Parse an Atom feed into Schema.org format. Args: root: XML root element feed_url: URL of the feed Returns: List of Schema.org formatted items
(root: ET.Element, feed_url: str | None = None)
| 379 | return result |
| 380 | |
| 381 | def parse_atom(root: ET.Element, feed_url: str | None = None) -> list[dict[str, Any]]: |
| 382 | """ |
| 383 | Parse an Atom feed into Schema.org format. |
| 384 | |
| 385 | Args: |
| 386 | root: XML root element |
| 387 | feed_url: URL of the feed |
| 388 | |
| 389 | Returns: |
| 390 | List of Schema.org formatted items |
| 391 | """ |
| 392 | result = [] |
| 393 | ns = {"atom": "http://www.w3.org/2005/Atom"} |
| 394 | |
| 395 | # Extract feed information |
| 396 | feed_title = safe_get_text(root.find('atom:title', ns)) |
| 397 | feed_subtitle = safe_get_text(root.find('atom:subtitle', ns)) |
| 398 | |
| 399 | # Extract feed link |
| 400 | feed_link = feed_url |
| 401 | for link in root.findall('atom:link', ns): |
| 402 | rel = link.get('rel', 'alternate') |
| 403 | if rel == 'self' and not feed_link: |
| 404 | feed_link = link.get('href') |
| 405 | elif rel == 'alternate': |
| 406 | feed_link = link.get('href') |
| 407 | break |
| 408 | |
| 409 | feed_link = fix_url(feed_link or "") |
| 410 | |
| 411 | # Create podcast series schema |
| 412 | podcast_series = { |
| 413 | "@type": "PodcastSeries", |
| 414 | "name": feed_title, |
| 415 | "description": feed_subtitle, |
| 416 | "url": feed_link |
| 417 | } |
| 418 | |
| 419 | # Process each entry |
| 420 | for entry in root.findall('atom:entry', ns): |
| 421 | try: |
| 422 | # Basic fields |
| 423 | title = safe_get_text(entry.find('atom:title', ns)) |
| 424 | summary = safe_get_text(entry.find('atom:summary', ns)) |
| 425 | published = safe_get_text(entry.find('atom:published', ns)) |
| 426 | updated = safe_get_text(entry.find('atom:updated', ns)) |
| 427 | |
| 428 | # Extract link (URL) |
| 429 | entry_url = None |
| 430 | for link in entry.findall('atom:link', ns): |
| 431 | rel = link.get('rel', 'alternate') |
| 432 | if rel == 'alternate': |
| 433 | entry_url = link.get('href') |
| 434 | break |
| 435 | |
| 436 | if not entry_url: |
| 437 | # Try any link |
| 438 | for link in entry.findall('atom:link', ns): |
no test coverage detected