kabigon 0.6.1__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
@@ -10,7 +10,9 @@ 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
15
+ from .twitter import TwitterLoader
14
16
  from .youtube import YoutubeLoader
15
17
  from .ytdlp import YtdlpLoader
16
18
 
kabigon/cli.py CHANGED
@@ -1,27 +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
+ from .twitter import TwitterLoader
9
10
  from .youtube import YoutubeLoader
10
11
  from .ytdlp import YtdlpLoader
11
12
 
12
13
 
13
- @click.command()
14
- @click.argument("url", type=click.STRING)
15
- def main(url: str) -> None:
14
+ def run(url: str) -> None:
16
15
  loader = Compose(
17
16
  [
17
+ PttLoader(),
18
+ TwitterLoader(),
18
19
  YoutubeLoader(),
19
20
  ReelLoader(),
20
21
  YtdlpLoader(),
21
22
  PDFLoader(),
22
- HttpxLoader(),
23
23
  PlaywrightLoader(),
24
24
  ]
25
25
  )
26
26
  result = loader.load(url)
27
27
  print(result)
28
+
29
+
30
+ def main() -> None:
31
+ typer.run(run)
kabigon/compose.py CHANGED
@@ -1,41 +1,14 @@
1
- from urllib.parse import urlparse
2
- from urllib.parse import urlunparse
3
-
4
1
  from loguru import logger
5
2
 
6
3
  from .errors import KabigonError
7
4
  from .loader import Loader
8
5
 
9
- REPLACEMENTS = {
10
- # fixupx.com seems better than api.fxtwitter.com
11
- "fixupx.com": [
12
- "twitter.com",
13
- "x.com",
14
- "fxtwitter.com",
15
- "vxtwitter.com",
16
- "fixvx.com",
17
- "twittpr.com",
18
- "api.fxtwitter.com",
19
- ]
20
- }
21
-
22
-
23
- def replace_domain(url: str) -> str:
24
- parsed = urlparse(url)
25
- for target, source in REPLACEMENTS.items():
26
- if parsed.netloc in source:
27
- fixed_url = parsed._replace(netloc=target)
28
- return urlunparse(fixed_url)
29
- return url
30
-
31
6
 
32
7
  class Compose(Loader):
33
8
  def __init__(self, loaders: list[Loader]) -> None:
34
9
  self.loaders = loaders
35
10
 
36
11
  def load(self, url: str) -> str:
37
- url = replace_domain(url)
38
-
39
12
  for loader in self.loaders:
40
13
  try:
41
14
  content = loader.load(url)
@@ -53,8 +26,6 @@ class Compose(Loader):
53
26
  raise KabigonError(f"Failed to load URL: {url}")
54
27
 
55
28
  async def async_load(self, url: str) -> str:
56
- url = replace_domain(url)
57
-
58
29
  for loader in self.loaders:
59
30
  try:
60
31
  content = await loader.async_load(url)
kabigon/errors.py CHANGED
@@ -21,3 +21,8 @@ class NotPDFError(KabigonError):
21
21
  class NotReelURLError(KabigonError):
22
22
  def __init__(self, url: str):
23
23
  super().__init__(f"URL is not an Instagram Reel: {url}")
24
+
25
+
26
+ class NotTwitterURLError(KabigonError):
27
+ def __init__(self, url: str):
28
+ super().__init__(f"URL is not a Twitter URL: {url}")
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)
kabigon/twitter.py ADDED
@@ -0,0 +1,46 @@
1
+ from urllib.parse import urlparse
2
+ from urllib.parse import urlunparse
3
+
4
+ from .errors import NotTwitterURLError
5
+ from .loader import Loader
6
+ from .playwright import PlaywrightLoader
7
+
8
+ TWITTER_DOMAINS = [
9
+ "twitter.com",
10
+ "x.com",
11
+ "fxtwitter.com",
12
+ "vxtwitter.com",
13
+ "fixvx.com",
14
+ "twittpr.com",
15
+ "api.fxtwitter.com",
16
+ "fixupx.com",
17
+ ]
18
+
19
+
20
+ def replace_domain(url: str, new_domain: str = "x.com") -> str:
21
+ return urlunparse(urlparse(url)._replace(netloc=new_domain))
22
+
23
+
24
+ def is_x_url(url: str) -> bool:
25
+ return urlparse(url).netloc in TWITTER_DOMAINS
26
+
27
+
28
+ class TwitterLoader(Loader):
29
+ def __init__(self) -> None:
30
+ self.playwright_loader = PlaywrightLoader(wait_until="networkidle")
31
+
32
+ def load(self, url: str) -> str:
33
+ if not is_x_url(url):
34
+ raise NotTwitterURLError(url)
35
+
36
+ url = replace_domain(url)
37
+
38
+ return self.playwright_loader.load(url)
39
+
40
+ async def async_load(self, url: str):
41
+ if not is_x_url(url):
42
+ raise NotTwitterURLError(url)
43
+
44
+ url = replace_domain(url)
45
+
46
+ return await self.playwright_loader.async_load(url)
@@ -1,22 +1,22 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: kabigon
3
- Version: 0.6.1
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
12
11
  Requires-Dist: markdownify>=0.14.1
13
12
  Requires-Dist: openai-whisper>=20240930
14
- Requires-Dist: playwright>=1.50.0
13
+ 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
- Requires-Dist: yt-dlp>=2025.1.26
19
+ Requires-Dist: yt-dlp>=2025.4.30
20
20
  Description-Content-Type: text/markdown
21
21
 
22
22
  # kabigon
@@ -0,0 +1,21 @@
1
+ kabigon/__init__.py,sha256=MUfTFUe5ezA249L2yuU5_2FiewLu86H3VsIpJSne2vQ,560
2
+ kabigon/cli.py,sha256=PJ0wnwp_AgHA54YxGr1jNJ_q3ls7fEymgTJaJxCVU7M,650
3
+ kabigon/compose.py,sha256=l2D5OK91VcN2a6DbjMdwBk3YSqzVV7fOVX0TqNm2gJo,1502
4
+ kabigon/errors.py,sha256=iri_YS71UsOHwaVtfy5IA6iUfq30DCsptZsChmZaZic,755
5
+ kabigon/firecrawl.py,sha256=Xnrlhco_R58x5kwGy9ZCKTnVqS4Pp-D3G0u-qnuGEsU,800
6
+ kabigon/httpx.py,sha256=Zup9DURyWLqoWzaxBbCYAaV-5LSlHUuAcNyyUsZTVag,696
7
+ kabigon/loader.py,sha256=KhOJvlzLDM0o05o4VqMRgiyRVjofkzuGEcihQagj-8M,452
8
+ kabigon/pdf.py,sha256=yJcgkdMMF52baFsFxJp9Jn89KsTKBboLTLwD3gs6U5U,1775
9
+ kabigon/playwright.py,sha256=MZ-r0Ej2wWAOJkDLwYRvO77wcDvh38KXz2wgDsCTgm0,1358
10
+ kabigon/ptt.py,sha256=S2d6SeFGxM4E4kg-n5blN2BX56CWj_eOhapJxzUnxu8,965
11
+ kabigon/py.typed,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
12
+ kabigon/reel.py,sha256=J2QOxGMYi_HaEscQPIipPEoHGN_iksGxR6pV_XvryME,929
13
+ kabigon/twitter.py,sha256=U07pa8xA0nHAaaDPeUelQRvXR5ZnUvYJZW35xRAvHA8,1114
14
+ kabigon/utils.py,sha256=eNTLtHLSB2erDac2HH3jWemgfr8Ou_ozwVb8h9BD-4g,922
15
+ kabigon/youtube.py,sha256=HoiFNq0ookPL7_rO_wloBaY8yTIX6xP8A77F7y02q64,1166
16
+ kabigon/ytdlp.py,sha256=_QRcyFx9s7NnI1MvcWdKKxlX-hHLnqtduCSL5_UH6dU,3140
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,,
@@ -1,19 +0,0 @@
1
- kabigon/__init__.py,sha256=_1LdKfp7qm0m7Fa_IJ9zcg4TRB14c9bB-r9M3fwASWI,498
2
- kabigon/cli.py,sha256=XSTyD1RFqq2Qok_52kSjJlBLUXl6t-K9QtsxCfB15o4,611
3
- kabigon/compose.py,sha256=5H_hWQ7ENUmEdeOpB2bkRv4U_U18_1IyNTran-cLYaM,2166
4
- kabigon/errors.py,sha256=3eSPQtLocreKuq9mhwRk7IMdA2xh8KkmIFEUhGmTPxg,618
5
- kabigon/firecrawl.py,sha256=Xnrlhco_R58x5kwGy9ZCKTnVqS4Pp-D3G0u-qnuGEsU,800
6
- kabigon/httpx.py,sha256=uDdLks6zVzirY7-mnsJkypX86kAI5XmUVfK-lFifdJA,895
7
- kabigon/loader.py,sha256=KhOJvlzLDM0o05o4VqMRgiyRVjofkzuGEcihQagj-8M,452
8
- kabigon/pdf.py,sha256=yJcgkdMMF52baFsFxJp9Jn89KsTKBboLTLwD3gs6U5U,1775
9
- kabigon/playwright.py,sha256=MZ-r0Ej2wWAOJkDLwYRvO77wcDvh38KXz2wgDsCTgm0,1358
10
- kabigon/py.typed,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
11
- kabigon/reel.py,sha256=J2QOxGMYi_HaEscQPIipPEoHGN_iksGxR6pV_XvryME,929
12
- kabigon/utils.py,sha256=eNTLtHLSB2erDac2HH3jWemgfr8Ou_ozwVb8h9BD-4g,922
13
- kabigon/youtube.py,sha256=HoiFNq0ookPL7_rO_wloBaY8yTIX6xP8A77F7y02q64,1166
14
- kabigon/ytdlp.py,sha256=_QRcyFx9s7NnI1MvcWdKKxlX-hHLnqtduCSL5_UH6dU,3140
15
- kabigon-0.6.1.dist-info/METADATA,sha256=78J35ClbTdy2-vU8GtlQWjgSTitntFI8J-Cz07FGqoo,1078
16
- kabigon-0.6.1.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
17
- kabigon-0.6.1.dist-info/entry_points.txt,sha256=O3FYAO9w-NQvlGMJrBvtrnGHSK2QkUnQBTa30YXRbVE,45
18
- kabigon-0.6.1.dist-info/licenses/LICENSE,sha256=H2T3_RTgmcngMeC7p_SXT3GwBLkd2DaNgAZuxulcfiA,1066
19
- kabigon-0.6.1.dist-info/RECORD,,