Class: ChunkerRuby::BaseSplitter
- Inherits:
-
Object
- Object
- ChunkerRuby::BaseSplitter
- Defined in:
- lib/chunker_ruby/base_splitter.rb
Direct Known Subclasses
Character, Code, HTML, JSONSplitter, Markdown, RecursiveCharacter, Semantic, Sentence, Separator, SlidingWindow, Token
Instance Attribute Summary collapse
-
#chunk_overlap ⇒ Object
readonly
Returns the value of attribute chunk_overlap.
-
#chunk_size ⇒ Object
readonly
Returns the value of attribute chunk_size.
Instance Method Summary collapse
-
#initialize(chunk_size: 1000, chunk_overlap: 200, **options) ⇒ BaseSplitter
constructor
A new instance of BaseSplitter.
- #split(text, metadata: {}) ⇒ Object
- #split_many(texts) ⇒ Object
Constructor Details
#initialize(chunk_size: 1000, chunk_overlap: 200, **options) ⇒ BaseSplitter
Returns a new instance of BaseSplitter.
7 8 9 10 11 12 13 14 |
# File 'lib/chunker_ruby/base_splitter.rb', line 7 def initialize(chunk_size: 1000, chunk_overlap: 200, **) raise ArgumentError, "chunk_size must be positive" unless chunk_size > 0 raise ArgumentError, "chunk_overlap must be non-negative" unless chunk_overlap >= 0 raise ArgumentError, "chunk_overlap must be less than chunk_size" unless chunk_overlap < chunk_size @chunk_size = chunk_size @chunk_overlap = chunk_overlap end |
Instance Attribute Details
#chunk_overlap ⇒ Object (readonly)
Returns the value of attribute chunk_overlap.
5 6 7 |
# File 'lib/chunker_ruby/base_splitter.rb', line 5 def chunk_overlap @chunk_overlap end |
#chunk_size ⇒ Object (readonly)
Returns the value of attribute chunk_size.
5 6 7 |
# File 'lib/chunker_ruby/base_splitter.rb', line 5 def chunk_size @chunk_size end |
Instance Method Details
#split(text, metadata: {}) ⇒ Object
16 17 18 |
# File 'lib/chunker_ruby/base_splitter.rb', line 16 def split(text, metadata: {}) raise NotImplementedError, "#{self.class}#split must be implemented" end |
#split_many(texts) ⇒ Object
20 21 22 |
# File 'lib/chunker_ruby/base_splitter.rb', line 20 def split_many(texts) texts.flat_map.with_index { |t, i| split(t, metadata: { doc_index: i }) } end |