kabigon 0.8.11__py3-none-any.whl → 0.8.13__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- kabigon/playwright.py +12 -2
- kabigon/twitter.py +2 -2
- kabigon/youtube.py +3 -5
- {kabigon-0.8.11.dist-info → kabigon-0.8.13.dist-info}/METADATA +2 -2
- {kabigon-0.8.11.dist-info → kabigon-0.8.13.dist-info}/RECORD +8 -8
- {kabigon-0.8.11.dist-info → kabigon-0.8.13.dist-info}/WHEEL +0 -0
- {kabigon-0.8.11.dist-info → kabigon-0.8.13.dist-info}/entry_points.txt +0 -0
- {kabigon-0.8.11.dist-info → kabigon-0.8.13.dist-info}/licenses/LICENSE +0 -0
kabigon/playwright.py
CHANGED
@@ -1,5 +1,7 @@
|
|
1
1
|
from typing import Literal
|
2
2
|
|
3
|
+
from loguru import logger
|
4
|
+
|
3
5
|
from .loader import Loader
|
4
6
|
from .utils import html_to_markdown
|
5
7
|
|
@@ -17,6 +19,7 @@ class PlaywrightLoader(Loader):
|
|
17
19
|
|
18
20
|
def load(self, url: str) -> str:
|
19
21
|
try:
|
22
|
+
from playwright.sync_api import TimeoutError
|
20
23
|
from playwright.sync_api import sync_playwright
|
21
24
|
except ImportError as e:
|
22
25
|
raise ImportError(
|
@@ -29,7 +32,10 @@ class PlaywrightLoader(Loader):
|
|
29
32
|
context = browser.new_context()
|
30
33
|
page = context.new_page()
|
31
34
|
|
32
|
-
|
35
|
+
try:
|
36
|
+
page.goto(url, timeout=self.timeout, wait_until=self.wait_until)
|
37
|
+
except TimeoutError as e:
|
38
|
+
logger.warning("TimeoutError: {}, (url: {}, timeout: {})", e, url, self.timeout)
|
33
39
|
|
34
40
|
content = page.content()
|
35
41
|
browser.close()
|
@@ -38,6 +44,7 @@ class PlaywrightLoader(Loader):
|
|
38
44
|
|
39
45
|
async def async_load(self, url: str) -> str:
|
40
46
|
try:
|
47
|
+
from playwright.async_api import TimeoutError
|
41
48
|
from playwright.async_api import async_playwright
|
42
49
|
except ImportError as e:
|
43
50
|
raise ImportError(
|
@@ -50,7 +57,10 @@ class PlaywrightLoader(Loader):
|
|
50
57
|
context = await browser.new_context()
|
51
58
|
page = await context.new_page()
|
52
59
|
|
53
|
-
|
60
|
+
try:
|
61
|
+
await page.goto(url, timeout=self.timeout, wait_until=self.wait_until)
|
62
|
+
except TimeoutError as e:
|
63
|
+
logger.warning("TimeoutError: {}, (url: {}, timeout: {})", e, url, self.timeout)
|
54
64
|
|
55
65
|
content = await page.content()
|
56
66
|
await browser.close()
|
kabigon/twitter.py
CHANGED
@@ -17,7 +17,7 @@ TWITTER_DOMAINS = [
|
|
17
17
|
|
18
18
|
|
19
19
|
def replace_domain(url: str, new_domain: str = "x.com") -> str:
|
20
|
-
return urlunparse(urlparse(url)._replace(netloc=new_domain))
|
20
|
+
return str(urlunparse(urlparse(url)._replace(netloc=new_domain)))
|
21
21
|
|
22
22
|
|
23
23
|
def check_x_url(url: str) -> None:
|
@@ -27,7 +27,7 @@ def check_x_url(url: str) -> None:
|
|
27
27
|
|
28
28
|
class TwitterLoader(Loader):
|
29
29
|
def __init__(self) -> None:
|
30
|
-
self.playwright_loader = PlaywrightLoader(wait_until="networkidle")
|
30
|
+
self.playwright_loader = PlaywrightLoader(wait_until="networkidle", timeout=10_000)
|
31
31
|
|
32
32
|
def load(self, url: str) -> str:
|
33
33
|
check_x_url(url)
|
kabigon/youtube.py
CHANGED
@@ -14,13 +14,11 @@ class YoutubeLoader(Loader):
|
|
14
14
|
def load(self, url: str) -> str:
|
15
15
|
video_id = aioytt.video_id.parse_video_id(url)
|
16
16
|
|
17
|
-
|
18
|
-
video_id, self.languages
|
19
|
-
)
|
17
|
+
fetched = YouTubeTranscriptApi().fetch(video_id, self.languages)
|
20
18
|
|
21
19
|
lines = []
|
22
|
-
for
|
23
|
-
text = str(
|
20
|
+
for snippet in fetched.snippets:
|
21
|
+
text = str(snippet.text).strip()
|
24
22
|
if text:
|
25
23
|
lines.append(text)
|
26
24
|
return "\n".join(lines)
|
@@ -1,6 +1,6 @@
|
|
1
1
|
Metadata-Version: 2.4
|
2
2
|
Name: kabigon
|
3
|
-
Version: 0.8.
|
3
|
+
Version: 0.8.13
|
4
4
|
Author-email: narumi <toucans-cutouts0f@icloud.com>
|
5
5
|
License-File: LICENSE
|
6
6
|
Requires-Python: >=3.10
|
@@ -12,7 +12,7 @@ Requires-Dist: markdownify>=0.14.1
|
|
12
12
|
Requires-Dist: pypdf>=5.3.0
|
13
13
|
Requires-Dist: rich>=13.9.4
|
14
14
|
Requires-Dist: typer>=0.15.3
|
15
|
-
Requires-Dist: youtube-transcript-api>=
|
15
|
+
Requires-Dist: youtube-transcript-api>=1.2.2
|
16
16
|
Requires-Dist: yt-dlp>=2025.4.30
|
17
17
|
Provides-Extra: all
|
18
18
|
Requires-Dist: openai-whisper>=20240930; extra == 'all'
|
@@ -5,17 +5,17 @@ kabigon/firecrawl.py,sha256=-5AI9tla_684dtpubY_BRudqLgw28158WdwA1RjJvAA,778
|
|
5
5
|
kabigon/httpx.py,sha256=Zup9DURyWLqoWzaxBbCYAaV-5LSlHUuAcNyyUsZTVag,696
|
6
6
|
kabigon/loader.py,sha256=W2ZdSINQKjG-5XNYweM8lIl19JA4wwd7AE2pueHeO70,380
|
7
7
|
kabigon/pdf.py,sha256=PNOx-Dz_VpN-EVnVar_wJJZdxWrDZrAFE-gSuUR9q7o,1870
|
8
|
-
kabigon/playwright.py,sha256=
|
8
|
+
kabigon/playwright.py,sha256=gECaDJhsIrY53lwG7OrH8gUHvIsVe7p5jgBTUy7KPJg,2413
|
9
9
|
kabigon/ptt.py,sha256=Gyp2nJrjptkjbwZJ9VEQHX0DEgKBe5QRQOmGVHUUgNA,896
|
10
10
|
kabigon/py.typed,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
|
11
11
|
kabigon/reel.py,sha256=qOwWCvcp7xNKg0JDunq_Bsl8yqqMzrnAOI9k5mSqrOU,874
|
12
|
-
kabigon/twitter.py,sha256
|
12
|
+
kabigon/twitter.py,sha256=-EcwbvgIEXXsGqDHaVZQ-VKJHbxhmT8FFn8bGTIdXiw,1067
|
13
13
|
kabigon/utils.py,sha256=eNTLtHLSB2erDac2HH3jWemgfr8Ou_ozwVb8h9BD-4g,922
|
14
|
-
kabigon/youtube.py,sha256=
|
14
|
+
kabigon/youtube.py,sha256=Vh4cBidScUruTgvm4BbTxqaEDxnotMI3uM3FuFlx5J0,1021
|
15
15
|
kabigon/youtube_ytdlp.py,sha256=Y6h55wYF-5PdxLRxsznFtJDypCedD8g-hY9fgCnfvbg,722
|
16
16
|
kabigon/ytdlp.py,sha256=4AjPY0W8IxbEhCNVXpi_KbBQLGrWn5ThrchhffcJIIU,1718
|
17
|
-
kabigon-0.8.
|
18
|
-
kabigon-0.8.
|
19
|
-
kabigon-0.8.
|
20
|
-
kabigon-0.8.
|
21
|
-
kabigon-0.8.
|
17
|
+
kabigon-0.8.13.dist-info/METADATA,sha256=37MjwEQKzcXbHdsF3mfIFePL3rN0n7xUJenVWkivz6w,1265
|
18
|
+
kabigon-0.8.13.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
|
19
|
+
kabigon-0.8.13.dist-info/entry_points.txt,sha256=O3FYAO9w-NQvlGMJrBvtrnGHSK2QkUnQBTa30YXRbVE,45
|
20
|
+
kabigon-0.8.13.dist-info/licenses/LICENSE,sha256=H2T3_RTgmcngMeC7p_SXT3GwBLkd2DaNgAZuxulcfiA,1066
|
21
|
+
kabigon-0.8.13.dist-info/RECORD,,
|
File without changes
|
File without changes
|
File without changes
|