Allow character detection dependencies to be optional in post-packaging steps

This commit is contained in:
Nate Prewitt
2024-05-14 14:59:26 -07:00
parent d6dded3f00
commit 555b870eb1
5 changed files with 60 additions and 23 deletions

View File

@@ -37,3 +37,23 @@ jobs:
- name: Run tests
run: |
make ci
no_chardet:
name: "No Character Detection"
runs-on: ubuntu-latest
strategy:
fail-fast: true
steps:
- uses: actions/checkout@8ade135a41bc03ea155e62e844d188df1ea18608
- name: 'Set up Python 3.8'
uses: actions/setup-python@82c7e631bb3cdc910f68e0081d67478d79c6982d
with:
python-version: '3.8'
- name: Install dependencies
run: |
make
python -m pip uninstall -y "charset_normalizer" "chardet"
- name: Run tests
run: |
make ci

View File

@@ -83,7 +83,11 @@ def check_compatibility(urllib3_version, chardet_version, charset_normalizer_ver
# charset_normalizer >= 2.0.0 < 4.0.0
assert (2, 0, 0) <= (major, minor, patch) < (4, 0, 0)
else:
raise Exception("You need either charset_normalizer or chardet installed")
warnings.warn(
"Unable to find acceptable character detection dependency "
"(chardet or charset_normalizer).",
RequestsDependencyWarning,
)
def _check_cryptography(cryptography_version):

View File

@@ -7,13 +7,28 @@ between Python 2 and Python 3. It remains for backwards
compatibility until the next major version.
"""
try:
import chardet
except ImportError:
import charset_normalizer as chardet
import importlib
import sys
# -------------------
# Character Detection
# -------------------
def _resolve_char_detection():
"""Find supported character detection libraries."""
chardet = None
for lib in ("chardet", "charset_normalizer"):
if chardet is None:
try:
chardet = importlib.import_module(lib)
except ImportError:
pass
return chardet
chardet = _resolve_char_detection()
# -------
# Pythons
# -------

View File

@@ -789,7 +789,12 @@ class Response:
@property
def apparent_encoding(self):
"""The apparent encoding, provided by the charset_normalizer or chardet libraries."""
return chardet.detect(self.content)["encoding"]
if chardet is not None:
return chardet.detect(self.content)["encoding"]
else:
# If no character detection library is available, we'll fall back
# to a standard Python utf-8 str.
return "utf-8"
def iter_content(self, chunk_size=1, decode_unicode=False):
"""Iterates over the response data. When stream=True is set on the

View File

@@ -1,13 +1,6 @@
import sys
try:
import chardet
except ImportError:
import warnings
import charset_normalizer as chardet
warnings.filterwarnings("ignore", "Trying to detect", module="charset_normalizer")
from .compat import chardet
# This code exists for backwards compatibility reasons.
# I don't like it either. Just look the other way. :)
@@ -20,11 +13,11 @@ for package in ("urllib3", "idna"):
if mod == package or mod.startswith(f"{package}."):
sys.modules[f"requests.packages.{mod}"] = sys.modules[mod]
target = chardet.__name__
for mod in list(sys.modules):
if mod == target or mod.startswith(f"{target}."):
imported_mod = sys.modules[mod]
sys.modules[f"requests.packages.{mod}"] = imported_mod
mod = mod.replace(target, "chardet")
sys.modules[f"requests.packages.{mod}"] = imported_mod
# Kinda cool, though, right?
if chardet is not None:
target = chardet.__name__
for mod in list(sys.modules):
if mod == target or mod.startswith(f"{target}."):
imported_mod = sys.modules[mod]
sys.modules[f"requests.packages.{mod}"] = imported_mod
mod = mod.replace(target, "chardet")
sys.modules[f"requests.packages.{mod}"] = imported_mod