anyio

Author: zzstoatzz

examples/refresh_chroma/refresh_collection.py

@@ -16,14 +16,7 @@
 def html_parser(html: str) -> str:
     import trafilatura
 
-    trafilatura_config = trafilatura.settings.use_config()
-    # disable signal, so it can run in a worker thread
-    # https://github.com/adbar/trafilatura/issues/202
-    trafilatura_config.set("DEFAULT", "EXTRACTION_TIMEOUT", "0")
-    return (
-        trafilatura.extract(html, config=trafilatura_config)
-        or BeautifulSoup(html, "html.parser").get_text()
-    )
+    return trafilatura.extract(html) or BeautifulSoup(html, "html.parser").get_text()
 
 
 raggy.settings.html_parser = html_parser

pyproject.toml

@@ -15,7 +15,6 @@ keywords = ["scraping", "nlp", "ai", "document-processing"]
 requires-python = ">=3.10"
 dependencies = [
     "aiofiles",
-    "anyio<4.0.0",
     "bs4",
     "chardet",
     "fake-useragent",
@@ -86,11 +85,7 @@ asyncio_mode = 'auto'
 filterwarnings = [
     "ignore:'crypt' is deprecated and slated for removal in Python 3.13:DeprecationWarning",
 ]
-env = [
-    'D:RAGGY_LOG_VERBOSE=1',
-    'D:RAGGY_LOG_LEVEL=DEBUG',
-    'PYTEST_TIMEOUT=20',
-]
+env = ['D:RAGGY_LOG_VERBOSE=1', 'D:RAGGY_LOG_LEVEL=DEBUG', 'PYTEST_TIMEOUT=20']
 
 [tool.ruff]
 extend-select = ["I"]
@@ -106,4 +101,4 @@ skip-magic-trailing-comma = false
 'tests/fixtures/*.py' = ['F403']
 
 [tool.setuptools_scm]
-write_to = "src/raggy/_version.py"
+write_to = "src/raggy/_version.py"