Module: Pdf2MarkdownOCR::LlmApi

Defined in:
lib/pdf2markdownOCR/llm_api.rb

Class Method Summary collapse

Class Method Details

.ocr_images(images) ⇒ Object 



32
33
34
35
36
37
38
39
40
41
42
43
44
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
 
# File 'lib/pdf2markdownOCR/llm_api.rb', line 32

def self.ocr_images(images)
  markdown_pages = []
  
  Pdf2MarkdownOCR.configuration.logger.info "OCR #{images.size} images"
  t1 = Process.clock_gettime(Process::CLOCK_MONOTONIC)
  hydra = Typhoeus::Hydra.new
  images.each_with_index do |image_path, index|

    payload = Pdf2MarkdownOCR::LlmApi.payload(image_path)
    request = Typhoeus::Request.new(
      "#{Pdf2MarkdownOCR.configuration.llm_api_url}/v1/chat/completions",
      method: :post,
      body: payload.to_json,
      headers: { "Content-Type" => "application/json" },
      timeout: 600 # Default OpenAI timeout is 600 seconds
    )

    request.on_complete do |response|
      if response.success?
        parsed_response = JSON.parse(response.body)
        markdown_page = parsed_response.dig("choices", 0, "message", "content") || ""

        if markdown_page && !markdown_page.empty?
          markdown_pages << { index: index, content: markdown_page }
        else
          Pdf2MarkdownOCR.configuration.logger.warn "Warning: No Markdown content generated for #{image_path}"
        end
      else
        Pdf2MarkdownOCR.configuration.logger.error "Error processing #{image_path}: #{response.return_message} (#{response.code})"
      end
    end



    hydra.queue(request)
  end
  hydra.run

  t2 = Process.clock_gettime(Process::CLOCK_MONOTONIC)
  Pdf2MarkdownOCR.configuration.logger.info "Total Image processing time: #{(t2 - t1).round(2)} seconds"

  markdown_content = markdown_pages.sort_by { |page| page[:index] }.map { |page| page[:content] }.join("\n\n---\n\n")
  markdown_content
end

.payload(image_path) ⇒ Object 



4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
 
# File 'lib/pdf2markdownOCR/llm_api.rb', line 4

def self.payload(image_path)
        
  image_url = Base64.strict_encode64(File.binread(image_path))

  payload = {
    model: Pdf2MarkdownOCR.configuration.llm_model,
    messages: [
      {
        role: "user",
        content: [
          {
            type: "text",
            text: "<image>\nFree OCR."
          },
          {
            type: "image_url",
            image_url: {
              url: "data:image/png;base64,#{image_url}"
            }
          },
        ]
      }
    ],
  }

  payload
end