silo/internal/ods/reader.go

package ods

import (
	"archive/zip"
	"bytes"
	"encoding/xml"
	"fmt"
	"io"
	"strconv"
	"strings"
)

// Read parses an ODS file from bytes and returns sheets and metadata.
func Read(data []byte) (*Workbook, error) {
	r, err := zip.NewReader(bytes.NewReader(data), int64(len(data)))
	if err != nil {
		return nil, fmt.Errorf("open zip: %w", err)
	}

	wb := &Workbook{
		Meta: make(map[string]string),
	}

	for _, f := range r.File {
		switch f.Name {
		case "content.xml":
			sheets, err := readContent(f)
			if err != nil {
				return nil, fmt.Errorf("read content.xml: %w", err)
			}
			wb.Sheets = sheets
		case "meta.xml":
			meta, err := readMeta(f)
			if err != nil {
				return nil, fmt.Errorf("read meta.xml: %w", err)
			}
			wb.Meta = meta
		}
	}

	return wb, nil
}

// readContent parses content.xml and extracts sheets.
func readContent(f *zip.File) ([]Sheet, error) {
	rc, err := f.Open()
	if err != nil {
		return nil, err
	}
	defer rc.Close()

	data, err := io.ReadAll(rc)
	if err != nil {
		return nil, err
	}

	return parseContentXML(data)
}

// parseContentXML extracts sheet data from content.xml bytes.
func parseContentXML(data []byte) ([]Sheet, error) {
	decoder := xml.NewDecoder(bytes.NewReader(data))

	var sheets []Sheet
	var currentSheet *Sheet
	var currentRow *Row
	var currentCellText strings.Builder
	var inTextP bool

	// Current cell attributes for the cell being parsed
	var cellValueType string
	var cellValue string
	var cellFormula string
	var cellRepeated int

	// Track row repeated
	var rowRepeated int

	for {
		tok, err := decoder.Token()
		if err == io.EOF {
			break
		}
		if err != nil {
			return nil, fmt.Errorf("xml decode: %w", err)
		}

		switch t := tok.(type) {
		case xml.StartElement:
			localName := t.Name.Local
			switch localName {
			case "table":
				name := getAttr(t.Attr, "name")
				sheets = append(sheets, Sheet{Name: name})
				currentSheet = &sheets[len(sheets)-1]

			case "table-column":
				if currentSheet != nil {
					col := Column{}
					vis := getAttr(t.Attr, "visibility")
					if vis == "collapse" || vis == "hidden" {
						col.Hidden = true
					}
					width := getAttrNS(t.Attr, "column-width")
					if width != "" {
						col.Width = width
					}
					// Handle repeated columns
					rep := getAttr(t.Attr, "number-columns-repeated")
					count := 1
					if rep != "" {
						if n, err := strconv.Atoi(rep); err == nil && n > 0 {
							count = n
						}
					}
					// Cap at reasonable max to avoid memory issues from huge repeated counts
					if count > 1024 {
						count = 1024
					}
					for i := 0; i < count; i++ {
						currentSheet.Columns = append(currentSheet.Columns, col)
					}
				}

			case "table-row":
				rowRepeated = 1
				rep := getAttr(t.Attr, "number-rows-repeated")
				if rep != "" {
					if n, err := strconv.Atoi(rep); err == nil && n > 0 {
						rowRepeated = n
					}
				}
				currentRow = &Row{}

			case "table-cell":
				cellValueType = getAttrNS(t.Attr, "value-type")
				cellValue = getAttrNS(t.Attr, "value")
				if cellValue == "" {
					cellValue = getAttrNS(t.Attr, "date-value")
				}
				cellFormula = getAttr(t.Attr, "formula")
				cellRepeated = 1
				rep := getAttr(t.Attr, "number-columns-repeated")
				if rep != "" {
					if n, err := strconv.Atoi(rep); err == nil && n > 0 {
						cellRepeated = n
					}
				}
				currentCellText.Reset()

			case "covered-table-cell":
				// Merged cell continuation -- treat as empty
				if currentRow != nil {
					rep := getAttr(t.Attr, "number-columns-repeated")
					count := 1
					if rep != "" {
						if n, err := strconv.Atoi(rep); err == nil && n > 0 {
							count = n
						}
					}
					if count > 1024 {
						count = 1024
					}
					for i := 0; i < count; i++ {
						currentRow.Cells = append(currentRow.Cells, Cell{Type: CellEmpty})
					}
				}

			case "p":
				inTextP = true
			}

		case xml.CharData:
			if inTextP {
				currentCellText.Write(t)
			}

		case xml.EndElement:
			localName := t.Name.Local
			switch localName {
			case "table":
				currentSheet = nil

			case "table-row":
				if currentRow != nil && currentSheet != nil {
					// Determine if the row is blank
					isBlank := true
					for _, c := range currentRow.Cells {
						if c.Type != CellEmpty && c.Value != "" {
							isBlank = false
							break
						}
					}
					currentRow.IsBlank = isBlank && len(currentRow.Cells) == 0

					// Cap row repeats to avoid memory blow-up from trailing empty rows
					if rowRepeated > 1 && isBlank {
						// Only emit one blank row for large repeats (trailing whitespace)
						if rowRepeated > 2 {
							rowRepeated = 1
						}
					}
					for i := 0; i < rowRepeated; i++ {
						rowCopy := Row{
							IsBlank: currentRow.IsBlank,
							Cells:   make([]Cell, len(currentRow.Cells)),
						}
						copy(rowCopy.Cells, currentRow.Cells)
						currentSheet.Rows = append(currentSheet.Rows, rowCopy)
					}
				}
				currentRow = nil

			case "table-cell":
				if currentRow != nil {
					cell := buildCell(cellValueType, cellValue, cellFormula, currentCellText.String())

					// Cap repeated to avoid memory issues from trailing empties
					if cellRepeated > 256 && cell.Type == CellEmpty && cell.Value == "" {
						cellRepeated = 1
					}
					for i := 0; i < cellRepeated; i++ {
						currentRow.Cells = append(currentRow.Cells, cell)
					}
				}
				cellValueType = ""
				cellValue = ""
				cellFormula = ""
				cellRepeated = 1
				currentCellText.Reset()

			case "p":
				inTextP = false
			}
		}
	}

	// Trim trailing empty rows from each sheet
	for i := range sheets {
		sheets[i].Rows = trimTrailingBlankRows(sheets[i].Rows)
	}

	// Trim trailing empty cells from each row
	for i := range sheets {
		for j := range sheets[i].Rows {
			sheets[i].Rows[j].Cells = trimTrailingEmptyCells(sheets[i].Rows[j].Cells)
		}
	}

	return sheets, nil
}

func buildCell(valueType, value, formula, text string) Cell {
	if formula != "" {
		return Cell{
			Type:    CellFormula,
			Formula: formula,
			Value:   text,
		}
	}

	switch valueType {
	case "float":
		// Prefer the office:value attribute for precision; fall back to text
		v := value
		if v == "" {
			v = text
		}
		return Cell{Type: CellFloat, Value: v}
	case "currency":
		v := value
		if v == "" {
			v = strings.TrimPrefix(text, "$")
			v = strings.ReplaceAll(v, ",", "")
		}
		return Cell{Type: CellCurrency, Value: v}
	case "date":
		v := value
		if v == "" {
			v = text
		}
		return Cell{Type: CellDate, Value: v}
	case "string":
		return Cell{Type: CellString, Value: text}
	default:
		if text != "" {
			return Cell{Type: CellString, Value: text}
		}
		return Cell{Type: CellEmpty}
	}
}

// readMeta parses meta.xml for custom Silo metadata.
func readMeta(f *zip.File) (map[string]string, error) {
	rc, err := f.Open()
	if err != nil {
		return nil, err
	}
	defer rc.Close()

	data, err := io.ReadAll(rc)
	if err != nil {
		return nil, err
	}

	return parseMetaXML(data)
}

func parseMetaXML(data []byte) (map[string]string, error) {
	decoder := xml.NewDecoder(bytes.NewReader(data))
	meta := make(map[string]string)

	var inUserDefined bool
	var userDefName string
	var textBuf strings.Builder

	for {
		tok, err := decoder.Token()
		if err == io.EOF {
			break
		}
		if err != nil {
			return nil, err
		}

		switch t := tok.(type) {
		case xml.StartElement:
			if t.Name.Local == "user-defined" {
				inUserDefined = true
				userDefName = getAttrNS(t.Attr, "name")
				textBuf.Reset()
			}
		case xml.CharData:
			if inUserDefined {
				textBuf.Write(t)
			}
		case xml.EndElement:
			if t.Name.Local == "user-defined" && inUserDefined {
				if userDefName == "_silo_meta" {
					// Parse key=value pairs
					for _, line := range strings.Split(textBuf.String(), "\n") {
						line = strings.TrimSpace(line)
						if idx := strings.Index(line, "="); idx > 0 {
							meta[line[:idx]] = line[idx+1:]
						}
					}
				} else if userDefName != "" {
					meta[userDefName] = textBuf.String()
				}
				inUserDefined = false
				userDefName = ""
			}
		}
	}

	return meta, nil
}

// getAttr returns the value of a local-name attribute (no namespace).
func getAttr(attrs []xml.Attr, localName string) string {
	for _, a := range attrs {
		if a.Name.Local == localName {
			return a.Value
		}
	}
	return ""
}

// getAttrNS returns the value of a local-name attribute, ignoring namespace.
func getAttrNS(attrs []xml.Attr, localName string) string {
	for _, a := range attrs {
		if a.Name.Local == localName {
			return a.Value
		}
	}
	return ""
}

func trimTrailingBlankRows(rows []Row) []Row {
	for len(rows) > 0 {
		last := rows[len(rows)-1]
		if last.IsBlank || isRowEmpty(last) {
			rows = rows[:len(rows)-1]
		} else {
			break
		}
	}
	return rows
}

func isRowEmpty(row Row) bool {
	for _, c := range row.Cells {
		if c.Type != CellEmpty && c.Value != "" {
			return false
		}
	}
	return true
}

func trimTrailingEmptyCells(cells []Cell) []Cell {
	for len(cells) > 0 {
		last := cells[len(cells)-1]
		if last.Type == CellEmpty && last.Value == "" {
			cells = cells[:len(cells)-1]
		} else {
			break
		}
	}
	return cells
}