graph TB
    subgraph "ChunkWise Library"
        A[Chunker] --> B[BaseChunker]
        B --> C[Strategies]
        B --> D[Language Support]
        B --> E[Tokenizers]

        subgraph "Strategies"
            C --> C1[Basic]
            C --> C2[Document Structure]
            C --> C3[Format-Specific]
            C --> C4[Semantic]
            C --> C5[LLM-Based]
            C --> C6[Retrieval-Optimized]
            C --> C7[Hybrid]

            C1 --> C1a[CharacterChunker]
            C1 --> C1b[TokenChunker]
            C1 --> C1c[WordChunker]
            C1 --> C1d[SentenceChunker]
            C1 --> C1e[ParagraphChunker]
            C1 --> C1f[RecursiveChunker]
            C1 --> C1g[SlidingWindowChunker]

            C2 --> C2a[MarkdownChunker]
            C2 --> C2b[HTMLChunker]
            C2 --> C2c[CodeChunker]

            C3 --> C3a[JSONChunker]
            C3 --> C3b[LaTeXChunker]
            C3 --> C3c[RegexChunker]

            C4 --> C4a[SemanticChunker]
            C4 --> C4b[ClusterChunker]
            C4 --> C4c[LateChunker]

            C5 --> C5a[AgenticChunker]
            C5 --> C5b[ContextualChunker]
            C5 --> C5c[DocumentSummaryChunker]

            C6 --> C6a[SentenceWindowChunker]
            C6 --> C6b[ParentDocumentChunker]
            C6 --> C6c[HierarchicalChunker]

            C7 --> C7a[HybridChunker]
            C7 --> C7b[AdaptiveChunker]
        end

        subgraph "Language Support"
            D --> D1[Arabic]
            D --> D2[English]
            D --> D3[Detector]

            D1 --> D1a[Preprocessor]
            D1 --> D1b[Sentence Splitter]
            D1 --> D1c[Tokenizer]
        end

        subgraph "External Integrations"
            E --> E1[Tiktoken]
            F[Embeddings] --> F1[SentenceTransformers]
            F --> F2[OpenAI Embeddings]
            G[LLM] --> G1[OpenAI]
            G --> G2[Anthropic]
        end
    end

    H[Input Text] --> A
    A --> I[Chunk Objects]
    I --> J[RAG Pipeline]

graph LR
    subgraph "Core Modules"
        A[chunkwise/] --> B[__init__.py]
        A --> C[base.py]
        A --> D[chunk.py]
        A --> E[chunker.py]
        A --> F[config.py]
        A --> G[exceptions.py]
    end

    subgraph "Strategies Module"
        H[strategies/] --> H1[fixed.py]
        H --> H2[sentence.py]
        H --> H3[paragraph.py]
        H --> H4[recursive.py]
        H --> H5[semantic.py]
        H --> H6[agentic.py]
        H --> H7[contextual.py]
        H --> H8[hierarchical.py]
        H --> H9[format_specific.py]
        H --> H10[document_structure.py]
    end

    subgraph "Language Module"
        I[language/] --> I1[detector.py]
        I --> I2[arabic/]
        I --> I3[english/]
    end

    subgraph "Support Modules"
        J[tokenizers/] --> J1[tiktoken_tokenizer.py]
        K[embeddings/] --> K1[sentence_transformers_embed.py]
        L[llm/] --> L1[openai_llm.py]
    end

    A --> H
    A --> I
    A --> J
    A --> K
    A --> L

flowchart LR
    A[Raw Text] --> B{Language Detection}
    B -->|Arabic| C[Arabic Preprocessor]
    B -->|English| D[English Preprocessor]
    B -->|Mixed| E[Mixed Handler]

    C --> F[Chunking Strategy]
    D --> F
    E --> F

    F --> G{Strategy Type}
    G -->|Basic| H[Fixed/Sentence/Recursive]
    G -->|Semantic| I[Embedding-Based]
    G -->|LLM| J[AI-Powered]
    G -->|Retrieval| K[Window/Hierarchical]

    H --> L[Chunk Objects]
    I --> L
    J --> L
    K --> L

    L --> M[Metadata Enrichment]
    M --> N[Output Chunks]