Ruby Git Diff Line信息解析器

Question

如何解析git diff的输出并获取行信息（即哪些行已被添加/修改）？

我想要类似的东西

raw = `git diff`
parsed = Git.Diff.parse(raw)
parsed.each do |file|
  file.each do |line|
     puts "#{file.name} - #{line.number} - #{line.type}"
  end
end

Edit:

Sample output

[
  {
    "file": "path/to/file1",
    "lines": [
      { number: "1", type: "modified"},
      { number: "4", type: "deleted"},
      { number: "9", type: "added"}
    ]
  },
  {
    "file": "path/to/file2",
    "lines": [
      { number: "4", type: "modified"},
      { number: "5", type: "added"}
    ]
  }
]

Answer 1

您需要的是将输出正确分组到文件块中并保留所需内容。

Getting the diff

你可以通过简单地运行一个来获得它

`git --diff`

What lines are needed?

• 从'diff --git'开始的行，您可以从中获取文件的名称
• 以'+ '开头的行是添加的行
• 以'- '开头的行是删除的行

How to group them?

对于这些东西，Enumerable#slice_before浮现在脑海中。

Putting it together

我最终得到了这个原型：

raw_data = `git diff`.split("\n")

# Keep what is needed
clean_data = raw_data.select { |li|
  li.starts_with?('diff --git') ||
  li.starts_with?('- ') ||
  li.starts_with?('+ ')
}

# Group the by file
# [[file_1, line1, line2, line3], [file_2, line1]]
file_data = clean_data.slice_before { |li| li.starts_with?('diff --git') }

# This is the output format
output = Hash.new {|h,k| h[k] = { added: 0, removed: 0 } }

# Populate the output
file_data.each_with_object(output) do |f_data, memo|
  file, *file_info = f_data
  file = file.split(' b/').first.gsub('diff --git a/', '')
  file_info.each { |f_info|
    memo[file][f_info[0] == '+' ? :added : :removed] += 1
  }
end

Output example

{
  "file_1" => { added: 1, removed: 12 },
  "file_2" => { added: 0, removed: 1 }
}

我相信它会变得更好:-)

Answer 2

这就是我最终的结果

class Parser
        def parse(text)
            if text.encoding.name != "UTF-8"
                encoded_text = @full_diff.encode("UTF-8", "binary", { :invalid => :replace, :undef => :replace })
            else
                encoded_text = text
            end

            hunks = []
            hunk = nil
            added_line_number = nil
            deleted_line_number = nil

            lines = encoded_text.strip.split("\n")
            lines.each_with_index do |line, index|
                if m = /^diff --git a\/(.*?) b\/(.*?)$/.match(line)
                    raise "Diff formatting error, 'diff --git' is the last line" if index + 1 >= lines.length

                    # new hunk
                    added_line_number = nil
                    delete_line_number = nil
                    hunk = Hunk.new(m[1], m[2])
                    hunk.type = hunk_type(lines[index + 1], m[1], m[2])
                    hunks.push(hunk)
                elsif /^Binary files /.match(line)
                    hunk.is_binary = true
                elsif m = /^@@ \-(\d+)(?:,\d+)? \+(\d+)(?:,\d+)? @@/.match(line)
                    # (e.g. @@ -19,6 +19,7 @@)
                    deleted_line_number = Integer(m[1])
                    added_line_number = Integer(m[2])
                else
                    if !added_line_number.nil?
                        if line.start_with?('+')
                            # added line
                            hunk.lines.push SourceLine.new(added_line_number, SourceLine::Type::Added, line[1..-1])
                            added_line_number += 1
                        elsif line.start_with?('-')
                            # deleted line
                            hunk.lines.push SourceLine.new(deleted_line_number, SourceLine::Type::Deleted, line[1..-1])
                            deleted_line_number += 1
                        else
                            # unmodified line
                            added_line_number += 1
                            deleted_line_number += 1
                        end
                    end
                end
            end
            hunks
        end

        def hunk_type(line, original, renamed)
            case line
            when /^new file /
                type = Hunk::Type::Added
            when /^deleted file /
                type = Hunk::Type::Deleted
            else
                type = original == renamed ? Hunk::Type::Modified : Hunk::Type::Renamed
            end
            type
        end
        private :hunk_type
    end
end

module Type
    Added = 'added'
    Deleted = 'deleted'
    Modified = 'modified'
    Renamed = 'renamed'
end

class Hunk
    module Type
        Added = 'added'
        Deleted = 'deleted'
        Modified = 'modified'
        Renamed = 'renamed'
    end

    attr_accessor :original_path, :renamed_path, :type, :lines, :is_binary
    alias_method :is_binary?, :is_binary

    def initialize(original_path, renamed_path)
        self.is_binary = false
        self.lines = []
        self.original_path = original_path
        self.renamed_path = renamed_path
    end
end

class SourceLine
    module Type
        Added = 'added'
        Deleted = 'deleted'
    end

    attr_accessor :number, :type, :text
    def initialize(number, type, text)
        self.number = number
        self.type = type
        self.text = text
    end
end