extraction

type CSVText ¶ added in v1.2.1

type CSVText struct {
	// contains filtered or unexported fields
}

CSVText converts CSV content into an indexable text representation.

The output contains three sections joined by newlines:

1. All column header names (if a header row is present).
2. Deduplicated string values from every non-numeric column.
3. The first few data rows written back as CSV.

A column is considered numeric when every non-empty value in that column can be parsed as a float64. Columns with at least one non-numeric value are treated as string columns.

func NewCSVText() *CSVText

func (c *CSVText) Text(content []byte) (string, error)

type DocumentText ¶

type DocumentText struct{}

DocumentText extracts plain text from binary document files using tabula.

func NewDocumentText() *DocumentText

func (d *DocumentText) Text(path string) (string, error)

type Extractors ¶ added in v1.2.1

type Extractors struct {
	// contains filtered or unexported fields
}

Extractors maps file extensions to text extractors.

func NewExtractors() *Extractors

func (e *Extractors) For(ext string) TextExtractor

type PDFTextRenderer ¶ added in v1.3.1

type PDFTextRenderer struct{}

PDFTextRenderer extracts text from individual PDF pages using tabula.

func NewPDFTextRenderer() *PDFTextRenderer

func (r *PDFTextRenderer) Close() error

func (r *PDFTextRenderer) PageCount(path string) (int, error)

func (r *PDFTextRenderer) Render(path string, page int) (string, error)

type PPTXTextRenderer ¶ added in v1.3.1

type PPTXTextRenderer struct{}

PPTXTextRenderer extracts text from individual presentation slides using tabula.

func NewPPTXTextRenderer() *PPTXTextRenderer

func (r *PPTXTextRenderer) Close() error

func (r *PPTXTextRenderer) PageCount(path string) (int, error)

func (r *PPTXTextRenderer) Render(path string, page int) (string, error)

type PageBoundary ¶ added in v1.3.1

type PageBoundary struct {
	Page       int // 1-based page number
	ByteOffset int // byte offset in the concatenated text
}

PageBoundary records where a page's text begins in a concatenated string.

type SinglePageTextRenderer ¶ added in v1.3.1

type SinglePageTextRenderer struct{}

SinglePageTextRenderer extracts text from document formats that are treated as a single page (DOCX, ODT, EPUB). PageCount always returns 1 and only page 1 can be rendered.

func NewSinglePageTextRenderer() *SinglePageTextRenderer

func (r *SinglePageTextRenderer) Close() error

func (r *SinglePageTextRenderer) PageCount(path string) (int, error)

func (r *SinglePageTextRenderer) Render(path string, page int) (string, error)

type SourceText ¶ added in v1.2.1

type SourceText struct{}

SourceText treats file content as plain text. Binary files (containing null bytes in the first 8 KB) produce empty text.

func NewSourceText() *SourceText

func (s *SourceText) Text(content []byte) (string, error)

type TextExtractor ¶ added in v1.2.1

type TextExtractor interface {
	Text(content []byte) (string, error)
}

TextExtractor converts raw file bytes into indexable plain text. Returns empty string when the content should be skipped.

type TextRenderer ¶ added in v1.3.1

type TextRenderer interface {
	io.Closer

	// PageCount returns the number of extractable pages in the document.
	PageCount(path string) (int, error)

	// Render returns the text content of the given 1-based page.
	Render(path string, page int) (string, error)
}

TextRenderer extracts text from individual document pages. For PDFs this means pages; for spreadsheets, sheets; for presentations, slides.

type TextRendererRegistry ¶ added in v1.3.1

type TextRendererRegistry struct {
	// contains filtered or unexported fields
}

TextRendererRegistry maps file extensions to TextRenderer implementations.

func NewTextRendererRegistry() *TextRendererRegistry

func (r *TextRendererRegistry) Close() error

func (r *TextRendererRegistry) For(ext string) (TextRenderer, bool)

func (r *TextRendererRegistry) Register(ext string, renderer TextRenderer)

func (r *TextRendererRegistry) Supports(ext string) bool

type XLSXTextRenderer ¶ added in v1.3.1

type XLSXTextRenderer struct{}

XLSXTextRenderer extracts text from individual spreadsheet sheets using tabula.

func NewXLSXTextRenderer() *XLSXTextRenderer

func (r *XLSXTextRenderer) Close() error

func (r *XLSXTextRenderer) PageCount(path string) (int, error)

func (r *XLSXTextRenderer) Render(path string, page int) (string, error)