kabigon 0.7.0__py3-none-any.whl → 0.8.0__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/__init__.py CHANGED
@@ -10,6 +10,7 @@ from .httpx import HttpxLoader
10
10
  from .loader import Loader
11
11
  from .pdf import PDFLoader
12
12
  from .playwright import PlaywrightLoader
13
+ from .ptt import PttLoader
13
14
  from .reel import ReelLoader
14
15
  from .twitter import TwitterLoader
15
16
  from .youtube import YoutubeLoader
kabigon/cli.py CHANGED
@@ -1,29 +1,31 @@
1
- import click
1
+ import typer
2
2
  from rich import print
3
3
 
4
4
  from .compose import Compose
5
- from .httpx import HttpxLoader
6
5
  from .pdf import PDFLoader
7
6
  from .playwright import PlaywrightLoader
7
+ from .ptt import PttLoader
8
8
  from .reel import ReelLoader
9
9
  from .twitter import TwitterLoader
10
10
  from .youtube import YoutubeLoader
11
11
  from .ytdlp import YtdlpLoader
12
12
 
13
13
 
14
- @click.command()
15
- @click.argument("url", type=click.STRING)
16
- def main(url: str) -> None:
14
+ def run(url: str) -> None:
17
15
  loader = Compose(
18
16
  [
17
+ PttLoader(),
19
18
  TwitterLoader(),
20
19
  YoutubeLoader(),
21
20
  ReelLoader(),
22
21
  YtdlpLoader(),
23
22
  PDFLoader(),
24
- HttpxLoader(),
25
23
  PlaywrightLoader(),
26
24
  ]
27
25
  )
28
26
  result = loader.load(url)
29
27
  print(result)
28
+
29
+
30
+ def main() -> None:
31
+ typer.run(run)
kabigon/httpx.py CHANGED
@@ -3,21 +3,18 @@ import httpx
3
3
  from .loader import Loader
4
4
  from .utils import html_to_markdown
5
5
 
6
- DEFAULT_HEADERS = {
7
- "Accept-Language": "zh-TW,zh;q=0.9,ja;q=0.8,en-US;q=0.7,en;q=0.6",
8
- "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/129.0.0.0 Safari/537.36", # noqa
9
- "Cookie": "over18=1", # Required for some sites like PTT
10
- }
11
-
12
6
 
13
7
  class HttpxLoader(Loader):
8
+ def __init__(self, headers: dict[str, str] | None = None) -> None:
9
+ self.headers = headers
10
+
14
11
  def load(self, url: str) -> str:
15
- response = httpx.get(url, headers=DEFAULT_HEADERS, follow_redirects=True)
12
+ response = httpx.get(url, headers=self.headers, follow_redirects=False)
16
13
  response.raise_for_status()
17
14
  return html_to_markdown(response.content)
18
15
 
19
16
  async def async_load(self, url: str) -> str:
20
17
  async with httpx.AsyncClient() as client:
21
- response = await client.get(url, headers=DEFAULT_HEADERS, follow_redirects=True)
18
+ response = await client.get(url, headers=self.headers, follow_redirects=True)
22
19
  response.raise_for_status()
23
20
  return html_to_markdown(response.content)
kabigon/ptt.py ADDED
@@ -0,0 +1,32 @@
1
+ from urllib.parse import urlparse
2
+
3
+ from .errors import NotTwitterURLError
4
+ from .httpx import HttpxLoader
5
+ from .loader import Loader
6
+
7
+
8
+ def is_ptt_url(url: str) -> bool:
9
+ return urlparse(url).netloc == "www.ptt.cc"
10
+
11
+
12
+ class PttLoader(Loader):
13
+ def __init__(self) -> None:
14
+ self.httpx_loader = HttpxLoader(
15
+ headers={
16
+ "Accept-Language": "zh-TW,zh;q=0.9,ja;q=0.8,en-US;q=0.7,en;q=0.6",
17
+ "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/129.0.0.0 Safari/537.36", # noqa
18
+ "Cookie": "over18=1",
19
+ }
20
+ )
21
+
22
+ def load(self, url: str) -> str:
23
+ if not is_ptt_url(url):
24
+ raise NotTwitterURLError(url)
25
+
26
+ return self.httpx_loader.load(url)
27
+
28
+ async def async_load(self, url: str):
29
+ if not is_ptt_url(url):
30
+ raise NotTwitterURLError(url)
31
+
32
+ return await self.httpx_loader.async_load(url)
@@ -1,11 +1,10 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: kabigon
3
- Version: 0.7.0
3
+ Version: 0.8.0
4
4
  Author-email: narumi <toucans-cutouts0f@icloud.com>
5
5
  License-File: LICENSE
6
6
  Requires-Python: >=3.10
7
7
  Requires-Dist: aioytt>=0.2.4
8
- Requires-Dist: click>=8.1.8
9
8
  Requires-Dist: firecrawl-py>=2.4.1
10
9
  Requires-Dist: httpx>=0.28.1
11
10
  Requires-Dist: loguru>=0.7.3
@@ -15,6 +14,7 @@ Requires-Dist: playwright>=1.52.0
15
14
  Requires-Dist: pypdf>=5.3.0
16
15
  Requires-Dist: rich>=13.9.4
17
16
  Requires-Dist: timeout-decorator>=0.5.0
17
+ Requires-Dist: typer>=0.15.3
18
18
  Requires-Dist: youtube-transcript-api>=0.6.3
19
19
  Requires-Dist: yt-dlp>=2025.4.30
20
20
  Description-Content-Type: text/markdown
@@ -1,20 +1,21 @@
1
- kabigon/__init__.py,sha256=L-MDyX7PrPZvq6mASyOCJnATebwn-Lj1Nobp9GAPeyU,533
2
- kabigon/cli.py,sha256=co9IczXx0bAnXo6EEzEdCMSfpBetkPe6pB5hd2jtzA8,675
1
+ kabigon/__init__.py,sha256=MUfTFUe5ezA249L2yuU5_2FiewLu86H3VsIpJSne2vQ,560
2
+ kabigon/cli.py,sha256=PJ0wnwp_AgHA54YxGr1jNJ_q3ls7fEymgTJaJxCVU7M,650
3
3
  kabigon/compose.py,sha256=l2D5OK91VcN2a6DbjMdwBk3YSqzVV7fOVX0TqNm2gJo,1502
4
4
  kabigon/errors.py,sha256=iri_YS71UsOHwaVtfy5IA6iUfq30DCsptZsChmZaZic,755
5
5
  kabigon/firecrawl.py,sha256=Xnrlhco_R58x5kwGy9ZCKTnVqS4Pp-D3G0u-qnuGEsU,800
6
- kabigon/httpx.py,sha256=uDdLks6zVzirY7-mnsJkypX86kAI5XmUVfK-lFifdJA,895
6
+ kabigon/httpx.py,sha256=Zup9DURyWLqoWzaxBbCYAaV-5LSlHUuAcNyyUsZTVag,696
7
7
  kabigon/loader.py,sha256=KhOJvlzLDM0o05o4VqMRgiyRVjofkzuGEcihQagj-8M,452
8
8
  kabigon/pdf.py,sha256=yJcgkdMMF52baFsFxJp9Jn89KsTKBboLTLwD3gs6U5U,1775
9
9
  kabigon/playwright.py,sha256=MZ-r0Ej2wWAOJkDLwYRvO77wcDvh38KXz2wgDsCTgm0,1358
10
+ kabigon/ptt.py,sha256=S2d6SeFGxM4E4kg-n5blN2BX56CWj_eOhapJxzUnxu8,965
10
11
  kabigon/py.typed,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
11
12
  kabigon/reel.py,sha256=J2QOxGMYi_HaEscQPIipPEoHGN_iksGxR6pV_XvryME,929
12
13
  kabigon/twitter.py,sha256=U07pa8xA0nHAaaDPeUelQRvXR5ZnUvYJZW35xRAvHA8,1114
13
14
  kabigon/utils.py,sha256=eNTLtHLSB2erDac2HH3jWemgfr8Ou_ozwVb8h9BD-4g,922
14
15
  kabigon/youtube.py,sha256=HoiFNq0ookPL7_rO_wloBaY8yTIX6xP8A77F7y02q64,1166
15
16
  kabigon/ytdlp.py,sha256=_QRcyFx9s7NnI1MvcWdKKxlX-hHLnqtduCSL5_UH6dU,3140
16
- kabigon-0.7.0.dist-info/METADATA,sha256=4rng3Fe89ifNd6bQoBggRhCoqTMIa4sWhbU6N75puzo,1078
17
- kabigon-0.7.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
18
- kabigon-0.7.0.dist-info/entry_points.txt,sha256=O3FYAO9w-NQvlGMJrBvtrnGHSK2QkUnQBTa30YXRbVE,45
19
- kabigon-0.7.0.dist-info/licenses/LICENSE,sha256=H2T3_RTgmcngMeC7p_SXT3GwBLkd2DaNgAZuxulcfiA,1066
20
- kabigon-0.7.0.dist-info/RECORD,,
17
+ kabigon-0.8.0.dist-info/METADATA,sha256=SAazEEnqklluyiH2psGQwUI40ee400GELlcxv0OyvuU,1079
18
+ kabigon-0.8.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
19
+ kabigon-0.8.0.dist-info/entry_points.txt,sha256=O3FYAO9w-NQvlGMJrBvtrnGHSK2QkUnQBTa30YXRbVE,45
20
+ kabigon-0.8.0.dist-info/licenses/LICENSE,sha256=H2T3_RTgmcngMeC7p_SXT3GwBLkd2DaNgAZuxulcfiA,1066
21
+ kabigon-0.8.0.dist-info/RECORD,,