kabigon 0.8.7__py3-none-any.whl → 0.8.9__py3-none-any.whl

kabigon/playwright.py

@@ -26,7 +26,8 @@ class PlaywrightLoader(Loader):
 
         with sync_playwright() as p:
             browser = p.chromium.launch(headless=self.browser_headless)
-            page = browser.new_page()
+            context = browser.new_context()
+            page = context.new_page()
 
             page.goto(url, timeout=self.timeout, wait_until=self.wait_until)
 
@@ -46,7 +47,8 @@ class PlaywrightLoader(Loader):
 
         async with async_playwright() as p:
             browser = await p.chromium.launch(headless=self.browser_headless)
-            page = await browser.new_page()
+            context = await browser.new_context()
+            page = await context.new_page()
 
             await page.goto(url, timeout=self.timeout, wait_until=self.wait_until)
 

kabigon/youtube_ytdlp.py

@@ -1,12 +1,20 @@
 from urllib.parse import urlparse
 
+from aioytt.video_id import ALLOWED_NETLOCS
+from aioytt.video_id import ALLOWED_SCHEMES
+
 from .loader import Loader
 from .ytdlp import YtdlpLoader
 
 
 def check_youtube_url(url: str) -> None:
-    if urlparse(url).netloc not in ["youtube.com", "youtu.be"]:
-        raise ValueError(f"URL must be from youtube.com or youtu.be, got {url}")
+    schema = urlparse(url).scheme
+    if schema not in ALLOWED_SCHEMES:
+        raise ValueError(f"URL scheme is not allowed: {schema}")
+
+    domain = urlparse(url).netloc
+    if domain not in ALLOWED_NETLOCS:
+        raise ValueError(f"URL domain is not allowed: {domain}")
 
 
 class YoutubeYtdlpLoader(Loader):

kabigon/ytdlp.py

@@ -1,6 +1,7 @@
 import hashlib
 import os
 import tempfile
+from pathlib import Path
 
 import yt_dlp
 from loguru import logger
@@ -8,9 +9,7 @@ from loguru import logger
 from .loader import Loader
 
 
-def download_audio(url: str) -> str:
-    filename = os.path.join(tempfile.gettempdir(), hashlib.sha512(url.encode("utf-8")).hexdigest())
-
+def download_audio(url: str, outtmpl: str | None = None) -> None:
     ydl_opts = {
         "format": "bestaudio/best",
         "postprocessors": [
@@ -20,16 +19,19 @@ def download_audio(url: str) -> str:
                 "preferredquality": "192",
             }
         ],
-        "outtmpl": filename,
-        "ffmpeg_location": os.getenv("FFMPEG_PATH", "ffmpeg"),
         "match_filter": yt_dlp.match_filter_func(["!is_live"]),
     }
 
+    if outtmpl is not None:
+        ydl_opts["outtmpl"] = outtmpl
+
+    ffmpeg_path = os.getenv("FFMPEG_PATH")
+    if ffmpeg_path is None:
+        ydl_opts["ffmpeg_location"] = ffmpeg_path
+
     with yt_dlp.YoutubeDL(ydl_opts) as ydl:
         ydl.download([url])
 
-    return filename + ".mp3"
-
 
 class YtdlpLoader(Loader):
     def __init__(self, model: str = "tiny") -> None:
@@ -44,12 +46,16 @@ class YtdlpLoader(Loader):
         self.load_audio = whisper.load_audio
 
     def load(self, url: str) -> str:
-        audio_file = download_audio(url)
-        audio = self.load_audio(audio_file)
+        temp_dir = Path(tempfile.gettempdir())
+        outtmpl = hashlib.sha512(url.encode("utf-8")).hexdigest()
+        path = (temp_dir / outtmpl).with_suffix(".mp3")
+
+        download_audio(url, outtmpl=outtmpl)
+        audio = self.load_audio(path)
 
         # Clean up the audio file
-        os.remove(audio_file)
+        os.remove(path)
 
-        logger.info("Transcribing audio file: {}", audio_file)
+        logger.info("Transcribing audio file: {}", path)
         result = self.model.transcribe(audio)
         return result.get("text", "")

{kabigon-0.8.7.dist-info → kabigon-0.8.9.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: kabigon
-Version: 0.8.7
+Version: 0.8.9
 Author-email: narumi <toucans-cutouts0f@icloud.com>
 License-File: LICENSE
 Requires-Python: >=3.10

{kabigon-0.8.7.dist-info → kabigon-0.8.9.dist-info}/RECORD

@@ -5,17 +5,17 @@ kabigon/firecrawl.py,sha256=-5AI9tla_684dtpubY_BRudqLgw28158WdwA1RjJvAA,778
 kabigon/httpx.py,sha256=Zup9DURyWLqoWzaxBbCYAaV-5LSlHUuAcNyyUsZTVag,696
 kabigon/loader.py,sha256=W2ZdSINQKjG-5XNYweM8lIl19JA4wwd7AE2pueHeO70,380
 kabigon/pdf.py,sha256=PNOx-Dz_VpN-EVnVar_wJJZdxWrDZrAFE-gSuUR9q7o,1870
-kabigon/playwright.py,sha256=RpP8qxSZe0p4j8hdRWbK7jJO9DNhIlYn77mXBDMAC3c,1865
+kabigon/playwright.py,sha256=etzxzAmwpATnB_ckxGU3IE4jgKBdlFDwtP1fN0u4Pr8,1959
 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=F9GpLa0iUy03wYU94RrrnrXa6ExqbG6CZpqx5bPENWE,1106
-kabigon/youtube_ytdlp.py,sha256=-kElnrP4cLhZz2kNtgp2-ijA8I-IpnRctZ1JNCDN7pw,503
-kabigon/ytdlp.py,sha256=OGVrKRGXY_eHePgDmN52Kkui12NN44LsIPonD_61QBM,1501
-kabigon-0.8.7.dist-info/METADATA,sha256=t5t54E7Mrl9AOmKoKp1CuJbWWRS2gr6jwrEwIX15tjc,1264
-kabigon-0.8.7.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-kabigon-0.8.7.dist-info/entry_points.txt,sha256=O3FYAO9w-NQvlGMJrBvtrnGHSK2QkUnQBTa30YXRbVE,45
-kabigon-0.8.7.dist-info/licenses/LICENSE,sha256=H2T3_RTgmcngMeC7p_SXT3GwBLkd2DaNgAZuxulcfiA,1066
-kabigon-0.8.7.dist-info/RECORD,,
+kabigon/youtube_ytdlp.py,sha256=Y6h55wYF-5PdxLRxsznFtJDypCedD8g-hY9fgCnfvbg,722
+kabigon/ytdlp.py,sha256=ex_at_1irwHgcl20Mm0mRZp-rTM92OHnVC5dWQCkI1c,1677
+kabigon-0.8.9.dist-info/METADATA,sha256=T8EHDM6dcghksnn6mOQ1548AvmYBfYlCefo4eIF9DBo,1264
+kabigon-0.8.9.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+kabigon-0.8.9.dist-info/entry_points.txt,sha256=O3FYAO9w-NQvlGMJrBvtrnGHSK2QkUnQBTa30YXRbVE,45
+kabigon-0.8.9.dist-info/licenses/LICENSE,sha256=H2T3_RTgmcngMeC7p_SXT3GwBLkd2DaNgAZuxulcfiA,1066
+kabigon-0.8.9.dist-info/RECORD,,