boyska/larigira
1
0
Fork 0
Code Issues 13 Pull requests Releases Activity

podcat: tolerate items without enclosures

Browse source
This commit is contained in:
boyska 2021-04-21 01:12:50 +02:00
parent e1729fed05
commit 924615b282
1 changed files with 11 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

14
larigira/audiogen_podcast.py
View file

@ -115,7 +115,10 @@ def get_item_date(el):
def get_audio_from_item(item): def get_audio_from_item(item):
encl = item.find("enclosure") encl = item.find("enclosure")
url = encl.get("url") if encl is not None:
url = encl.get("url")
else:
return None
audio_args = {} audio_args = {}
if item.find("duration") is not None: if item.find("duration") is not None:
duration_parts = item.findtext("duration").split(":") duration_parts = item.findtext("duration").split(":")
@ -125,8 +128,11 @@ def get_audio_from_item(item):
if total_seconds: if total_seconds:
audio_args["duration"] = total_seconds audio_args["duration"] = total_seconds
else: else:
for child in item.xpath("group/content"): contents = item.xpath("group/content")
if child.get("url") == url: if not contents:
contents = item.xpath("content")
for child in contents:
if child.get("url") == url and child.get("duration") is not None:
audio_args["duration"] = int(float(child.get("duration"))) audio_args["duration"] = int(float(child.get("duration")))
break break
return Audio(url, **audio_args) return Audio(url, **audio_args)
@ -137,6 +143,8 @@ def get_urls(tree):
for it in items: for it in items:
# title = it.find("title").text # title = it.find("title").text
audio = get_audio_from_item(it) audio = get_audio_from_item(it)
if audio is None:
continue
if audio.date is None: if audio.date is None:
audio.date = get_item_date(it) audio.date = get_item_date(it)
yield audio yield audio