fix premium videos length and view count scraping failure

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "yourtube"
-version = "0.5.3"
+version = "0.5.4"
 description = "Better youtube recommendations"
 authors = ["Filip Sondej <filipsondej@protonmail.com>"]
 license = "LGPL-2.1"

yourtube/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.5.3"
+__version__ = "0.5.4"
 
 import os
 import subprocess

yourtube/debug.ipynb

@@ -1,5 +1,17 @@
 {
  "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "try:\n",
+    "    %load_ext lab_black\n",
+    "except ModuleNotFoundError:\n",
+    "    print(\"nb_black not installed\")"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -13,7 +25,7 @@
     "sys.path.append(os.path.abspath(\"..\"))\n",
     "import matplotlib.pyplot as plt\n",
     "\n",
-    "from yourtube.scraping import get_content, scrape_content"
+    "from yourtube.scraping import *"
    ]
   },
   {

yourtube/file_operations.py

@@ -43,7 +43,7 @@
         it is absent if the video hasn't been watched
     view_count:
         number of views on youtube
-        can be None if the video is premium (so the views are hidden)
+        can be None if the video is premium (so the views are ambiguous)
     like_count:
         number of likes on youtube
         can be None if likes are disabled
@@ -54,6 +54,7 @@
         category of the video
     length:
         video length in seconds
+        can be None if the video is premium (so the views are ambiguous)
     keywords:
         keywords of the video as a list of strings
         can be an empty list if there are no keywords

yourtube/scraping.py

@@ -46,13 +46,12 @@ def get_title(content):
 def get_view_count(content):
     candidates = re.findall(r'"viewCount":"([0-9]+)"', content.text)
     candidates = set(candidates)
-    assert len(candidates) <= 1
-    if candidates:
-        view_count = candidates.pop()
-        return int(view_count)
-    else:
-        # view count is absent for example in premium videos
+    assert 1 <= len(candidates) <= 2
+    if len(candidates) == 2:
+        # premium videos list 2 different video versions
         return None
+    view_count = candidates.pop()
+    return int(view_count)
 
 
 def get_like_count(content):
@@ -99,7 +98,10 @@ def get_category(content):
 def get_length(content):
     candidates = re.findall(r'"videoDetails":.*?"lengthSeconds":"(.*?)"', content.text)
     candidates = set(candidates)
-    assert len(candidates) == 1
+    assert 1 <= len(candidates) <= 2
+    if len(candidates) == 2:
+        # premium videos list 2 different lengths
+        return None
     length = candidates.pop()
     return int(length)