kabigon 0.7.0__py3-none-any.whl → 0.8.1__py3-none-any.whl

kabigon/__init__.py

@@ -10,6 +10,7 @@ from .httpx import HttpxLoader
 from .loader import Loader
 from .pdf import PDFLoader
 from .playwright import PlaywrightLoader
+from .ptt import PttLoader
 from .reel import ReelLoader
 from .twitter import TwitterLoader
 from .youtube import YoutubeLoader

kabigon/cli.py

@@ -1,29 +1,31 @@
-import click
+import typer
 from rich import print
 
 from .compose import Compose
-from .httpx import HttpxLoader
 from .pdf import PDFLoader
 from .playwright import PlaywrightLoader
+from .ptt import PttLoader
 from .reel import ReelLoader
 from .twitter import TwitterLoader
 from .youtube import YoutubeLoader
 from .ytdlp import YtdlpLoader
 
 
-@click.command()
-@click.argument("url", type=click.STRING)
-def main(url: str) -> None:
+def run(url: str) -> None:
     loader = Compose(
         [
+            PttLoader(),
             TwitterLoader(),
             YoutubeLoader(),
             ReelLoader(),
             YtdlpLoader(),
             PDFLoader(),
-            HttpxLoader(),
             PlaywrightLoader(),
         ]
     )
     result = loader.load(url)
     print(result)
+
+
+def main() -> None:
+    typer.run(run)

kabigon/compose.py

@@ -1,6 +1,5 @@
 from loguru import logger
 
-from .errors import KabigonError
 from .loader import Loader
 
 
@@ -23,7 +22,7 @@ class Compose(Loader):
             except Exception as e:
                 logger.info("[{}] Failed to load URL: {}, got error: {}", loader.__class__.__name__, url, e)
 
-        raise KabigonError(f"Failed to load URL: {url}")
+        raise Exception(f"Failed to load URL: {url}")
 
     async def async_load(self, url: str) -> str:
         for loader in self.loaders:
@@ -40,4 +39,4 @@ class Compose(Loader):
             except Exception as e:
                 logger.info("[{}] Failed to load URL: {}, got error: {}", loader.__class__.__name__, url, e)
 
-        raise KabigonError(f"Failed to load URL: {url}")
+        raise Exception(f"Failed to load URL: {url}")

kabigon/firecrawl.py

@@ -2,8 +2,6 @@ import os
 
 from firecrawl import FirecrawlApp
 
-from .errors import FirecrawlError
-from .errors import FirecrawlKeyError
 from .loader import Loader
 
 
@@ -13,7 +11,7 @@ class FirecrawlLoader(Loader):
 
         api_key = os.getenv("FIRECRAWL_API_KEY")
         if not api_key:
-            raise FirecrawlKeyError()
+            raise ValueError("FIRECRAWL_API_KEY is not set.")
 
         self.app = FirecrawlApp(api_key=api_key)
 
@@ -25,7 +23,7 @@ class FirecrawlLoader(Loader):
         )
 
         if not result.success:
-            raise FirecrawlError(url, result.error)
+            raise Exception(f"Failed to load URL: {url}, got: {result.error}")
 
         return result.markdown
 

kabigon/httpx.py

@@ -3,21 +3,18 @@ import httpx
 from .loader import Loader
 from .utils import html_to_markdown
 
-DEFAULT_HEADERS = {
-    "Accept-Language": "zh-TW,zh;q=0.9,ja;q=0.8,en-US;q=0.7,en;q=0.6",
-    "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
-    "Cookie": "over18=1",  # Required for some sites like PTT
-}
-
 
 class HttpxLoader(Loader):
+    def __init__(self, headers: dict[str, str] | None = None) -> None:
+        self.headers = headers
+
     def load(self, url: str) -> str:
-        response = httpx.get(url, headers=DEFAULT_HEADERS, follow_redirects=True)
+        response = httpx.get(url, headers=self.headers, follow_redirects=False)
         response.raise_for_status()
         return html_to_markdown(response.content)
 
     async def async_load(self, url: str) -> str:
         async with httpx.AsyncClient() as client:
-            response = await client.get(url, headers=DEFAULT_HEADERS, follow_redirects=True)
+            response = await client.get(url, headers=self.headers, follow_redirects=True)
             response.raise_for_status()
             return html_to_markdown(response.content)

kabigon/pdf.py

@@ -6,7 +6,6 @@ from typing import Any
 import httpx
 from pypdf import PdfReader
 
-from .errors import NotPDFError
 from .loader import Loader
 
 DEFAULT_HEADERS = {
@@ -15,6 +14,11 @@ DEFAULT_HEADERS = {
 }
 
 
+class NotPDFError(Exception):
+    def __init__(self, url: str) -> None:
+        super().__init__(f"URL is not a PDF: {url}")
+
+
 class PDFLoader(Loader):
     def load(self, url_or_file: str) -> str:
         if not url_or_file.startswith("http"):

kabigon/ptt.py

@@ -0,0 +1,30 @@
+from urllib.parse import urlparse
+
+from .httpx import HttpxLoader
+from .loader import Loader
+
+
+def check_ptt_url(url: str) -> None:
+    if urlparse(url).netloc != "www.ptt.cc":
+        raise ValueError(f"URL must be from ptt.cc, got {url}")
+
+
+class PttLoader(Loader):
+    def __init__(self) -> None:
+        self.httpx_loader = HttpxLoader(
+            headers={
+                "Accept-Language": "zh-TW,zh;q=0.9,ja;q=0.8,en-US;q=0.7,en;q=0.6",
+                "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
+                "Cookie": "over18=1",
+            }
+        )
+
+    def load(self, url: str) -> str:
+        check_ptt_url(url)
+
+        return self.httpx_loader.load(url)
+
+    async def async_load(self, url: str):
+        check_ptt_url(url)
+
+        return await self.httpx_loader.async_load(url)

kabigon/reel.py

@@ -1,11 +1,11 @@
-from .errors import NotReelURLError
 from .httpx import HttpxLoader
 from .loader import Loader
 from .ytdlp import YtdlpLoader
 
 
-def is_reel_url(url: str) -> bool:
-    return url.startswith("https://www.instagram.com/reel")
+def check_reel_url(url: str) -> None:
+    if not url.startswith("https://www.instagram.com/reel"):
+        raise ValueError(f"URL is not an Instagram Reel: {url}")
 
 
 class ReelLoader(Loader):
@@ -14,8 +14,7 @@ class ReelLoader(Loader):
         self.ytdlp_loader = YtdlpLoader()
 
     def load(self, url: str) -> str:
-        if not is_reel_url(url):
-            raise NotReelURLError(url)
+        check_reel_url(url)
 
         audio_content = self.ytdlp_loader.load(url)
         html_content = self.httpx_loader.load(url)
@@ -23,8 +22,7 @@ class ReelLoader(Loader):
         return f"{audio_content}\n\n{html_content}"
 
     async def async_load(self, url: str):
-        if not is_reel_url(url):
-            raise NotReelURLError(url)
+        check_reel_url(url)
 
         audio_content = await self.ytdlp_loader.async_load(url)
         html_content = await self.httpx_loader.async_load(url)

kabigon/twitter.py

@@ -1,7 +1,6 @@
 from urllib.parse import urlparse
 from urllib.parse import urlunparse
 
-from .errors import NotTwitterURLError
 from .loader import Loader
 from .playwright import PlaywrightLoader
 
@@ -21,8 +20,9 @@ def replace_domain(url: str, new_domain: str = "x.com") -> str:
     return urlunparse(urlparse(url)._replace(netloc=new_domain))
 
 
-def is_x_url(url: str) -> bool:
-    return urlparse(url).netloc in TWITTER_DOMAINS
+def check_x_url(url: str) -> None:
+    if urlparse(url).netloc not in TWITTER_DOMAINS:
+        raise ValueError(f"URL is not a Twitter URL: {url}")
 
 
 class TwitterLoader(Loader):
@@ -30,16 +30,14 @@ class TwitterLoader(Loader):
         self.playwright_loader = PlaywrightLoader(wait_until="networkidle")
 
     def load(self, url: str) -> str:
-        if not is_x_url(url):
-            raise NotTwitterURLError(url)
+        check_x_url(url)
 
         url = replace_domain(url)
 
         return self.playwright_loader.load(url)
 
     async def async_load(self, url: str):
-        if not is_x_url(url):
-            raise NotTwitterURLError(url)
+        check_x_url(url)
 
         url = replace_domain(url)
 

{kabigon-0.7.0.dist-info → kabigon-0.8.1.dist-info}/METADATA

@@ -1,11 +1,10 @@
 Metadata-Version: 2.4
 Name: kabigon
-Version: 0.7.0
+Version: 0.8.1
 Author-email: narumi <toucans-cutouts0f@icloud.com>
 License-File: LICENSE
 Requires-Python: >=3.10
 Requires-Dist: aioytt>=0.2.4
-Requires-Dist: click>=8.1.8
 Requires-Dist: firecrawl-py>=2.4.1
 Requires-Dist: httpx>=0.28.1
 Requires-Dist: loguru>=0.7.3
@@ -15,6 +14,7 @@ Requires-Dist: playwright>=1.52.0
 Requires-Dist: pypdf>=5.3.0
 Requires-Dist: rich>=13.9.4
 Requires-Dist: timeout-decorator>=0.5.0
+Requires-Dist: typer>=0.15.3
 Requires-Dist: youtube-transcript-api>=0.6.3
 Requires-Dist: yt-dlp>=2025.4.30
 Description-Content-Type: text/markdown

kabigon-0.8.1.dist-info/RECORD

@@ -0,0 +1,20 @@
+kabigon/__init__.py,sha256=MUfTFUe5ezA249L2yuU5_2FiewLu86H3VsIpJSne2vQ,560
+kabigon/cli.py,sha256=PJ0wnwp_AgHA54YxGr1jNJ_q3ls7fEymgTJaJxCVU7M,650
+kabigon/compose.py,sha256=Kb6_-SNeh08QELMF-r3mWxasDTxJBuJJQamFPLcwQ1I,1463
+kabigon/firecrawl.py,sha256=-5AI9tla_684dtpubY_BRudqLgw28158WdwA1RjJvAA,778
+kabigon/httpx.py,sha256=Zup9DURyWLqoWzaxBbCYAaV-5LSlHUuAcNyyUsZTVag,696
+kabigon/loader.py,sha256=KhOJvlzLDM0o05o4VqMRgiyRVjofkzuGEcihQagj-8M,452
+kabigon/pdf.py,sha256=PNOx-Dz_VpN-EVnVar_wJJZdxWrDZrAFE-gSuUR9q7o,1870
+kabigon/playwright.py,sha256=MZ-r0Ej2wWAOJkDLwYRvO77wcDvh38KXz2wgDsCTgm0,1358
+kabigon/ptt.py,sha256=Gyp2nJrjptkjbwZJ9VEQHX0DEgKBe5QRQOmGVHUUgNA,896
+kabigon/py.typed,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+kabigon/reel.py,sha256=qOwWCvcp7xNKg0JDunq_Bsl8yqqMzrnAOI9k5mSqrOU,874
+kabigon/twitter.py,sha256=aRqAiFxIwln6lteWdoF6SmvbzO62yBTQRzcB7UcVJwk,1046
+kabigon/utils.py,sha256=eNTLtHLSB2erDac2HH3jWemgfr8Ou_ozwVb8h9BD-4g,922
+kabigon/youtube.py,sha256=HoiFNq0ookPL7_rO_wloBaY8yTIX6xP8A77F7y02q64,1166
+kabigon/ytdlp.py,sha256=_QRcyFx9s7NnI1MvcWdKKxlX-hHLnqtduCSL5_UH6dU,3140
+kabigon-0.8.1.dist-info/METADATA,sha256=8nZX3ukADj-qndGHMMuZluG_HdI8-wqNu8QJBoplPRc,1079
+kabigon-0.8.1.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+kabigon-0.8.1.dist-info/entry_points.txt,sha256=O3FYAO9w-NQvlGMJrBvtrnGHSK2QkUnQBTa30YXRbVE,45
+kabigon-0.8.1.dist-info/licenses/LICENSE,sha256=H2T3_RTgmcngMeC7p_SXT3GwBLkd2DaNgAZuxulcfiA,1066
+kabigon-0.8.1.dist-info/RECORD,,

kabigon/errors.py

@@ -1,28 +0,0 @@
-class KabigonError(Exception):
-    pass
-
-
-class FirecrawlKeyError(KabigonError):
-    def __init__(self) -> None:
-        super().__init__("FIRECRAWL_API_KEY is not set.")
-
-
-class FirecrawlError(KabigonError):
-    def __init__(self, url: str, error: str) -> None:
-        msg = f"Failed to load URL: {url}, got: {error}"
-        super().__init__(msg)
-
-
-class NotPDFError(KabigonError):
-    def __init__(self, url: str) -> None:
-        super().__init__(f"URL is not a PDF: {url}")
-
-
-class NotReelURLError(KabigonError):
-    def __init__(self, url: str):
-        super().__init__(f"URL is not an Instagram Reel: {url}")
-
-
-class NotTwitterURLError(KabigonError):
-    def __init__(self, url: str):
-        super().__init__(f"URL is not a Twitter URL: {url}")

kabigon-0.7.0.dist-info/RECORD

@@ -1,20 +0,0 @@
-kabigon/__init__.py,sha256=L-MDyX7PrPZvq6mASyOCJnATebwn-Lj1Nobp9GAPeyU,533
-kabigon/cli.py,sha256=co9IczXx0bAnXo6EEzEdCMSfpBetkPe6pB5hd2jtzA8,675
-kabigon/compose.py,sha256=l2D5OK91VcN2a6DbjMdwBk3YSqzVV7fOVX0TqNm2gJo,1502
-kabigon/errors.py,sha256=iri_YS71UsOHwaVtfy5IA6iUfq30DCsptZsChmZaZic,755
-kabigon/firecrawl.py,sha256=Xnrlhco_R58x5kwGy9ZCKTnVqS4Pp-D3G0u-qnuGEsU,800
-kabigon/httpx.py,sha256=uDdLks6zVzirY7-mnsJkypX86kAI5XmUVfK-lFifdJA,895
-kabigon/loader.py,sha256=KhOJvlzLDM0o05o4VqMRgiyRVjofkzuGEcihQagj-8M,452
-kabigon/pdf.py,sha256=yJcgkdMMF52baFsFxJp9Jn89KsTKBboLTLwD3gs6U5U,1775
-kabigon/playwright.py,sha256=MZ-r0Ej2wWAOJkDLwYRvO77wcDvh38KXz2wgDsCTgm0,1358
-kabigon/py.typed,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
-kabigon/reel.py,sha256=J2QOxGMYi_HaEscQPIipPEoHGN_iksGxR6pV_XvryME,929
-kabigon/twitter.py,sha256=U07pa8xA0nHAaaDPeUelQRvXR5ZnUvYJZW35xRAvHA8,1114
-kabigon/utils.py,sha256=eNTLtHLSB2erDac2HH3jWemgfr8Ou_ozwVb8h9BD-4g,922
-kabigon/youtube.py,sha256=HoiFNq0ookPL7_rO_wloBaY8yTIX6xP8A77F7y02q64,1166
-kabigon/ytdlp.py,sha256=_QRcyFx9s7NnI1MvcWdKKxlX-hHLnqtduCSL5_UH6dU,3140
-kabigon-0.7.0.dist-info/METADATA,sha256=4rng3Fe89ifNd6bQoBggRhCoqTMIa4sWhbU6N75puzo,1078
-kabigon-0.7.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-kabigon-0.7.0.dist-info/entry_points.txt,sha256=O3FYAO9w-NQvlGMJrBvtrnGHSK2QkUnQBTa30YXRbVE,45
-kabigon-0.7.0.dist-info/licenses/LICENSE,sha256=H2T3_RTgmcngMeC7p_SXT3GwBLkd2DaNgAZuxulcfiA,1066
-kabigon-0.7.0.dist-info/RECORD,,