifest_filter"

data/lib/bibliothecary/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Bibliothecary
-  VERSION = "14.2.0"
+  VERSION = "15.0.0"
 end

data/lib/bibliothecary.rb

@@ -9,13 +9,9 @@ require "bibliothecary/runner"
 require "bibliothecary/exceptions"
 require "bibliothecary/file_info"
 require "bibliothecary/related_files_info"
-require "bibliothecary/purl_util"
 require "find"
 require "tomlrb"
 
-Dir[File.expand_path("bibliothecary/multi_parsers/*.rb", __dir__)].each do |file|
-  require file
-end
 Dir[File.expand_path("bibliothecary/parsers/*.rb", __dir__)].each do |file|
   require file
 end
@@ -100,19 +96,16 @@ module Bibliothecary
   end
 
   def self.runner
-    configuration
-    @runner
+    @runner ||= Runner.new(configuration)
   end
 
   def self.configuration
     @configuration ||= Configuration.new
-    @runner = Runner.new(@configuration)
-    @configuration
   end
 
   def self.reset
-    @configuration = Configuration.new
-    @runner = Runner.new(@configuration)
+    @configuration = nil
+    @runner = nil
   end
 
   def self.configure

data/lib/dockerfile_parser.rb

@@ -4,7 +4,7 @@ class DockerfileParser
   end
 
   def parse
-    fromlines = @file_contents.split("\n").select { |line| line.strip =~ /^\FROM/i }
+    fromlines = @file_contents.split("\n").select { |line| line.strip =~ /^FROM/i }
     
     fromlines.map do |line|
       line = line.strip.split(' ')

data/lib/modelfile_parser.rb

@@ -4,10 +4,10 @@ class ModelfileParser
   end
 
   def parse
-    fromlines = @file_contents.split("\n").select { |line| line.strip =~ /^\FROM/i }
+    fromlines = @file_contents.split("\n").select { |line| line.strip =~ /^FROM/i }
 
     fromlines.map do |line|
-      line = line.strip.split(' ')
+      line = line.strip.split
 
       # Remove the FROM keyword
       line.shift
@@ -23,19 +23,19 @@ class ModelfileParser
       model_ref = line[0]
 
       # Check if it's a file path (local GGUF or directory)
-      if model_ref =~ /\.(gguf|safetensors)$/i || model_ref.start_with?('./', '/')
+      if model_ref =~ /\.(gguf|safetensors)$/i || model_ref.start_with?("./", "/")
         {
           name: File.basename(model_ref),
-          requirement: 'local',
-          type: 'runtime'
+          requirement: "local",
+          type: "runtime",
         }
       else
         # It's a registry model (e.g., llama3.2 or llama3.2:latest)
-        parts = model_ref.split(':')
+        parts = model_ref.split(":")
         {
           name: parts[0],
-          requirement: parts[1] || 'latest',
-          type: 'runtime'
+          requirement: parts[1] || "latest",
+          type: "runtime",
         }
       end
     end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: ecosystems-bibliothecary
 version: !ruby/object:Gem::Version
-  version: 14.2.0
+  version: 15.0.0
 platform: ruby
 authors:
 - Andrew Nesbitt
@@ -23,20 +23,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: commander
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
 - !ruby/object:Gem::Dependency
   name: csv
   requirement: !ruby/object:Gem::Requirement
@@ -51,20 +37,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: deb_control
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
 - !ruby/object:Gem::Dependency
   name: json
   requirement: !ruby/object:Gem::Requirement
@@ -79,20 +51,6 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '2.8'
-- !ruby/object:Gem::Dependency
-  name: librariesio-gem-parser
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
 - !ruby/object:Gem::Dependency
   name: ox
   requirement: !ruby/object:Gem::Requirement
@@ -107,20 +65,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 2.8.1
-- !ruby/object:Gem::Dependency
-  name: packageurl-ruby
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
 - !ruby/object:Gem::Dependency
   name: racc
   requirement: !ruby/object:Gem::Requirement
@@ -135,20 +79,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: sdl4r
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
 - !ruby/object:Gem::Dependency
   name: tomlrb
   requirement: !ruby/object:Gem::Requirement
@@ -163,47 +93,18 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '2.0'
-- !ruby/object:Gem::Dependency
-  name: typhoeus
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
 email:
 - andrewnez@gmail.com
 executables:
 - bibliothecary
-- console
-- setup
 extensions: []
 extra_rdoc_files: []
 files:
-- ".codeclimate.yml"
-- ".github/CONTRIBUTING.md"
-- ".github/workflows/ci.yml"
-- ".gitignore"
-- ".rspec"
-- ".rubocop.yml"
-- ".ruby-version"
-- ".tidelift"
 - CHANGELOG.md
-- CODE_OF_CONDUCT.md
-- Gemfile
 - LICENSE.txt
 - README.md
-- Rakefile
 - bibliothecary.gemspec
 - bin/bibliothecary
-- bin/console
-- bin/setup
 - lib/bibliothecary.rb
 - lib/bibliothecary/analyser.rb
 - lib/bibliothecary/analyser/analysis.rb
@@ -214,11 +115,6 @@ files:
 - lib/bibliothecary/dependency.rb
 - lib/bibliothecary/exceptions.rb
 - lib/bibliothecary/file_info.rb
-- lib/bibliothecary/multi_parsers/bundler_like_manifest.rb
-- lib/bibliothecary/multi_parsers/cyclonedx.rb
-- lib/bibliothecary/multi_parsers/dependencies_csv.rb
-- lib/bibliothecary/multi_parsers/json_runtime.rb
-- lib/bibliothecary/multi_parsers/spdx.rb
 - lib/bibliothecary/parser_result.rb
 - lib/bibliothecary/parsers/actions.rb
 - lib/bibliothecary/parsers/bentoml.rb
@@ -228,6 +124,7 @@ files:
 - lib/bibliothecary/parsers/clojars.rb
 - lib/bibliothecary/parsers/cocoapods.rb
 - lib/bibliothecary/parsers/cog.rb
+- lib/bibliothecary/parsers/conan.rb
 - lib/bibliothecary/parsers/conda.rb
 - lib/bibliothecary/parsers/cpan.rb
 - lib/bibliothecary/parsers/cran.rb
@@ -254,14 +151,11 @@ files:
 - lib/bibliothecary/parsers/shard.rb
 - lib/bibliothecary/parsers/swift_pm.rb
 - lib/bibliothecary/parsers/vcpkg.rb
-- lib/bibliothecary/purl_util.rb
 - lib/bibliothecary/related_files_info.rb
 - lib/bibliothecary/runner.rb
-- lib/bibliothecary/runner/multi_manifest_filter.rb
 - lib/bibliothecary/version.rb
 - lib/dockerfile_parser.rb
 - lib/modelfile_parser.rb
-- lib/sdl_parser.rb
 homepage: https://github.com/ecosyste-ms/bibliothecary
 licenses:
 - AGPL-3.0

data/.codeclimate.yml

@@ -1,25 +0,0 @@
----
-engines:
-  duplication:
-    enabled: true
-    config:
-      languages:
-      - ruby
-      - javascript
-      - python
-      - php
-  fixme:
-    enabled: true
-  rubocop:
-    enabled: true
-ratings:
-  paths:
-  - "**.inc"
-  - "**.js"
-  - "**.jsx"
-  - "**.module"
-  - "**.php"
-  - "**.py"
-  - "**.rb"
-exclude_paths:
-- spec/