Improved Youtube search
This commit is contained in:
pluja
parent
1e4c4c82e3
commit
a3f4d89781
@ -7,7 +7,6 @@ from requests_futures.sessions import FuturesSession
|
||||
from werkzeug.datastructures import Headers
|
||||
from concurrent.futures import as_completed
|
||||
from werkzeug.utils import secure_filename
|
||||
from youtube_data import videos as ytvids
|
||||
from youtube_search import YoutubeSearch
|
||||
from werkzeug.urls import url_parse
|
||||
from youtube_dl import YoutubeDL
|
||||
@ -24,6 +23,10 @@ import bleach
|
||||
import urllib
|
||||
import json
|
||||
import re
|
||||
#########################################
|
||||
from youtube_data import videos as ytvids
|
||||
from youtube_data import search as yts
|
||||
#########################################
|
||||
##########################
|
||||
#### Config variables ####
|
||||
##########################
|
||||
@ -225,32 +228,14 @@ def ytsearch():
|
||||
channels = []
|
||||
videos = []
|
||||
|
||||
searchTerm = form.channelId.data
|
||||
search = YoutubeSearch(searchTerm)
|
||||
chnns = search.channels_to_dict()
|
||||
vids = search.videos_to_dict()
|
||||
searchTerms = form.channelId.data
|
||||
page = 1
|
||||
autocorrect = 1
|
||||
sort = 0
|
||||
filters = {"time":0, "type":0, "duration":0}
|
||||
results = yts.search_by_terms(searchTerms, page, autocorrect, sort, filters)
|
||||
|
||||
for v in vids:
|
||||
videos.append({
|
||||
'channelName':v['channel'],
|
||||
'videoTitle':v['title'],
|
||||
'description':Markup(v['long_desc']),
|
||||
'id':v['id'],
|
||||
'videoThumb': v['thumbnails'][-1],
|
||||
'channelUrl':v['url_suffix'],
|
||||
'channelId': v['channelId'],
|
||||
'views':v['views'],
|
||||
'timeStamp':v['publishedText']
|
||||
})
|
||||
|
||||
for c in chnns:
|
||||
channels.append({
|
||||
'username':c['name'],
|
||||
'channelId':c['id'],
|
||||
'thumbnail':'https:{}'.format(c['thumbnails'][0]),
|
||||
'subCount':c['suscriberCountText'].split(" ")[0]
|
||||
})
|
||||
return render_template('ytsearch.html', form=form, btform=button_form, channels=channels, videos=videos, restricted=config['restrictPublicUsage'], config=config)
|
||||
return render_template('ytsearch.html', form=form, btform=button_form, results=results, restricted=config['restrictPublicUsage'], config=config)
|
||||
|
||||
else:
|
||||
return render_template('ytsearch.html', form=form)
|
||||
|
||||
@ -16,7 +16,7 @@
|
||||
</div>
|
||||
</div>
|
||||
<div class="extra content">
|
||||
{% if video.views == "Livestream" %}
|
||||
{% if video.isLive == "Livestream" or video.isLive %}
|
||||
<span class="right floated">
|
||||
<i class="red circle icon"></i>
|
||||
{{video.views}}
|
||||
@ -28,7 +28,7 @@
|
||||
</span>
|
||||
{% endif %}
|
||||
|
||||
{% if video.timeStamp == "Scheduled" %}
|
||||
{% if video.timeStamp == "Scheduled" or video.isUpcoming %}
|
||||
<span class="right floated">
|
||||
<i class="blue clock icon"></i>
|
||||
{{video.timeStamp}}
|
||||
|
||||
@ -14,37 +14,48 @@
|
||||
<p>{{ form.submit() }}</p>
|
||||
</form>
|
||||
|
||||
{% if channels %}
|
||||
{% if results.channels %}
|
||||
<h3 class="ui dividing header">Users</h3>
|
||||
<div class="ui relaxed divided list">
|
||||
|
||||
{% for res in channels %}
|
||||
{% for res in results.channels %}
|
||||
<div class="item">
|
||||
{% if restricted or current_user.is_authenticated %}
|
||||
<div class="right floated content">
|
||||
{% if not current_user.is_following_yt(res.channelId) %}
|
||||
<p>
|
||||
<div class="image">
|
||||
<img src="{{ url_for('img', url=res.thumbnail) }}">
|
||||
</div>
|
||||
<div class="content">
|
||||
<a class = "header" href="{{ url_for('channel', id=res.channelId)}}">{{res.username}}</a>
|
||||
<div class="meta">
|
||||
<span>{{res.description}}</span>
|
||||
</div>
|
||||
<div class="description">
|
||||
<p></p>
|
||||
</div>
|
||||
<div class="extra">
|
||||
<div class="ui label">
|
||||
<i class="user icon"></i> {{res.suscribers}}
|
||||
</div>
|
||||
|
||||
<div class="ui label">
|
||||
<i class="video icon"></i> {{res.videos}}
|
||||
</div>
|
||||
|
||||
{% if restricted or current_user.is_authenticated %}
|
||||
<div class="right floated content">
|
||||
{% if not current_user.is_following_yt(res.channelId) %}
|
||||
<form action="{{ url_for('ytfollow', channelId=res.channelId) }}" method="post">
|
||||
{{ btform.hidden_tag() }}
|
||||
{{ btform.submit(value='Follow') }}
|
||||
</form>
|
||||
</p>
|
||||
{% else %}
|
||||
<p>
|
||||
{% else %}
|
||||
<form action="{{ url_for('ytunfollow', channelId=res.channelId) }}" method="post">
|
||||
{{ btform.hidden_tag() }}
|
||||
{{ btform.submit(value='Unfollow') }}
|
||||
</form>
|
||||
</p>
|
||||
{% endif %}
|
||||
</div>
|
||||
{% endif %}
|
||||
</div>
|
||||
{% endif %}
|
||||
<img alt="Avatar" class="ui avatar image" src="{{ res.thumbnail }}">
|
||||
<div class="content">
|
||||
<a class = "header" href="{{ url_for('channel', id=res.channelId)}}">{{res.username}}</a>
|
||||
<div class="description"><div class="ui label">
|
||||
<i class="user icon"></i> {{res.subCount}}
|
||||
</div></div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
{% endfor %}
|
||||
@ -53,10 +64,10 @@
|
||||
|
||||
|
||||
<div class="ui middle aligned divided list">
|
||||
{% if videos %}
|
||||
{% if results.videos %}
|
||||
<h3 class="ui dividing header">Videos</h3>
|
||||
<div class="ui centered cards">
|
||||
{% for video in videos %}
|
||||
{% for video in results.videos %}
|
||||
{% include '_video_item.html' %}
|
||||
{% endfor %}
|
||||
</div>
|
||||
|
||||
130
youtube_data/proto.py
Normal file
130
youtube_data/proto.py
Normal file
@ -0,0 +1,130 @@
|
||||
from math import ceil
|
||||
import base64
|
||||
import io
|
||||
|
||||
# FROM https://github.com/user234683/youtube-local/blob/master/youtube/proto.py
|
||||
|
||||
def byte(n):
|
||||
return bytes((n,))
|
||||
|
||||
|
||||
def varint_encode(offset):
|
||||
'''In this encoding system, for each 8-bit byte, the first bit is 1 if there are more bytes, and 0 is this is the last one.
|
||||
The next 7 bits are data. These 7-bit sections represent the data in Little endian order. For example, suppose the data is
|
||||
aaaaaaabbbbbbbccccccc (each of these sections is 7 bits). It will be encoded as:
|
||||
1ccccccc 1bbbbbbb 0aaaaaaa
|
||||
|
||||
This encoding is used in youtube parameters to encode offsets and to encode the length for length-prefixed data.
|
||||
See https://developers.google.com/protocol-buffers/docs/encoding#varints for more info.'''
|
||||
needed_bytes = ceil(offset.bit_length()/7) or 1 # (0).bit_length() returns 0, but we need 1 in that case.
|
||||
encoded_bytes = bytearray(needed_bytes)
|
||||
for i in range(0, needed_bytes - 1):
|
||||
encoded_bytes[i] = (offset & 127) | 128 # 7 least significant bits
|
||||
offset = offset >> 7
|
||||
encoded_bytes[-1] = offset & 127 # leave first bit as zero for last byte
|
||||
|
||||
return bytes(encoded_bytes)
|
||||
|
||||
|
||||
def varint_decode(encoded):
|
||||
decoded = 0
|
||||
for i, byte in enumerate(encoded):
|
||||
decoded |= (byte & 127) << 7*i
|
||||
|
||||
if not (byte & 128):
|
||||
break
|
||||
return decoded
|
||||
|
||||
|
||||
def string(field_number, data):
|
||||
data = as_bytes(data)
|
||||
return _proto_field(2, field_number, varint_encode(len(data)) + data)
|
||||
nested = string
|
||||
|
||||
def uint(field_number, value):
|
||||
return _proto_field(0, field_number, varint_encode(value))
|
||||
|
||||
|
||||
|
||||
|
||||
def _proto_field(wire_type, field_number, data):
|
||||
''' See https://developers.google.com/protocol-buffers/docs/encoding#structure '''
|
||||
return varint_encode( (field_number << 3) | wire_type) + data
|
||||
|
||||
|
||||
|
||||
def percent_b64encode(data):
|
||||
return base64.urlsafe_b64encode(data).replace(b'=', b'%3D')
|
||||
|
||||
|
||||
def unpadded_b64encode(data):
|
||||
return base64.urlsafe_b64encode(data).replace(b'=', b'')
|
||||
|
||||
def as_bytes(value):
|
||||
if isinstance(value, str):
|
||||
return value.encode('utf-8')
|
||||
return value
|
||||
|
||||
|
||||
def read_varint(data):
|
||||
result = 0
|
||||
i = 0
|
||||
while True:
|
||||
try:
|
||||
byte = data.read(1)[0]
|
||||
except IndexError:
|
||||
if i == 0:
|
||||
raise EOFError()
|
||||
raise Exception('Unterminated varint starting at ' + str(data.tell() - i))
|
||||
result |= (byte & 127) << 7*i
|
||||
if not byte & 128:
|
||||
break
|
||||
|
||||
i += 1
|
||||
return result
|
||||
|
||||
|
||||
def read_group(data, end_sequence):
|
||||
start = data.tell()
|
||||
index = data.original.find(end_sequence, start)
|
||||
if index == -1:
|
||||
raise Exception('Unterminated group')
|
||||
data.seek(index + len(end_sequence))
|
||||
return data.original[start:index]
|
||||
|
||||
def read_protobuf(data):
|
||||
data_original = data
|
||||
data = io.BytesIO(data)
|
||||
data.original = data_original
|
||||
while True:
|
||||
try:
|
||||
tag = read_varint(data)
|
||||
except EOFError:
|
||||
break
|
||||
wire_type = tag & 7
|
||||
field_number = tag >> 3
|
||||
|
||||
if wire_type == 0:
|
||||
value = read_varint(data)
|
||||
elif wire_type == 1:
|
||||
value = data.read(8)
|
||||
elif wire_type == 2:
|
||||
length = read_varint(data)
|
||||
value = data.read(length)
|
||||
elif wire_type == 3:
|
||||
end_bytes = encode_varint((field_number << 3) | 4)
|
||||
value = read_group(data, end_bytes)
|
||||
elif wire_type == 5:
|
||||
value = data.read(4)
|
||||
else:
|
||||
raise Exception("Unknown wire type: " + str(wire_type) + ", Tag: " + bytes_to_hex(succinct_encode(tag)) + ", at position " + str(data.tell()))
|
||||
yield (wire_type, field_number, value)
|
||||
|
||||
def parse(data):
|
||||
return {field_number: value for _, field_number, value in read_protobuf(data)}
|
||||
|
||||
def b64_to_bytes(data):
|
||||
if isinstance(data, bytes):
|
||||
data = data.decode('ascii')
|
||||
data = data.replace("%3D", "=")
|
||||
return base64.urlsafe_b64decode(data + "="*((4 - len(data)%4)%4) )
|
||||
171
youtube_data/search.py
Normal file
171
youtube_data/search.py
Normal file
@ -0,0 +1,171 @@
|
||||
from bs4 import BeautifulSoup as bs
|
||||
from youtube_data import proto
|
||||
from flask import Markup
|
||||
import urllib.parse
|
||||
import requests
|
||||
import base64
|
||||
import json
|
||||
|
||||
def page_number_to_sp_parameter(page, autocorrect, sort, filters):
|
||||
offset = (int(page) - 1)*20 # 20 results per page
|
||||
autocorrect = proto.nested(8, proto.uint(1, 1 - int(autocorrect) ))
|
||||
filters_enc = proto.nested(2, proto.uint(1, filters['time']) + proto.uint(2, filters['type']) + proto.uint(3, filters['duration']))
|
||||
result = proto.uint(1, sort) + filters_enc + autocorrect + proto.uint(9, offset) + proto.string(61, b'')
|
||||
return base64.urlsafe_b64encode(result).decode('ascii')
|
||||
|
||||
def search_by_terms(search_terms, page, autocorrect, sort, filters):
|
||||
url = "https://www.youtube.com/results?search_query=" + urllib.parse.quote_plus(search_terms)
|
||||
headers = {
|
||||
'Host': 'www.youtube.com',
|
||||
'User-Agent': 'Mozilla/5.0 (Windows NT 6.1; Win64; x64)',
|
||||
'Accept': '*/*',
|
||||
'Accept-Language': 'en-US,en;q=0.5',
|
||||
'X-YouTube-Client-Name': '1',
|
||||
'X-YouTube-Client-Version': '2.20180418',
|
||||
}
|
||||
url += "&pbj=1&sp=" + page_number_to_sp_parameter(page, autocorrect, sort, filters).replace("=", "%3D")
|
||||
content = requests.get(url, headers=headers).text
|
||||
|
||||
info = json.loads(content)
|
||||
videos = get_videos_from_search(info)
|
||||
channels = get_channels_from_search(info)
|
||||
|
||||
results = {
|
||||
"videos": videos,
|
||||
"channels": channels
|
||||
}
|
||||
return results
|
||||
|
||||
def get_channels_from_search(search):
|
||||
results = []
|
||||
search = search[1]['response']
|
||||
primaryContents = search['contents']['twoColumnSearchResultsRenderer']['primaryContents']
|
||||
items = primaryContents['sectionListRenderer']['contents'][0]['itemSectionRenderer']['contents']
|
||||
|
||||
for item in items:
|
||||
try:
|
||||
item['channelRenderer']
|
||||
channel = get_channel_renderer_item_info(item['channelRenderer'])
|
||||
results.append(channel)
|
||||
except KeyError:
|
||||
continue
|
||||
return results
|
||||
|
||||
def get_channel_renderer_item_info(item):
|
||||
try:
|
||||
suscribers = item['subscriberCountText']['simpleText'].split(" ")[0]
|
||||
except:
|
||||
suscribers = "?"
|
||||
|
||||
try:
|
||||
description = get_description_snippet_text(item['descriptionSnippet']['runs'])
|
||||
except KeyError:
|
||||
description = ""
|
||||
|
||||
try:
|
||||
channel = {
|
||||
"channelId": item['channelId'],
|
||||
"username": item['title']['simpleText'],
|
||||
"thumbnail": "https:{}".format(item['thumbnail']['thumbnails'][0]['url'].replace("/", "~")),
|
||||
"description": Markup(str(description)),
|
||||
"suscribers": suscribers,
|
||||
"videos": item['videoCountText']['runs'][0]['text']
|
||||
}
|
||||
except KeyError:
|
||||
channel = {
|
||||
"channelId": item['channelId'],
|
||||
"username": item['title']['simpleText'],
|
||||
"avatar": item['thumbnail']['thumbnails'][0]['url'],
|
||||
"suscribers": suscribers
|
||||
}
|
||||
return channel
|
||||
|
||||
def get_videos_from_search(search):
|
||||
latest = []
|
||||
results = []
|
||||
search = search[1]['response']
|
||||
primaryContents = search['contents']['twoColumnSearchResultsRenderer']['primaryContents']
|
||||
items = primaryContents['sectionListRenderer']['contents'][0]['itemSectionRenderer']['contents']
|
||||
|
||||
for item in items:
|
||||
try:
|
||||
item['videoRenderer']
|
||||
video = get_video_renderer_item_info(item['videoRenderer'])
|
||||
results.append(video)
|
||||
except KeyError:
|
||||
continue
|
||||
|
||||
# Sometimes Youtube will return an empty query. Try again.
|
||||
return results
|
||||
|
||||
def get_description_snippet_text(ds):
|
||||
string = ""
|
||||
for t in ds:
|
||||
try:
|
||||
if t['bold']:
|
||||
text = "<b>"+t['text']+"</b>"
|
||||
else:
|
||||
text = t['text']
|
||||
except:
|
||||
text = t['text']
|
||||
string = string + text
|
||||
return string
|
||||
|
||||
def get_video_renderer_item_info(item):
|
||||
published = ""
|
||||
views = ""
|
||||
isLive = False
|
||||
isUpcoming = False
|
||||
|
||||
thumbnailOverlays = item['thumbnailOverlays']
|
||||
try:
|
||||
if 'UPCOMING' in str(thumbnailOverlays):
|
||||
start_time = item['upcomingEventData']['startTime']
|
||||
isUpcoming = True
|
||||
views = "-"
|
||||
published = "Scheduled"
|
||||
except KeyError:
|
||||
isUpcoming = False
|
||||
|
||||
try:
|
||||
if 'LIVE' in str(thumbnailOverlays):
|
||||
isLive = True
|
||||
try:
|
||||
views = item['viewCountText']['simpleText']
|
||||
except:
|
||||
views = "Live"
|
||||
try:
|
||||
duration = item['lengthText']['simpleText']
|
||||
except:
|
||||
duration = "-"
|
||||
if published != "Scheduled":
|
||||
try:
|
||||
published = item['publishedTimeText']['simpleText']
|
||||
except KeyError:
|
||||
published = "None"
|
||||
except:
|
||||
isUpcoming = False
|
||||
isLive = False
|
||||
|
||||
if not isUpcoming and not isLive:
|
||||
views = item['viewCountText']['simpleText']
|
||||
published = item['publishedTimeText']['simpleText']
|
||||
duration = item['lengthText']['simpleText']
|
||||
|
||||
video = {
|
||||
'videoTitle':item['title']['runs'][0]['text'],
|
||||
'description':Markup(str(get_description_snippet_text(item['descriptionSnippet']['runs']))),
|
||||
'views':views,
|
||||
'timeStamp':published,
|
||||
'duration':duration,
|
||||
'channelName':item['ownerText']['runs'][0]['text'],
|
||||
'authorUrl':"/channel/{}".format(item['ownerText']['runs'][0]['navigationEndpoint']['browseEndpoint']['browseId']),
|
||||
'channelId':item['ownerText']['runs'][0]['navigationEndpoint']['browseEndpoint']['browseId'],
|
||||
'id':item['videoId'],
|
||||
'videoUrl':"/watch?v={}".format(item['videoId']),
|
||||
'isLive':isLive,
|
||||
'isUpcoming':isUpcoming,
|
||||
'videoThumb':item['thumbnail']['thumbnails'][0]['url']
|
||||
}
|
||||
return video
|
||||
|
||||
@ -11,7 +11,6 @@ def get_renderer_key(renderer, key):
|
||||
return k[key]
|
||||
|
||||
def get_video_primary_info(datad, datai):
|
||||
|
||||
contents = datai["contents"]["twoColumnWatchNextResults"]['results']['results']['contents']
|
||||
item = get_renderer_key(contents, "videoPrimaryInfoRenderer")
|
||||
details = datad['videoDetails']
|
||||
|
||||
Reference in New Issue
Block a user