mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-13 14:44:29 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

fix: Add change to calculate wget folder when there is a port present

Browse source
This commit is contained in:
Cristian 2020-07-17 16:55:56 -05:00
parent 58ac44c867
commit 23e6803f02
6 changed files with 19 additions and 16 deletions
Download patch file Download diff file Expand all files Collapse all files

16
tests/test_args.py
View file

@ -4,25 +4,25 @@ import json
from .fixtures import *
def test_depth_flag_is_accepted(process):
arg_process = subprocess.run(["archivebox", "add", "http://localhost:8080/static/example.com.html", "--depth=0"], capture_output=True)
arg_process = subprocess.run(["archivebox", "add", "http://127.0.0.1:8080/static/example.com.html", "--depth=0"], capture_output=True)
assert 'unrecognized arguments: --depth' not in arg_process.stderr.decode("utf-8")
def test_depth_flag_fails_if_it_is_not_0_or_1(process):
arg_process = subprocess.run(["archivebox", "add", "http://localhost:8080/static/example.com.html", "--depth=5"], capture_output=True)
arg_process = subprocess.run(["archivebox", "add", "http://127.0.0.1:8080/static/example.com.html", "--depth=5"], capture_output=True)
assert 'invalid choice' in arg_process.stderr.decode("utf-8")
arg_process = subprocess.run(["archivebox", "add", "http://localhost:8080/static/example.com.html", "--depth=-1"], capture_output=True)
arg_process = subprocess.run(["archivebox", "add", "http://127.0.0.1:8080/static/example.com.html", "--depth=-1"], capture_output=True)
assert 'invalid choice' in arg_process.stderr.decode("utf-8")
def test_depth_flag_0_crawls_only_the_arg_page(tmp_path, process):
arg_process = subprocess.run(["archivebox", "add", "http://localhost:8080/static/example.com.html", "--depth=0"], capture_output=True)
arg_process = subprocess.run(["archivebox", "add", "http://127.0.0.1:8080/static/example.com.html", "--depth=0"], capture_output=True)
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"] == "localhost:8080/static/example.com.html"
assert output_json["base_url"] == "127.0.0.1:8080/static/example.com.html"
def test_depth_flag_1_crawls_the_page_AND_links(tmp_path, process):
arg_process = subprocess.run(["archivebox", "add", "http://localhost:8080/static/example.com.html", "--depth=1"], capture_output=True)
arg_process = subprocess.run(["archivebox", "add", "http://127.0.0.1:8080/static/example.com.html", "--depth=1"], capture_output=True)
with open(tmp_path / "index.json", "r") as f:
archive_file = f.read()
assert "http://localhost:8080/static/example.com.html" in archive_file
assert "http://localhost:8080/static/iana.org.html" in archive_file
assert "http://127.0.0.1:8080/static/example.com.html" in archive_file
assert "http://127.0.0.1:8080/static/iana.org.html" in archive_file