Alexander Czyrny commited on
Extracting publish date (#794)
Browse files* Extracting publish date
* Added unit test for .publish_date attribute.
- pytube/__main__.py +9 -0
- pytube/extract.py +19 -0
- tests/test_extract.py +7 -0
pytube/__main__.py
CHANGED
|
@@ -262,6 +262,15 @@ class YouTube:
|
|
| 262 |
|
| 263 |
return f"https://img.youtube.com/vi/{self.video_id}/maxresdefault.jpg"
|
| 264 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 265 |
@property
|
| 266 |
def title(self) -> str:
|
| 267 |
"""Get the video title.
|
|
|
|
| 262 |
|
| 263 |
return f"https://img.youtube.com/vi/{self.video_id}/maxresdefault.jpg"
|
| 264 |
|
| 265 |
+
@property
|
| 266 |
+
def publish_date(self):
|
| 267 |
+
"""Get the publish date.
|
| 268 |
+
|
| 269 |
+
:rtype: datetime
|
| 270 |
+
|
| 271 |
+
"""
|
| 272 |
+
return extract.publish_date(self.watch_html)
|
| 273 |
+
|
| 274 |
@property
|
| 275 |
def title(self) -> str:
|
| 276 |
"""Get the video title.
|
pytube/extract.py
CHANGED
|
@@ -4,6 +4,7 @@ import json
|
|
| 4 |
import logging
|
| 5 |
import re
|
| 6 |
from collections import OrderedDict
|
|
|
|
| 7 |
from typing import Any
|
| 8 |
from typing import Dict
|
| 9 |
from typing import List
|
|
@@ -22,6 +23,24 @@ from pytube.helpers import regex_search
|
|
| 22 |
logger = logging.getLogger(__name__)
|
| 23 |
|
| 24 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 25 |
def recording_available(watch_html):
|
| 26 |
"""Check if live stream recording is available.
|
| 27 |
|
|
|
|
| 4 |
import logging
|
| 5 |
import re
|
| 6 |
from collections import OrderedDict
|
| 7 |
+
from datetime import datetime
|
| 8 |
from typing import Any
|
| 9 |
from typing import Dict
|
| 10 |
from typing import List
|
|
|
|
| 23 |
logger = logging.getLogger(__name__)
|
| 24 |
|
| 25 |
|
| 26 |
+
def publish_date(watch_html: str):
|
| 27 |
+
"""Extract publish date
|
| 28 |
+
:param str watch_html:
|
| 29 |
+
The html contents of the watch page.
|
| 30 |
+
:rtype: str
|
| 31 |
+
:returns:
|
| 32 |
+
Publish date of the video.
|
| 33 |
+
"""
|
| 34 |
+
try:
|
| 35 |
+
result = regex_search(
|
| 36 |
+
r"(?<=itemprop=\"datePublished\" content=\")\d{4}-\d{2}-\d{2}",
|
| 37 |
+
watch_html, group=0
|
| 38 |
+
)
|
| 39 |
+
except RegexMatchError:
|
| 40 |
+
return None
|
| 41 |
+
return datetime.strptime(result, '%Y-%m-%d')
|
| 42 |
+
|
| 43 |
+
|
| 44 |
def recording_available(watch_html):
|
| 45 |
"""Check if live stream recording is available.
|
| 46 |
|
tests/test_extract.py
CHANGED
|
@@ -1,5 +1,6 @@
|
|
| 1 |
# -*- coding: utf-8 -*-
|
| 2 |
"""Unit tests for the :module:`extract <extract>` module."""
|
|
|
|
| 3 |
import pytest
|
| 4 |
|
| 5 |
from pytube import extract
|
|
@@ -64,6 +65,12 @@ def test_recording_available(cipher_signature):
|
|
| 64 |
assert extract.recording_available(cipher_signature.watch_html)
|
| 65 |
|
| 66 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
def test_not_recording_available(missing_recording):
|
| 68 |
assert not extract.recording_available(missing_recording['watch_html'])
|
| 69 |
|
|
|
|
| 1 |
# -*- coding: utf-8 -*-
|
| 2 |
"""Unit tests for the :module:`extract <extract>` module."""
|
| 3 |
+
from datetime import datetime
|
| 4 |
import pytest
|
| 5 |
|
| 6 |
from pytube import extract
|
|
|
|
| 65 |
assert extract.recording_available(cipher_signature.watch_html)
|
| 66 |
|
| 67 |
|
| 68 |
+
def test_publish_date(cipher_signature):
|
| 69 |
+
expected = datetime(2019, 12, 5)
|
| 70 |
+
assert cipher_signature.publish_date == expected
|
| 71 |
+
assert extract.publish_date('') is None
|
| 72 |
+
|
| 73 |
+
|
| 74 |
def test_not_recording_available(missing_recording):
|
| 75 |
assert not extract.recording_available(missing_recording['watch_html'])
|
| 76 |
|