Add TMiR 2025-02 transcript, add script to automate transcript processing (#341)

* AI gen transcript processing script

Thanks Claude, seems to work

* Add TMiR 2025-02 transcript

Author: vcarl

package.json

@@ -11,7 +11,8 @@
     "lint": "next lint",
     "test": "next lint && prettier --check .",
     "prettier": "prettier --write .",
-    "install-if-package-changed": "git diff-tree -r --name-only --no-commit-id ORIG_HEAD HEAD | grep --quiet yarn.lock && yarn install || exit 0"
+    "install-if-package-changed": "git diff-tree -r --name-only --no-commit-id ORIG_HEAD HEAD | grep --quiet yarn.lock && yarn install || exit 0",
+    "process-tmir": "node --experimental-strip-types scripts/process-tmir.ts"
   },
   "husky": {
     "hooks": {

scripts/process-tmir.ts

@@ -0,0 +1,127 @@
+#!/usr/bin/env node
+
+/**
+ * Transcript Converter
+ *
+ * Transforms raw transcript markdown into a publishing-ready format with
+ * a table of contents and properly formatted timestamps.
+ *
+ * Usage:
+ * cat | npm run process-tmir | pbcopy
+ */
+
+// Read from stdin
+let rawTranscript = "";
+process.stdin.on("data", (chunk) => {
+  rawTranscript += chunk;
+});
+
+process.stdin.on("end", () => {
+  const processedTranscript = processTranscript(rawTranscript);
+  process.stdout.write(processedTranscript);
+});
+
+function processTranscript(input: string): string {
+  // Split the input into lines for processing
+  const lines = input.split("\n");
+
+  // Extract section headers for table of contents
+  const sectionHeaders = extractSectionHeaders(lines);
+  const tableOfContents = generateTableOfContents(sectionHeaders);
+
+  // Process the transcript lines
+  const processedLines = lines.map((line) => {
+    // Process section headers (remove timestamp from the header itself)
+    if (line.startsWith("## ")) {
+      return processSectionHeader(line);
+    }
+
+    // Process speaker lines and other content (move timestamp to end)
+    return processContentLine(line);
+  });
+
+  // Combine the table of contents with the processed transcript
+  return tableOfContents + "\n\n" + processedLines.join("\n");
+}
+
+interface SectionHeader {
+  timestamp: string;
+  title: string;
+  anchorLink: string;
+}
+
+function extractSectionHeaders(lines: string[]): SectionHeader[] {
+  const headers: SectionHeader[] = [];
+
+  lines.forEach((line) => {
+    if (line.startsWith("## ")) {
+      // Extract timestamp and title
+      const timestampMatch = line.match(/\[(\d{2}:\d{2}:\d{2})\]/);
+      if (!timestampMatch) return;
+
+      const timestamp = timestampMatch[1];
+
+      // Extract the title (everything after the timestamp)
+      let title = line.replace(/## \[\d{2}:\d{2}:\d{2}\] /, "");
+
+      // For the anchor link generation, we need to handle colons which are common in titles
+      // but need to be removed for proper anchor links
+
+      // Create an anchor link from the title
+      // Extract all text from the title, including text within markdown links
+      // First, replace all markdown links [text](url) with just the text
+      let plainTitle = title.replace(/\[(.*?)\]\(.*?\)/g, "$1");
+
+      // Convert to lowercase, remove special chars, and replace spaces with hyphens
+      let anchorLink = plainTitle
+        .toLowerCase()
+        .replace(/[^\w\s-]/g, "")
+        .replace(/\s+/g, "-");
+
+      headers.push({
+        timestamp,
+        title,
+        anchorLink,
+      });
+    }
+  });
+
+  return headers;
+}
+
+function generateTableOfContents(headers: SectionHeader[]): string {
+  return headers
+    .map(
+      (header) =>
+        `- [[${header.timestamp}](#${header.anchorLink})] ${header.title}`,
+    )
+    .join("\n");
+}
+
+function processSectionHeader(line: string): string {
+  // Extract timestamp
+  const timestampMatch = line.match(/\[(\d{2}:\d{2}:\d{2})\]/);
+  if (!timestampMatch) return line;
+
+  // Remove timestamp from the header
+  return line.replace(/\[\d{2}:\d{2}:\d{2}\] /, "");
+}
+
+function processContentLine(line: string): string {
+  // Skip empty lines or lines without timestamps
+  if (!line.trim() || !line.includes("[") || !line.includes("]")) {
+    return line;
+  }
+
+  // Extract timestamp
+  const timestampMatch = line.match(/\[(\d{2}:\d{2}:\d{2})\]/);
+  if (!timestampMatch) return line;
+
+  const timestamp = timestampMatch[0];
+
+  // Remove timestamp from the beginning
+  const contentWithoutTimestamp = line.replace(timestamp + " ", "");
+
+  // Add timestamp to the end
+  return contentWithoutTimestamp + " " + timestamp;
+}