package notion
import (
"encoding/json"
"fmt"
"strings"
)
// =============================================================================
// Markdown Converter for Notion Blocks
// Converts Notion block JSON to readable Markdown format
// =============================================================================
// BlocksToMarkdown converts Notion blocks JSON response to Markdown
func BlocksToMarkdown(jsonStr string) string {
var data map[string]any
if err := json.Unmarshal([]byte(jsonStr), &data); err != nil {
return jsonStr
}
results, ok := data["results"].([]any)
if !ok {
return ""
}
var sb strings.Builder
var listStack []string // Track nested list types
for _, item := range results {
block, ok := item.(map[string]any)
if !ok {
continue
}
blockType := getString(block, "type")
md := blockToMarkdown(block, blockType, &listStack)
if md != "" {
sb.WriteString(md)
}
}
// Add pagination info if has_more
if hasMore, ok := data["has_more"].(bool); ok && hasMore {
if cursor, ok := data["next_cursor"].(string); ok {
sb.WriteString(fmt.Sprintf("\n---\n[more: %s]\n", cursor))
}
}
return strings.TrimSuffix(sb.String(), "\n")
}
// blockToMarkdown converts a single block to Markdown
func blockToMarkdown(block map[string]any, blockType string, listStack *[]string) string {
content, ok := block[blockType].(map[string]any)
if !ok && blockType != "unsupported" {
return ""
}
switch blockType {
// Headings
case "heading_1":
return fmt.Sprintf("# %s\n\n", extractRichTextMD(content))
case "heading_2":
return fmt.Sprintf("## %s\n\n", extractRichTextMD(content))
case "heading_3":
return fmt.Sprintf("### %s\n\n", extractRichTextMD(content))
// Paragraph
case "paragraph":
text := extractRichTextMD(content)
if text == "" {
return "\n"
}
return fmt.Sprintf("%s\n\n", text)
// Lists
case "bulleted_list_item":
return fmt.Sprintf("- %s\n", extractRichTextMD(content))
case "numbered_list_item":
return fmt.Sprintf("1. %s\n", extractRichTextMD(content))
case "to_do":
checked := false
if c, ok := content["checked"].(bool); ok {
checked = c
}
checkbox := "[ ]"
if checked {
checkbox = "[x]"
}
return fmt.Sprintf("- %s %s\n", checkbox, extractRichTextMD(content))
// Quote & Callout
case "quote":
text := extractRichTextMD(content)
lines := strings.Split(text, "\n")
var quoted []string
for _, line := range lines {
quoted = append(quoted, "> "+line)
}
return strings.Join(quoted, "\n") + "\n\n"
case "callout":
icon := extractIcon(content)
text := extractRichTextMD(content)
if icon != "" {
return fmt.Sprintf("> %s %s\n\n", icon, text)
}
return fmt.Sprintf("> %s\n\n", text)
// Code
case "code":
lang := ""
if l, ok := content["language"].(string); ok {
lang = l
}
text := extractRichTextMD(content)
return fmt.Sprintf("```%s\n%s\n```\n\n", lang, text)
// Divider
case "divider":
return "---\n\n"
// Toggle
case "toggle":
text := extractRichTextMD(content)
hasChildren := false
if hc, ok := block["has_children"].(bool); ok {
hasChildren = hc
}
if hasChildren {
return fmt.Sprintf("<details>\n<summary>%s</summary>\n\n[nested content]\n</details>\n\n", text)
}
return fmt.Sprintf("<details>\n<summary>%s</summary>\n</details>\n\n", text)
// Image
case "image":
url := extractFileURL(content)
caption := extractCaption(content)
if caption != "" {
return fmt.Sprintf("\n\n", caption, url)
}
return fmt.Sprintf("\n\n", url)
// Bookmark & Link
case "bookmark":
url := ""
if u, ok := content["url"].(string); ok {
url = u
}
caption := extractCaption(content)
if caption != "" {
return fmt.Sprintf("[%s](%s)\n\n", caption, url)
}
return fmt.Sprintf("<%s>\n\n", url)
case "link_preview":
url := ""
if u, ok := content["url"].(string); ok {
url = u
}
return fmt.Sprintf("<%s>\n\n", url)
// Embed & Video
case "embed", "video":
url := extractFileURL(content)
return fmt.Sprintf("<%s>\n\n", url)
// File & PDF
case "file", "pdf":
url := extractFileURL(content)
caption := extractCaption(content)
if caption != "" {
return fmt.Sprintf("[%s](%s)\n\n", caption, url)
}
return fmt.Sprintf("[file](%s)\n\n", url)
// Table (simplified - just indicate table exists)
case "table":
return "[table]\n\n"
case "table_row":
cells := extractTableRow(content)
return "| " + strings.Join(cells, " | ") + " |\n"
// Child page/database
case "child_page":
title := ""
if t, ok := content["title"].(string); ok {
title = t
}
return fmt.Sprintf("**%s**\n\n", title)
case "child_database":
title := ""
if t, ok := content["title"].(string); ok {
title = t
}
return fmt.Sprintf("**%s**\n\n", title)
// Synced block
case "synced_block":
return "[synced block]\n\n"
// Column
case "column_list":
return "" // Column containers don't render
case "column":
return "" // Columns don't render directly
// Equation
case "equation":
expr := ""
if e, ok := content["expression"].(string); ok {
expr = e
}
return fmt.Sprintf("$$%s$$\n\n", expr)
// Breadcrumb & TOC
case "breadcrumb":
return "[breadcrumb]\n"
case "table_of_contents":
return "[table of contents]\n\n"
// Link to page
case "link_to_page":
return "[link to page]\n\n"
// Unsupported
case "unsupported":
return "[unsupported block]\n\n"
default:
return fmt.Sprintf("[%s]\n\n", blockType)
}
}
// =============================================================================
// Helper functions for Markdown extraction
// =============================================================================
// extractRichTextMD extracts rich text with Markdown formatting
func extractRichTextMD(content map[string]any) string {
arr, ok := content["rich_text"].([]any)
if !ok {
return ""
}
var parts []string
for _, item := range arr {
textObj, ok := item.(map[string]any)
if !ok {
continue
}
plainText := ""
if pt, ok := textObj["plain_text"].(string); ok {
plainText = pt
}
// Apply annotations
if annotations, ok := textObj["annotations"].(map[string]any); ok {
plainText = applyAnnotations(plainText, annotations)
}
// Handle links
if href, ok := textObj["href"].(string); ok && href != "" {
plainText = fmt.Sprintf("[%s](%s)", plainText, href)
}
parts = append(parts, plainText)
}
return strings.Join(parts, "")
}
// applyAnnotations applies Markdown formatting based on Notion annotations
func applyAnnotations(text string, annotations map[string]any) string {
if text == "" {
return text
}
// Bold
if bold, ok := annotations["bold"].(bool); ok && bold {
text = "**" + text + "**"
}
// Italic
if italic, ok := annotations["italic"].(bool); ok && italic {
text = "*" + text + "*"
}
// Strikethrough
if strike, ok := annotations["strikethrough"].(bool); ok && strike {
text = "~~" + text + "~~"
}
// Code
if code, ok := annotations["code"].(bool); ok && code {
text = "`" + text + "`"
}
// Underline (not standard MD, use HTML)
if underline, ok := annotations["underline"].(bool); ok && underline {
text = "<u>" + text + "</u>"
}
return text
}
// extractIcon extracts emoji or icon from callout
func extractIcon(content map[string]any) string {
icon, ok := content["icon"].(map[string]any)
if !ok {
return ""
}
if emoji, ok := icon["emoji"].(string); ok {
return emoji
}
return ""
}
// extractFileURL extracts URL from file/image block
func extractFileURL(content map[string]any) string {
// Try external URL first
if external, ok := content["external"].(map[string]any); ok {
if url, ok := external["url"].(string); ok {
return url
}
}
// Try file URL (uploaded to Notion)
if file, ok := content["file"].(map[string]any); ok {
if url, ok := file["url"].(string); ok {
return url
}
}
// Direct URL field
if url, ok := content["url"].(string); ok {
return url
}
return ""
}
// extractCaption extracts caption from block
func extractCaption(content map[string]any) string {
if caption, ok := content["caption"].([]any); ok {
return extractPlainText(caption)
}
return ""
}
// extractTableRow extracts cells from table_row block
func extractTableRow(content map[string]any) []string {
cells, ok := content["cells"].([]any)
if !ok {
return nil
}
var result []string
for _, cell := range cells {
if cellArr, ok := cell.([]any); ok {
result = append(result, extractPlainText(cellArr))
}
}
return result
}
