KevinMidboe/linguist
1
0
Fork 0
mirror of https://github.com/KevinMidboe/linguist.git synced 2025-10-29 17:50:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

Change find_by_filename api to return all matching languages

Browse Source
This commit is contained in:
Joshua Peek
2012-08-03 13:53:12 -05:00
parent 4a06d2ea7e
commit 6014bd015e
3 changed files with 25 additions and 103 deletions
Download Patch File Download Diff File Expand all files Collapse all files

29
lib/linguist/blob_helper.rb
View File

@@ -410,14 +410,15 @@ module Linguist
def guess_language
return if binary_mime_type?
# Disambiguate between multiple language extensions
disambiguate_extension_language ||
possible_languages = Language.find_by_filename(name.to_s)
# See if there is a Language for the extension
Language.find_by_filename(name.to_s) ||
# Try to detect Language from shebang line
shebang_language
if possible_languages.length > 1
if result = Classifier.classify(Samples::DATA, data, possible_languages.map(&:name)).first
Language[result[0]]
end
else
possible_languages.first || shebang_language
end
end
# Internal: Get the lexer of the blob.
@@ -427,20 +428,6 @@ module Linguist
language ? language.lexer : Pygments::Lexer.find_by_name('Text only')
end
# Internal: Disambiguates between multiple language extensions.
#
# Returns a Language or nil.
def disambiguate_extension_language
if Language.ambiguous?(extname)
possible_languages = Language.all.select { |l| l.extensions.include?(extname) }.map(&:name)
if possible_languages.any?
if result = Classifier.classify(Samples::DATA, data, possible_languages).first
Language[result[0]]
end
end
end
end
# Internal: Extract the script name from the shebang line
#
# Requires Blob#data

36
lib/linguist/language.rb
View File

@@ -15,8 +15,8 @@ module Linguist
@index = {}
@name_index = {}
@alias_index = {}
@extension_index = {}
@filename_index = {}
@extension_index = Hash.new { |h,k| h[k] = [] }
@filename_index = Hash.new { |h,k| h[k] = [] }
# Valid Languages types
TYPES = [:data, :markup, :programming]
@@ -60,13 +60,7 @@ module Linguist
raise ArgumentError, "Extension is missing a '.': #{extension.inspect}"
end
unless ambiguous?(extension)
# Index the extension with a leading ".": ".rb"
@extension_index[extension] = language
# Index the extension without a leading ".": "rb"
@extension_index[extension.sub(/^\./, '')] = language
end
@extension_index[extension] << language
end
language.overrides.each do |extension|
@@ -82,7 +76,7 @@ module Linguist
end
language.filenames.each do |filename|
@filename_index[filename] = language
@filename_index[filename] << language
end
language
@@ -123,33 +117,19 @@ module Linguist
@alias_index[name]
end
# Public: Look up Language by extension.
#
# extension - The extension String. May include leading "."
#
# Examples
#
# Language.find_by_extension('.rb')
# # => #<Language name="Ruby">
#
# Returns the Language or nil if none was found.
def self.find_by_extension(extension)
@extension_index[extension]
end
# Public: Look up Language by filename.
# Public: Look up Languages by filename.
#
# filename - The path String.
#
# Examples
#
# Language.find_by_filename('foo.rb')
# # => #<Language name="Ruby">
# # => [#<Language name="Ruby">]
#
# Returns the Language or nil if none was found.
# Returns all matching Languages or [] if none were found.
def self.find_by_filename(filename)
basename, extname = File.basename(filename), File.extname(filename)
@filename_index[basename] || @extension_index[extname]
@filename_index[basename] + @extension_index[extname]
end
# Public: Look up Language by its name or lexer.