doc2text

type AbstractNum struct {
	AbstractNumID int              `xml:"abstractNumId,attr"`
	Levels        []NumberingLevel `xml:"lvl"`
}

type AbstractNumIdVal struct {
	Val int `xml:"val,attr"`
}

type ContentElement struct {
	Type    string     `json:"type"`
	Level   int        `json:"level,omitempty"`
	Text    string     `json:"text,omitempty"`
	Rows    [][]string `json:"rows,omitempty"`
	Items   []string   `json:"items,omitempty"`
	Ordered bool       `json:"ordered,omitempty"`
}

type Document struct {
	Paragraphs []Paragraph `xml:"body>p"`
	Tables     []Table     `xml:"body>tbl"`
}

type DocumentOutput struct {
	Metadata Metadata         `json:"metadata"`
	Content  []ContentElement `json:"content"`
}

type IlvlVal struct {
	Val int `xml:"val,attr"`
}

type JcVal struct {
	Val string `xml:"val,attr"`
}

type Metadata struct {
	Source      string `json:"source"`
	ExtractedAt string `json:"extracted_at"`
}

type Num struct {
	NumID         int              `xml:"numId,attr"`
	AbstractNumId AbstractNumIdVal `xml:"abstractNumId"`
}

type NumFmtVal struct {
	Val string `xml:"val,attr"`
}

type NumIdVal struct {
	Val int `xml:"val,attr"`
}

type NumPr struct {
	Ilvl  IlvlVal  `xml:"ilvl"`
	NumId NumIdVal `xml:"numId"`
}

type Numbering struct {
	AbstractNums []AbstractNum `xml:"abstractNum"`
	Nums         []Num         `xml:"num"`
}

type NumberingLevel struct {
	Level     int       `xml:"ilvl,attr"`
	NumFormat NumFmtVal `xml:"numFmt"` // "decimal", "bullet", "lowerLetter", etc.
}

type PStyleVal struct {
	Val string `xml:"val,attr"`
}

type Paragraph struct {
	Properties ParagraphProperties `xml:"pPr"`
	Runs       []Run               `xml:"r"`
}

type ParagraphProperties struct {
	PStyle PStyleVal `xml:"pStyle"`
	Jc     JcVal     `xml:"jc"`
	NumPr  NumPr     `xml:"numPr"`
}

type Run struct {
	Properties RunProperties `xml:"rPr"`
	Texts      []Text        `xml:"t"`
}

type RunProperties struct {
	Bold      bool   `xml:"b"`
	Italic    bool   `xml:"i"`
	FontSize  string `xml:"sz"`
	FontColor string `xml:"color"`
}

type Table struct {
	Rows []TableRow `xml:"tr"`
}

type TableCell struct {
	Text string `xml:"p>r>t"`
}

type TableRow struct {
	Cells []TableCell `xml:"tc"`
}

type Text struct {
	Text string `xml:",chardata"`
}