import type {
ProseMirrorDoc,
ProseMirrorNode,
TranscriptUtterance,
} from "./api.js";
export function proseMirrorToMarkdown(doc: ProseMirrorDoc | undefined): string {
if (!doc || !doc.content) {
return "";
}
return processNodes(doc.content);
}
function processNodes(nodes: ProseMirrorNode[]): string {
return nodes.map(processNode).join("");
}
function processNode(node: ProseMirrorNode): string {
switch (node.type) {
case "heading": {
const level = (node.attrs?.level as number) ?? 1;
const text = node.content ? processNodes(node.content) : "";
return `${"#".repeat(level)} ${text}\n\n`;
}
case "paragraph": {
const text = node.content ? processNodes(node.content) : "";
return `${text}\n\n`;
}
case "bulletList": {
if (!node.content) return "";
const items = node.content
.filter((item) => item.type === "listItem")
.map((item) => {
const text = item.content ? processNodes(item.content).trim() : "";
return `- ${text}`;
});
return items.join("\n") + "\n\n";
}
case "orderedList": {
if (!node.content) return "";
const items = node.content
.filter((item) => item.type === "listItem")
.map((item, index) => {
const text = item.content ? processNodes(item.content).trim() : "";
return `${index + 1}. ${text}`;
});
return items.join("\n") + "\n\n";
}
case "listItem": {
return node.content ? processNodes(node.content) : "";
}
case "codeBlock": {
const lang = (node.attrs?.language as string) ?? "";
const code = node.content ? processNodes(node.content) : "";
return `\`\`\`${lang}\n${code}\`\`\`\n\n`;
}
case "blockquote": {
const text = node.content ? processNodes(node.content) : "";
return text
.split("\n")
.map((line) => `> ${line}`)
.join("\n") + "\n\n";
}
case "horizontalRule":
return "---\n\n";
case "hardBreak":
return "\n";
case "text": {
let text = node.text ?? "";
// Handle marks (bold, italic, etc.) if present
const marks = (node as unknown as { marks?: { type: string }[] }).marks;
if (marks) {
for (const mark of marks) {
switch (mark.type) {
case "bold":
case "strong":
text = `**${text}**`;
break;
case "italic":
case "em":
text = `*${text}*`;
break;
case "code":
text = `\`${text}\``;
break;
case "strike":
text = `~~${text}~~`;
break;
}
}
}
return text;
}
default:
// Recursively process unknown nodes that might have content
if (node.content) {
return processNodes(node.content);
}
return "";
}
}
export function transcriptToMarkdown(
utterances: TranscriptUtterance[]
): string {
if (!utterances || utterances.length === 0) {
return "";
}
const lines: string[] = ["# Transcript\n"];
for (const utterance of utterances) {
const time = new Date(utterance.start_timestamp).toLocaleTimeString();
const source = utterance.source === "microphone" ? "🎤" : "💻";
lines.push(`**[${time}]** ${source} ${utterance.text}\n`);
}
return lines.join("\n");
}
export function transcriptToPlainText(
utterances: TranscriptUtterance[]
): string {
if (!utterances || utterances.length === 0) {
return "";
}
return utterances.map((u) => u.text).join(" ");
}
/**
* Converts transcript to speaker-attributed format for AI processing.
* Uses [ME] for microphone (the user) and [PARTICIPANT] for system audio (others).
*/
export function transcriptWithSpeakers(
utterances: TranscriptUtterance[]
): string {
if (!utterances || utterances.length === 0) {
return "";
}
const lines: string[] = [];
let currentSpeaker: string | null = null;
let currentTexts: string[] = [];
// Merge consecutive utterances from the same speaker
for (const utterance of utterances) {
const speaker = utterance.source === "microphone" ? "[ME]" : "[PARTICIPANT]";
if (speaker === currentSpeaker) {
currentTexts.push(utterance.text);
} else {
// Flush previous speaker's text
if (currentSpeaker && currentTexts.length > 0) {
lines.push(`${currentSpeaker} ${currentTexts.join(" ")}`);
}
currentSpeaker = speaker;
currentTexts = [utterance.text];
}
}
// Flush final speaker's text
if (currentSpeaker && currentTexts.length > 0) {
lines.push(`${currentSpeaker} ${currentTexts.join(" ")}`);
}
return lines.join("\n\n");
}
