mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-16 16:14:28 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

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
Download patch file Download diff file Expand all files Collapse all files

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]
with open(archived_item_path / "index.json", "r") as 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