forks/archivebox
1
0
Fork 0
Code Releases Activity

feat: Enable depth=1 functionality

Browse source
This commit is contained in:
Cristian 2020-07-07 10:07:44 -05:00
parent 2db0324539
commit 32e790979e
2 changed files with 16 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

16
archivebox/cli/archivebox_add.py
View file

@ -61,14 +61,14 @@ def main(args: Optional[List[str]]=None, stdin: Optional[IO]=None, pwd: Optional
index_only=command.index_only, index_only=command.index_only,
out_dir=pwd or OUTPUT_DIR, out_dir=pwd or OUTPUT_DIR,
) )
#if command.depth == 1: if command.depth == 1:
# add( add(
# import_str=None, import_str=None,
# import_path=command.import_path, import_path=command.import_path,
# update_all=command.update_all, update_all=command.update_all,
# index_only=command.index_only, index_only=command.index_only,
# out_dir=pwd or OUTPUT_DIR, out_dir=pwd or OUTPUT_DIR,
# ) )
if __name__ == '__main__': if __name__ == '__main__':

9
tests/test_args.py
View file

@ -12,4 +12,11 @@ def test_depth_flag_0_crawls_only_the_arg_page(tmp_path, process):
archived_item_path = list(tmp_path.glob('archive/**/*'))[0] archived_item_path = list(tmp_path.glob('archive/**/*'))[0]
with open(archived_item_path / "index.json", "r") as f: with open(archived_item_path / "index.json", "r") as f:
output_json = json.load(f) output_json = json.load(f)
assert output_json["base_url"] == "example.com" assert output_json["base_url"] == "example.com"
def test_depth_flag_1_crawls_the_page_AND_links(tmp_path, process):
arg_process = subprocess.run(["archivebox", "add", "https://example.com", "--depth=1"], capture_output=True)
with open(tmp_path / "index.json", "r") as f:
archive_file = f.read()
assert "https://example.com" in archive_file
assert "https://www.iana.org/domains/example" in archive_file