adf.parser.ts•9.13 kB
/**
* ADF (Atlassian Document Format) Parser
*
* Converts JIRA's complex ADF object structures to readable markdown text.
* Handles backward compatibility with string descriptions.
*/
/**
* ADF Node structure representing Atlassian Document Format nodes
*/
export interface ADFNode {
type: string;
content?: ADFNode[];
text?: string;
attrs?: Record<string, unknown>;
marks?: Array<{ type: string; attrs?: Record<string, unknown> }>;
}
/**
* ADF Document structure (top-level document with version)
*/
export interface ADFDocument extends ADFNode {
type: "doc";
version: number;
content: ADFNode[];
}
/**
* Text mark for formatting (bold, italic, code, etc.)
*/
export interface ADFMark {
type: string;
attrs?: Record<string, unknown>;
}
/**
* Type guard to check if value is a string
*/
function isString(value: unknown): value is string {
return typeof value === "string";
}
/**
* Type guard to check if value is null or undefined
*/
function isNullish(value: unknown): value is null | undefined {
return value === null || value === undefined;
}
/**
* Type guard to check if value is an object (not null)
*/
function isObject(value: unknown): value is Record<string, unknown> {
return typeof value === "object" && value !== null;
}
/**
* Type guard to check if value is an ADF node
*/
function isADFNode(value: unknown): value is ADFNode {
return isObject(value) && "type" in value && isString(value.type);
}
/**
* Type guard to check if value is an ADF document
*/
function isADFDocument(value: unknown): value is ADFDocument {
return (
isADFNode(value) &&
value.type === "doc" &&
"version" in value &&
typeof value.version === "number" &&
"content" in value &&
Array.isArray(value.content)
);
}
/**
* Type guard to check if value is a non-document ADF node
*/
function isNonDocumentADFNode(value: unknown): value is ADFNode {
return isADFNode(value) && value.type !== "doc";
}
/**
* Type guard to check if string is empty or whitespace only
*/
function isEmptyString(value: string): boolean {
return value.trim() === "";
}
/**
* Converts ADF objects to markdown format preserving structure and formatting
*/
export class ADFToMarkdownParser {
/**
* Parse ADF content to markdown
* @param adf - ADF object or string (for backward compatibility)
* @returns Formatted markdown string
*/
parse(adf: ADFNode | string | null | undefined): string {
// Handle backward compatibility with string descriptions
if (isString(adf)) return adf;
if (isNullish(adf)) return "";
return this.parseNode(adf);
}
/**
* Parse a single ADF node
*/
private parseNode(node: ADFNode): string {
switch (node.type) {
case "doc":
return this.parseContent(node.content || []);
case "paragraph":
return `${this.parseContent(node.content || [])}\n\n`;
case "text":
return this.formatText(node.text || "", node.marks || []);
case "codeBlock":
return this.parseCodeBlock(node);
case "bulletList":
return this.parseBulletList(node);
case "orderedList":
return this.parseOrderedList(node);
case "listItem":
return this.parseListItem(node);
case "heading":
return this.parseHeading(node);
case "blockquote":
return this.parseBlockquote(node);
case "hardBreak":
return "\n";
case "rule":
return "\n---\n\n";
default:
// For unknown node types, try to parse content if available
return this.parseContent(node.content || []);
}
}
/**
* Parse array of content nodes
*/
private parseContent(content: ADFNode[]): string {
return content.map((node) => this.parseNode(node)).join("");
}
/**
* Format text with marks (bold, italic, code, etc.)
*/
private formatText(text: string, marks: ADFMark[]): string {
if (!marks || marks.length === 0) return text;
let formattedText = text;
// Apply marks in order
for (const mark of marks) {
switch (mark.type) {
case "strong":
formattedText = `**${formattedText}**`;
break;
case "em":
formattedText = `*${formattedText}*`;
break;
case "code":
formattedText = `\`${formattedText}\``;
break;
case "strike":
formattedText = `~~${formattedText}~~`;
break;
case "link": {
const href = mark.attrs?.href;
if (href && isString(href)) {
formattedText = `[${formattedText}](${href})`;
}
break;
}
// Add more mark types as needed
default:
// Unknown mark type, keep text as-is
break;
}
}
return formattedText;
}
/**
* Parse code block with language support
*/
private parseCodeBlock(node: ADFNode): string {
const language = node.attrs?.language || "";
const content = this.parseContent(node.content || []);
return `\`\`\`${language}\n${content}\n\`\`\`\n\n`;
}
/**
* Parse bullet list
*/
private parseBulletList(node: ADFNode): string {
const items = (node.content || [])
.map((item) => this.parseListItem(item, "- "))
.join("");
return `${items}\n`;
}
/**
* Parse ordered list
*/
private parseOrderedList(node: ADFNode): string {
const items = (node.content || [])
.map((item, index) => this.parseListItem(item, `${index + 1}. `))
.join("");
return `${items}\n`;
}
/**
* Parse list item with prefix
*/
private parseListItem(node: ADFNode, prefix = "- "): string {
const content = this.parseContent(node.content || []);
return `${prefix}${content.trim()}\n\n`;
}
/**
* Parse heading with level support
*/
private parseHeading(node: ADFNode): string {
const level = node.attrs?.level || 1;
const hashes = "#".repeat(Math.min(level as number, 6));
const content = this.parseContent(node.content || []);
return `${hashes} ${content}\n\n`;
}
/**
* Parse blockquote
*/
private parseBlockquote(node: ADFNode): string {
const content = this.parseContent(node.content || []);
const lines = content.split("\n");
const quotedLines = lines.map((line) => `> ${line}`).join("\n");
return `${quotedLines}\n\n`;
}
/**
* Extract plain text only (alternative format for simple use cases)
*/
extractPlainText(adf: ADFNode | string | null | undefined): string {
if (isString(adf)) return adf;
if (isNullish(adf)) return "";
return this.extractTextFromNode(adf);
}
/**
* Recursively extract text from ADF node
*/
private extractTextFromNode(node: ADFNode): string {
if (node.type === "text") {
return node.text || "";
}
if (node.content) {
return node.content
.map((childNode) => this.extractTextFromNode(childNode))
.join("");
}
return "";
}
}
/**
* Default parser instance for convenience
*/
export const adfParser = new ADFToMarkdownParser();
/**
* Convenience function for parsing ADF to markdown
* @param adf - ADF object or string
* @returns Formatted markdown string
*/
export function parseADF(adf: ADFNode | string | null | undefined): string {
return adfParser.parse(adf);
}
/**
* Convenience function for extracting plain text from ADF
* @param adf - ADF object or string
* @returns Plain text string
*/
export function extractTextFromADF(
adf: ADFNode | string | null | undefined,
): string {
return adfParser.extractPlainText(adf);
}
/**
* Convert plain text to ADF document format
* @param text - Plain text string
* @returns ADF document structure
*/
export function textToADF(text: string | null | undefined): ADFDocument | null {
if (isNullish(text) || isEmptyString(text)) {
return null;
}
// Split text into paragraphs (by double newlines or single newlines)
const paragraphs = text
.split(/\n\s*\n/)
.map((p) => p.trim())
.filter((p) => p.length > 0);
if (paragraphs.length === 0) {
return null;
}
const content: ADFNode[] = paragraphs.map((paragraph) => ({
type: "paragraph",
content: [
{
type: "text",
text: paragraph,
},
],
}));
return {
version: 1,
type: "doc",
content,
};
}
/**
* Convert text to ADF or return existing ADF if already in correct format
* @param input - Text string or existing ADF document/node
* @returns ADF document structure or null
*/
export function ensureADFFormat(
input: string | ADFDocument | ADFNode | null | undefined,
): ADFDocument | null {
if (isNullish(input)) {
return null;
}
// If it's already an ADF document, return as-is
if (isADFDocument(input)) {
return input;
}
// If it's an ADF node but not a document, wrap it in a document
if (isNonDocumentADFNode(input)) {
return {
version: 1,
type: "doc",
content: [input],
};
}
// If it's a string, convert to ADF
if (isString(input)) {
return textToADF(input);
}
return null;
}