Class: Mathpix::MCP::Tools::BatchConvertTool

Inherits:
BaseTool
  • Object
show all
Defined in:
lib/mathpix/mcp/tools/batch_convert_tool.rb

Overview

Batch Convert Tool

Converts multiple images in batch for efficiency Thin delegate to Mathpix::Client#snap with batch processing

Class Method Summary collapse

Class Method Details

.call(image_paths:, server_context:, formats: nil, parallel: false, max_parallel: 3, output_dir: nil) ⇒ Object 



45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
 
# File 'lib/mathpix/mcp/tools/batch_convert_tool.rb', line 45

def self.call(image_paths:, server_context:, formats: nil, parallel: false, max_parallel: 3, output_dir: nil)
  safe_execute do
    client = mathpix_client(server_context)

    # Extract formats or use defaults
    output_formats = extract_formats(formats, client)
    dir = output_dir && !output_dir.empty? ? File.expand_path(output_dir) : artifact_dir

    # Normalize paths
    normalized_paths = image_paths.map do |path|
      url?(path) ? path : normalize_path(path)
    end

    # Process images (concurrently when requested)
    results =
      if parallel
        process_batch_parallel(client, normalized_paths, output_formats, max_parallel.to_i, dir)
      else
        process_batch_sequential(client, normalized_paths, output_formats, dir)
      end

    # Format response
    response_data = {
      success: true,
      batch_size: image_paths.length,
      formats: output_formats,
      parallel: parallel,
      results: results,
      summary: {
        total: results.length,
        successful: results.count { |r| r[:success] },
        failed: results.count { |r| !r[:success] }
      }
    }

    json_response(response_data)
  end
end

.convert_one(client, path, index, formats, dir) ⇒ Object

Convert a single image, writing its recognized content to a file and returning a result/error hash (never raises). The OCR text/latex is saved to disk rather than inlined so a large batch can't overflow the model context.



88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
 
# File 'lib/mathpix/mcp/tools/batch_convert_tool.rb', line 88

def self.convert_one(client, path, index, formats, dir)
  result = client.snap(path, formats: formats)
  contents = {
    'latex' => result.latex,
    'text' => result.text,
    'mathml' => result.mathml,
    'asciimath' => result.asciimath
  }.compact
  stem = url?(path) ? "image_#{index}" : File.basename(path, File.extname(path))
  saved = write_artifacts(contents, File.join(dir, "mathpix_#{sanitize(stem)}.tex"))
  {
    index: index,
    image_path: path,
    success: true,
    confidence: result.confidence,
    saved_files: saved,
    preview: preview_of(result.latex || result.text, 200)
  }
rescue Mathpix::Error => e
  {
    index: index,
    image_path: path,
    success: false,
    error: e.message
  }
end

.process_batch_parallel(client, paths, formats, max_parallel, dir) ⇒ Object

Bounded thread pool: at most max_parallel concurrent HTTP requests. Results are written back by original index so ordering is preserved.



121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
 
# File 'lib/mathpix/mcp/tools/batch_convert_tool.rb', line 121

def self.process_batch_parallel(client, paths, formats, max_parallel, dir)
  max_parallel = 1 if max_parallel < 1
  results = Array.new(paths.length)
  queue = Queue.new
  paths.each_with_index { |path, i| queue << [path, i] }

  worker_count = [max_parallel, paths.length].min
  workers = Array.new(worker_count) do
    Thread.new do
      loop do
        path, index = begin
          queue.pop(true) # non-blocking; raises ThreadError when empty
        rescue ThreadError
          break
        end
        results[index] = convert_one(client, path, index, formats, dir)
      end
    end
  end

  workers.each(&:join)
  results
end

.process_batch_sequential(client, paths, formats, dir) ⇒ Object 



115
116
117
 
# File 'lib/mathpix/mcp/tools/batch_convert_tool.rb', line 115

def self.process_batch_sequential(client, paths, formats, dir)
  paths.map.with_index { |path, index| convert_one(client, path, index, formats, dir) }
end