remove _wget (#61356)

lyuwenyu · web-flow · commit 4c0888d7b8f1 · 2024-02-04T13:00:41.000+08:00
* remove _wget

* remove _wget

* remove wget test
diff --git a/python/paddle/hapi/hub.py b/python/paddle/hapi/hub.py
@@ -117,7 +117,6 @@ def _get_cache_or_reload(repo, force_reload, verbose=True, source='github'):
             hub_dir,
             check_exist=not force_reload,
             decompress=False,
-            method=('wget' if source == 'gitee' else 'get'),
         )
         shutil.move(fpath, cached_file)
 
diff --git a/python/paddle/utils/download.py b/python/paddle/utils/download.py
@@ -15,14 +15,11 @@
 import hashlib
 import os
 import os.path as osp
-import shlex
 import shutil
-import subprocess
 import sys
 import tarfile
 import time
 import zipfile
-from urllib.parse import urlparse
 
 import httpx
 
@@ -198,40 +195,7 @@ def _get_download(url, fullname):
         return False
 
 
-def _wget_download(url: str, fullname: str):
-    try:
-        assert urlparse(url).scheme in (
-            'http',
-            'https',
-        ), 'Only support https and http url'
-        # using wget to download url
-        tmp_fullname = shlex.quote(fullname + "_tmp")
-        url = shlex.quote(url)
-        # –user-agent
-        command = f'wget -O {tmp_fullname} -t {DOWNLOAD_RETRY_LIMIT} {url}'
-        subprc = subprocess.Popen(
-            command, shell=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE
-        )
-        _ = subprc.communicate()
-
-        if subprc.returncode != 0:
-            raise RuntimeError(
-                f'{command} failed. Please make sure `wget` is installed or {url} exists'
-            )
-
-        shutil.move(tmp_fullname, fullname)
-
-    except Exception as e:  # requests.exceptions.ConnectionError
-        logger.info(f"Downloading {url} failed with exception {str(e)}")
-        return False
-
-    return fullname
-
-
-_download_methods = {
-    'get': _get_download,
-    'wget': _wget_download,
-}
+_download_methods = {'get': _get_download}
 
 
 def _download(url, path, md5sum=None, method='get'):
diff --git a/test/legacy_test/test_download.py b/test/legacy_test/test_download.py
@@ -120,14 +120,6 @@ def test_retry_exception(
                 './test',
             )
 
-    def test_wget_download_error(
-        self,
-    ):
-        with self.assertRaises(RuntimeError):
-            from paddle.utils.download import _download
-
-            _download('www.baidu', './test', method='wget')
-
     def test_download_methods(
         self,
     ):
@@ -136,14 +128,9 @@ def test_download_methods(
             "https://paddle-hapi.bj.bcebos.com/unittest/files.zip",
         ]
 
-        import sys
-
         from paddle.utils.download import _download
 
-        if sys.platform == 'linux':
-            methods = ['wget', 'get']
-        else:
-            methods = ['get']
+        methods = ['get']
 
         for url in urls:
             for method in methods:

Original file line number	Diff line number	Diff line change
`@@ -117,7 +117,6 @@ def _get_cache_or_reload(repo, force_reload, verbose=True, source='github'):`
`117`	`117`	`hub_dir,`
`118`	`118`	`check_exist=not force_reload,`
`119`	`119`	`decompress=False,`
`120`		`- method=('wget' if source == 'gitee' else 'get'),`
`121`	`120`	`)`
`122`	`121`	`shutil.move(fpath, cached_file)`
`123`	`122`