Class: Ragnar::Chunker

Inherits:
Object
  • Object
show all
Defined in:
lib/ragnar/chunker.rb

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(chunk_size: Ragnar::DEFAULT_CHUNK_SIZE, chunk_overlap: Ragnar::DEFAULT_CHUNK_OVERLAP) ⇒ Chunker

Returns a new instance of Chunker.



5
6
7
8
9
10
11
12
13
14
 
# File 'lib/ragnar/chunker.rb', line 5

def initialize(chunk_size: Ragnar::DEFAULT_CHUNK_SIZE, chunk_overlap: Ragnar::DEFAULT_CHUNK_OVERLAP)
  @chunk_size = chunk_size
  @chunk_overlap = chunk_overlap
  # Use RecursiveCharacterTextSplitter for better chunking
  @splitter = Baran::RecursiveCharacterTextSplitter.new(
    chunk_size: chunk_size,
    chunk_overlap: chunk_overlap,
    separators: ["\n\n", "\n", ". ", " ", ""]
  )
end

Instance Attribute Details

#chunk_overlap ⇒ Object (readonly)

Returns the value of attribute chunk_overlap.



3
4
5
 
# File 'lib/ragnar/chunker.rb', line 3

def chunk_overlap
  @chunk_overlap
end

#chunk_size ⇒ Object (readonly)

Returns the value of attribute chunk_size.



3
4
5
 
# File 'lib/ragnar/chunker.rb', line 3

def chunk_size
  @chunk_size
end

Class Method Details

.semantic_chunker(model: nil) ⇒ Object 



87
88
89
90
91
92
93
94
95
 
# File 'lib/ragnar/chunker.rb', line 87

def self.semantic_chunker(model: nil)
  # Future enhancement: Use more sophisticated chunking with semantic boundaries
  # Could use sentence embeddings to find natural break points
  Baran::RecursiveCharacterTextSplitter.new(
    chunk_size: Ragnar::DEFAULT_CHUNK_SIZE,
    chunk_overlap: Ragnar::DEFAULT_CHUNK_OVERLAP,
    separators: ["\n\n", "\n", ". ", " ", ""]
  )
end

Instance Method Details

#chunk_documents(documents) ⇒ Object 



64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
 
# File 'lib/ragnar/chunker.rb', line 64

def chunk_documents(documents)
  all_chunks = []
  
  documents.each do |doc|
    if doc.is_a?(String)
      # If it's a file path
      if File.exist?(doc)
        all_chunks.concat(chunk_file(doc))
      else
        # Treat as raw text
        all_chunks.concat(chunk_text(doc))
      end
    elsif doc.is_a?(Hash)
      # If it's a document hash with text and metadata
      text = doc[:text] || doc["text"]
      metadata = doc[:metadata] || doc["metadata"] || {}
      all_chunks.concat(chunk_text(text, metadata))
    end
  end
  
  all_chunks
end

#chunk_file(file_path) ⇒ Object 



47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
 
# File 'lib/ragnar/chunker.rb', line 47

def chunk_file(file_path)
  unless File.exist?(file_path)
    raise "File not found: #{file_path}"
  end
  
  text = File.read(file_path, encoding: 'utf-8', invalid: :replace, undef: :replace)
  
  metadata = {
    file_path: File.absolute_path(file_path),
    file_name: File.basename(file_path),
    file_size: File.size(file_path),
    file_modified: File.mtime(file_path).to_s
  }
  
  chunk_text(text, metadata)
end

#chunk_text(text, metadata = {}) ⇒ Object 



16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
 
# File 'lib/ragnar/chunker.rb', line 16

def chunk_text(text, metadata = {})
  return [] if text.nil? || text.strip.empty?
  
  # Use Baran to split the text into chunks
  chunks = @splitter.chunks(text)
  
  # Add metadata to each chunk
  # Baran returns chunks as hashes with :text and :cursor keys
  chunks.map.with_index do |chunk_data, index|
    # Extract the actual text from the chunk
    chunk_text = if chunk_data.is_a?(Hash)
      chunk_data[:text] || chunk_data["text"]
    else
      chunk_data.to_s
    end
    
    {
      text: chunk_text,
      index: index,
      metadata: metadata.merge(
        chunk_index: index,
        total_chunks: chunks.size,
        chunk_size: chunk_text.size
      )
    }
  end
rescue => e
  puts "Error chunking text: #{e.message}"
  []
end