From 2e8ab33259c285f71625fff546cf1fd3022f959d Mon Sep 17 00:00:00 2001
From: pluja <github_pluja@r3d.red>
Date: Wed, 7 Oct 2020 16:22:16 +0200
Subject: [PATCH] Adds support for Youtube comments

---
 app/routes.py                     |  14 +++-
 app/templates/_video_comment.html |  25 ++++++
 app/templates/video.html          |  13 +++-
 youtube_data/comments.py          | 123 ++++++++++++++++++++++++++++++
 4 files changed, 172 insertions(+), 3 deletions(-)
 create mode 100644 app/templates/_video_comment.html
 create mode 100644 youtube_data/comments.py

diff --git a/app/routes.py b/app/routes.py
index 42106d6..3d503ec 100644
--- a/app/routes.py
+++ b/app/routes.py
@@ -30,6 +30,7 @@ import os
 #########################################
 from youtube_data import videos as ytvids
 from youtube_data import search as yts
+from youtube_data import comments as ytcomments
 #########################################
 
 cache = Cache(config={'CACHE_TYPE': 'simple'})
@@ -437,6 +438,16 @@ def watch():
     else:
         vid_urls = get_best_urls(info['video']['urls'])
 
+    # Get comments
+    try:
+        comments = ytcomments.video_comments(id)
+        if comments:
+            comments.sort(key=lambda x: x['likes'], reverse=True)
+        else:
+            comments = False
+    except:
+        comments = False
+
     video={
         'title':info['video']['title'],
         'description':Markup(markupString(info['video']['description'])),
@@ -451,7 +462,8 @@ def watch():
         'isUpcoming': info['video']['isUpcoming'],
         'thumbnail': info['video']['thumbnail'],
         'nginxAudioUrl': audioUrl,
-        'premieres': info['video']['premieres']
+        'premieres': info['video']['premieres'],
+        'comments': comments
     }
     return render_template("video.html", video=video, title='{}'.format(video['title']), config=config, urls=vid_urls)
 
diff --git a/app/templates/_video_comment.html b/app/templates/_video_comment.html
new file mode 100644
index 0000000..ee10a7a
--- /dev/null
+++ b/app/templates/_video_comment.html
@@ -0,0 +1,25 @@
+<div class="comment">
+    <a class="avatar" style="width: 32px; height: 32px;"><img src="{{ comment.thumbnail }}"></a>
+    <div class="content">
+        {% if comment.authorIsChannelOwner %}
+        
+        <a class="author" href="{{comment.channel}}"><i class="red user circle icon"></i>{{comment.author}}</a>
+        {% else %}
+        <a class="author" href="{{comment.channel}}">{{comment.author}}</a>
+        {% endif %}
+    <div class="metadata">
+        <div class="date">{{comment.date}}</div>
+        <div class="rating">
+            <i class="thumbs up icon"></i>
+            {{comment.likes}}
+        </div>
+        <div class="rating">
+            <i class="comment icon"></i>
+            {{comment.replies}}
+        </div>
+    </div>
+    <div class="text">
+        {{comment.text}}
+    </div>
+    </div>
+</div>
\ No newline at end of file
diff --git a/app/templates/video.html b/app/templates/video.html
index 9294895..228c27f 100644
--- a/app/templates/video.html
+++ b/app/templates/video.html
@@ -88,8 +88,17 @@
         <div class="ui raised segment break-word">
             <p>{{video.description}}</p>
         </div>
-      </div>
-</div>
+    </div>
+
+    {% if comments != False %}
+    <div class="ui comments">
+        <h3 class="ui dividing header">Comments</h3>
+        {% for comment in video.comments %}
+                {% include '_video_comment.html' %}
+        {% endfor %}
+    </div>
+    {%endif%}
+            
 <script src="{{ url_for('static',filename='video.min.js') }}"></script>
 <script src="{{ url_for('static',filename='videojs-http-streaming.min.js')}}"></script>
 <script>
diff --git a/youtube_data/comments.py b/youtube_data/comments.py
new file mode 100644
index 0000000..0b3d07b
--- /dev/null
+++ b/youtube_data/comments.py
@@ -0,0 +1,123 @@
+from youtube_data import proto
+import json
+import base64
+import urllib
+import requests
+import re
+import bleach
+from flask import Markup
+
+URL_ORIGIN = "/https://www.youtube.com"
+
+def make_comment_ctoken(video_id, sort=0, offset=0, lc='', secret_key=''):
+    video_id = proto.as_bytes(video_id)
+    secret_key = proto.as_bytes(secret_key)
+    
+
+    page_info = proto.string(4,video_id) + proto.uint(6, sort)
+    offset_information = proto.nested(4, page_info) + proto.uint(5, offset)
+    if secret_key:
+        offset_information = proto.string(1, secret_key) + offset_information
+
+    page_params = proto.string(2, video_id)
+    if lc:
+        page_params += proto.string(6, proto.percent_b64encode(proto.string(15, lc)))
+
+    result = proto.nested(2, page_params) + proto.uint(3,6) + proto.nested(6, offset_information)
+    return base64.urlsafe_b64encode(result).decode('ascii')
+
+def comment_replies_ctoken(video_id, comment_id, max_results=500):  
+
+    params = proto.string(2, comment_id) + proto.uint(9, max_results)
+    params = proto.nested(3, params)
+    
+    result = proto.nested(2, proto.string(2, video_id)) + proto.uint(3,6) + proto.nested(6, params)
+    return base64.urlsafe_b64encode(result).decode('ascii')
+
+
+
+mobile_headers = {
+    'User-Agent': 'Mozilla/5.0 (iPhone; CPU iPhone OS 10_3_1 like Mac OS X) AppleWebKit/603.1.30 (KHTML, like Gecko) Version/10.0 Mobile/14E304 Safari/602.1',
+    'Accept': '*/*',
+    'Accept-Language': 'en-US,en;q=0.5',
+    'X-YouTube-Client-Name': '2',
+    'X-YouTube-Client-Version': '2.20180823',
+}
+def request_comments(ctoken, replies=False):
+    if replies: # let's make it use different urls for no reason despite all the data being encoded
+        base_url = "https://m.youtube.com/watch_comment?action_get_comment_replies=1&ctoken="
+    else:
+        base_url = "https://m.youtube.com/watch_comment?action_get_comments=1&ctoken="
+    url = base_url + ctoken.replace("=", "%3D") + "&pbj=1"
+
+    for i in range(0,8):    # don't retry more than 8 times
+        content = requests.get(url, headers=mobile_headers).text
+        if content[0:4] == b")]}'":             # random closing characters included at beginning of response for some reason
+            content = content[4:]
+        elif content[0:10] == b'\n<!DOCTYPE':   # occasionally returns html instead of json for no reason
+            content = b''
+            print("got <!DOCTYPE>, retrying")
+            continue
+        break
+
+    polymer_json = json.loads(content)
+    return polymer_json
+
+
+def single_comment_ctoken(video_id, comment_id):
+    page_params = proto.string(2, video_id) + proto.string(6, proto.percent_b64encode(proto.string(15, comment_id)))
+
+    result = proto.nested(2, page_params) + proto.uint(3,6)
+    return base64.urlsafe_b64encode(result).decode('ascii')
+
+
+def concat_texts(strings):
+    '''Concatenates strings. Returns None if any of the arguments are None'''
+    result = ''
+    for string in strings:
+        if string['text'] is None:
+            return None
+        result += string['text']
+    return result
+
+def parse_comment(raw_comment):
+    cmnt = {}
+    raw_comment = raw_comment['commentThreadRenderer']['comment']['commentRenderer']
+    cmnt['author'] = raw_comment['authorText']['runs'][0]['text']
+    cmnt['thumbnail'] = raw_comment['authorThumbnail']['thumbnails'][0]['url']
+    cmnt['channel'] = raw_comment['authorEndpoint']['commandMetadata']['webCommandMetadata']['url']
+    cmnt['text'] = Markup(bleach.linkify(concat_texts(raw_comment['contentText']['runs']).replace("\n", "<br>")))
+    cmnt['date'] = raw_comment['publishedTimeText']['runs'][0]['text']
+    
+    try:
+        cmnt['likes'] = raw_comment['likeCount']
+    except:
+        cmnt['likes'] = 0
+
+    try:
+        cmnt['replies'] = raw_comment['replyCount']
+    except:
+        cmnt['replies'] = 0
+
+    cmnt['authorIsChannelOwner'] = raw_comment['authorIsChannelOwner']
+    try:
+        cmnt['pinned'] = raw_comment['pinnedCommentBadge']
+        cmnt['pinned'] = True
+    except:
+        cmnt['pinned'] = False
+    return cmnt
+
+def post_process_comments_info(comments_info):
+    comments = []
+    for comment in comments_info[1]['response']['continuationContents']['commentSectionContinuation']['items']:        
+        comments.append(parse_comment(comment))
+    return comments
+        
+
+
+def video_comments(video_id, sort=0, offset=0, lc='', secret_key=''):
+    comments_info = request_comments(make_comment_ctoken(video_id, sort, offset, lc, secret_key))
+    comments_info = post_process_comments_info(comments_info)
+    return comments_info
+    return {}
+