Module: Commiti::GitReader

Defined in:
lib/services/git/git_reader.rb

Constant Summary collapse

MAX_DIFF_BYTES = 
50_000
TRUNCATION_NOTICE = 
"\n# ... diff clipped by Commiti to preserve context under size limit\n"
LOCKFILE_PATTERNS = 
[
  /Gemfile\.lock/,
  /package-lock\.json/,
  /yarn\.lock/,
  /pnpm-lock\.yaml/,
  /composer\.lock/,
  /mix\.lock/,
  /Cargo\.lock/,
  /Pipfile\.lock/
].freeze

Class Method Summary collapse

Class Method Details

.append_notice(clipped_diff, max_bytes:) ⇒ Object 



177
178
179
180
181
182
183
184
185
186
187
 
# File 'lib/services/git/git_reader.rb', line 177

def self.append_notice(clipped_diff, max_bytes:)
  safe_clipped = clipped_diff.to_s
  return safe_clipped if safe_clipped.bytesize >= max_bytes && max_bytes <= TRUNCATION_NOTICE.bytesize

  return safe_clipped + TRUNCATION_NOTICE if safe_clipped.bytesize + TRUNCATION_NOTICE.bytesize <= max_bytes

  available = max_bytes - TRUNCATION_NOTICE.bytesize
  return safe_clipped.byteslice(0, max_bytes) if available <= 0

  safe_clipped.byteslice(0, available) + TRUNCATION_NOTICE
end

.branch_diff(base_branch: 'main') ⇒ Object 



21
22
23
24
25
26
27
28
29
30
31
 
# File 'lib/services/git/git_reader.rb', line 21

def self.branch_diff(base_branch: 'main')
  raise 'Invalid branch name.' unless base_branch.match?(%r{\A[a-zA-Z0-9_\-./]+\z})

  # Strip context lines using -U0 and filter out binary/lockfile noise
  diff, status = Open3.capture2('git', 'diff', '-U0', "#{base_branch}...HEAD")
  raise "Failed to read branch diff against '#{base_branch}'." unless status.success?
  raise "No diff found against '#{base_branch}'." if diff.strip.empty?

  filtered_diff = filter_diff_noise(diff)
  clip_diff_context(filtered_diff, max_bytes: MAX_DIFF_BYTES)
end

.clip_chunks(chunks, max_bytes:) ⇒ Object 



98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
 
# File 'lib/services/git/git_reader.rb', line 98

def self.clip_chunks(chunks, max_bytes:)
  output = +''

  chunks.each do |chunk|
    remaining = max_bytes - output.bytesize
    break if remaining <= 0

    chunk_text = chunk[:lines].join
    if chunk_text.bytesize <= remaining
      output << chunk_text
      next
    end

    output << clip_single_chunk(chunk[:lines], max_bytes: remaining)
    break
  end

  if output.empty?
    first_chunk_text = chunks.first[:lines].join
    return first_chunk_text.byteslice(0, max_bytes)
  end

  output
end

.clip_diff_context(diff, max_bytes:) ⇒ Object 



38
39
40
41
42
43
44
45
46
47
48
49
 
# File 'lib/services/git/git_reader.rb', line 38

def self.clip_diff_context(diff, max_bytes:)
  return diff if diff.bytesize <= max_bytes

  chunks = split_by_file(diff)
  clipped = if chunks.empty?
              diff.byteslice(0, max_bytes)
            else
              clip_chunks(chunks, max_bytes: max_bytes)
            end

  append_notice(clipped, max_bytes: max_bytes)
end

.clip_single_chunk(lines, max_bytes:) ⇒ Object 



123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
 
# File 'lib/services/git/git_reader.rb', line 123

def self.clip_single_chunk(lines, max_bytes:)
  output = +''
  return output if max_bytes <= 0

  header_lines = []
  hunks = []
  current_hunk = nil
  in_hunks = false

  lines.each do |line|
    if line.start_with?('@@')
      in_hunks = true
      current_hunk = [line]
      hunks << current_hunk
      next
    end

    if in_hunks
      current_hunk << line
    else
      header_lines << line
    end
  end

  header_lines.each do |line|
    break if output.bytesize + line.bytesize > max_bytes

    output << line
  end

  return output if hunks.empty?

  hunks.each do |hunk|
    hunk_text = hunk.join
    if output.bytesize + hunk_text.bytesize <= max_bytes
      output << hunk_text
      next
    end

    hunk_header = hunk.first
    break if output.bytesize + hunk_header.bytesize > max_bytes

    output << hunk_header
    hunk[1..].to_a.each do |line|
      break if output.bytesize + line.bytesize > max_bytes

      output << line
    end
    break
  end

  output
end

.recent_commits(count: 10) ⇒ Object 



33
34
35
36
 
# File 'lib/services/git/git_reader.rb', line 33

def self.recent_commits(count: 10)
  out, = Open3.capture2('git', 'log', '--oneline', "-#{count}")
  out
end

.split_by_file(diff) ⇒ Object

Returns [{ path: String, lines: Array<String> }]



94
95
96
 
# File 'lib/services/git/git_reader.rb', line 94

def self.split_by_file(diff)
  Commiti::DiffParser.split_by_file_lines(diff)
end

.staged_diff ⇒ Object 



11
12
13
14
15
16
17
18
19
 
# File 'lib/services/git/git_reader.rb', line 11

def self.staged_diff
  # Strip context lines using -U0 and filter out binary/lockfile noise
  diff, status = Open3.capture2('git', 'diff', '--cached', '-U0')
  raise 'Failed to read staged diff.' unless status.success?
  raise 'No staged changes. Run `git add` first.' if diff.strip.empty?

  filtered_diff = filter_diff_noise(diff)
  clip_diff_context(filtered_diff, max_bytes: MAX_DIFF_BYTES)
end