typo

Signed-off-by: ugrnm <ultrageranium@bleu255.com>

README

@@ -15,6 +15,7 @@ There are two scripts:
 * download_loooooops.py
   * gets latest toots tagged with #looptoper
   * download all the attachment if new
+  * transcodes to opus
   * generates new playlist and tells ezstream to reload it
 
 
@@ -22,7 +23,7 @@ USAGE
 -----
 
 Quick and dirty, two tmux panes:
-* `while true; do ./download_loooooops.py && sleep 1d; done`
+* `while true; do ./download_loooooops.py </dev/null && sleep 1d; done`
 * `stream_loooooops.sh`
 
 You can start whichever you want firt.
@@ -31,7 +32,7 @@ You can start whichever you want firt.
 AGPL + DONATIONWARE
 -------------------
 
-devi is dual licensed AGPL and donationware.
+radio-looptober is dual licensed AGPL and donationware.
 If you find radio-looptober useful and can afford it, a donation to LURK is
 greatly appreciated :)
 

TODO

@@ -2,9 +2,7 @@ TODO
 ====
 
 * cache the json parsing/process locally for faster restart tests
-* only download MP3 files
-* extract relevant json stuff to fill in or update ID3 tags
 * 80c :)
-* convert the downloaded mp3 files to opus
 * get rid of request, just use urlib or something
 * build the loop inside the download script
+* prune media if original toot was deleted

download_loooooops.py

@@ -9,10 +9,31 @@ from urllib.parse import urlparse
 #   remote_url
 #   description
 
-output_dir = "loooooops"
+year = "2022"
+
+output_dir = os.path.join("loooooops", year)
+
+bitrate = "128k"
+
+def transcode_media(path, media_item, metadata):
+    infile = os.path.join(path, media_item)
+    outfile = os.path.join(path, media_item + ".opus")
+    if not os.path.exists(outfile):
+        print("transcodeing to {}".format(outfile))
+        pid = os.fork()
+        if pid == 0:
+            artist = metadata["creator"]
+            title = metadata["url"]
+            comment = metadata["description"]
+            date = metadata["date"]
+            os.execlp("ffmpeg", "ffmpeg", "-hide_banner", "-loglevel", "error", "-i", infile, "-map_metadata", "-1", "-metadata", "artist={}".format(artist), "-metadata", "title={}".format(title), "-metadata", "creation_time={}".format(date), "-map_chapters", "-1", "-ac", "2", "-af", "loudnorm=dual_mono=true", "-b:a", bitrate, "-y", outfile)
+	    # never reached
+        else:
+            os.wait()
 
 def grab_media(path, url):
 
+  try:
     media_item = urlparse(url).path.split('/')[-1]
 
     headers = {
@@ -20,13 +41,19 @@ def grab_media(path, url):
     'From': 'post.lurk.org/@lurk'  # This is another valid field
     } 
 
-    if not os.path.exists(os.path.join(path, media_item)):
+    if os.path.exists(os.path.join(path, media_item)):
+        return media_item
+    else:
         response = requests.get(url, headers=headers, stream=True)
         if response.ok:
             with open(os.path.join(path, media_item), 'wb') as media_file:
                 shutil.copyfileobj(response.raw, media_file)
                 print('Downloaded media {} from {}'.format(media_item, urlparse(url).netloc))
                 return media_item
+  except requests.exceptions.ConnectionError as e:
+    # maybe transient network issues
+    print(e)
+    sleep(60)
 
 #This pages through all the looptober tag and collects the json in 'data'
 there_is_more = True
@@ -63,7 +90,7 @@ for collection in data:
                 i['created_at'][:-1]).astimezone(
                 datetime.timezone.utc)
 
-            if creation_date.strftime('%Y') == "2022": #we only take entries from this year
+            if creation_date.strftime('%Y') == year: #we only take entries from this year
                 stuff = {}
                 stuff["url"] = i["url"]
                 stuff["description"] = i["content"]
@@ -76,13 +103,9 @@ for collection in data:
                     i["account"]["username"],
                     len(i["media_attachments"])))
 
-if not os.path.exists(output_dir):
-    os.mkdir(output_dir)
-
 for l in looooops:
     path = os.path.join(output_dir,"{}_{}".format(l['creator'], l['id']))
-    if not os.path.exists(path):
-        os.mkdir(path)
+    os.makedirs(path, exist_ok=True)
 
     print("\n")
     print("Downloading looops by ***{}***".format(l['creator']))
@@ -92,10 +115,13 @@ for l in looooops:
         else: 
             url = a['url']
 
-        grab_media(path, url)
+        media_item = grab_media(path, url)
+        if media_item:
+            transcode_media(path, media_item, l)
 
 # Once we've done everythin we generate a playlist and ask ezstream
 # to reload it
-os.system('find . -iname "*mp3" > playlist_loooooops.m3u'\
-    '&& kill -s HUP `pidof ezstream`')
-
+# this is not an injection vulnerability as output_dir is under
+# our control
+os.system('find {} -iname "*opus" > playlist_loooooops.m3u'\
+    '&& kill -s HUP `pidof ezstream`'.format(output_dir))