[bugfix] download_needed param error in kernels_output (#629)

When I use the "kaggle kernel output" to download files, it always
re-downloads files that already exist. Additionally, I found some issues
in the kernels_output method.

1) When calling download_needed, the first parameter is url, but it
expects a response object instead.

2) After fixing issue 1, the method does skip existing files. However,
it’s still slow, and system monitoring tools show high network usage,
indicating that the files are actually being downloaded again. The
download_needed method only requires response.headers, but requests.get
downloads the entire file by default. This can be fixed by setting the
stream=True parameter to change the download behavior.

Author: memset-c

kaggle/api/kaggle_api_extended.py

@@ -2616,8 +2616,8 @@ def kernels_output(self, kernel, path, force=False, quiet=True):
         for item in response['files']:
             outfile = os.path.join(target_dir, item['fileName'])
             outfiles.append(outfile)
-            download_response = requests.get(item['url'])
-            if force or self.download_needed(item, outfile, quiet):
+            download_response = requests.get(item['url'], stream=True)
+            if force or self.download_needed(download_response, outfile, quiet):
                 os.makedirs(os.path.split(outfile)[0], exist_ok=True)
                 with open(outfile, 'wb') as out:
                     out.write(download_response.content)

src/kaggle/api/kaggle_api_extended.py

@@ -2600,8 +2600,8 @@ def kernels_output(self, kernel, path, force=False, quiet=True):
         for item in response['files']:
             outfile = os.path.join(target_dir, item['fileName'])
             outfiles.append(outfile)
-            download_response = requests.get(item['url'])
-            if force or self.download_needed(item, outfile, quiet):
+            download_response = requests.get(item['url'], stream=True)
+            if force or self.download_needed(download_response, outfile, quiet):
                 os.makedirs(os.path.split(outfile)[0], exist_ok=True)
                 with open(outfile, 'wb') as out:
                     out.write(download_response.content)