2021-02-26 19:07:26 +00:00
|
|
|
#!/usr/bin/python3
|
|
|
|
|
|
|
|
|
|
import time
|
|
|
|
|
from flask import url_for
|
2023-06-19 21:29:13 +00:00
|
|
|
from .util import live_server_setup, wait_for_all_checks
|
2022-02-23 09:49:25 +00:00
|
|
|
from changedetectionio import html_tools
|
2021-02-26 19:07:26 +00:00
|
|
|
|
2021-05-08 01:29:41 +00:00
|
|
|
def test_setup(live_server):
|
|
|
|
|
live_server_setup(live_server)
|
2021-02-26 19:07:26 +00:00
|
|
|
|
|
|
|
|
# Unit test of the stripper
|
2021-02-27 19:30:06 +00:00
|
|
|
# Always we are dealing in utf-8
|
2021-02-26 19:07:26 +00:00
|
|
|
def test_strip_text_func():
|
2023-03-18 19:36:26 +00:00
|
|
|
from ..processors import text_json_diff as fetch_site_status
|
|
|
|
|
|
2021-02-26 19:07:26 +00:00
|
|
|
|
|
|
|
|
test_content = """
|
|
|
|
|
Some content
|
|
|
|
|
is listed here
|
2021-02-27 08:05:25 +00:00
|
|
|
|
2021-02-26 19:07:26 +00:00
|
|
|
but sometimes we want to remove the lines.
|
2021-02-27 08:05:25 +00:00
|
|
|
|
2021-02-26 19:07:26 +00:00
|
|
|
but not always."""
|
|
|
|
|
|
|
|
|
|
ignore_lines = ["sometimes"]
|
|
|
|
|
|
2021-02-27 19:30:06 +00:00
|
|
|
fetcher = fetch_site_status.perform_site_check(datastore=False)
|
2022-02-23 09:49:25 +00:00
|
|
|
stripped_content = html_tools.strip_ignore_text(test_content, ignore_lines)
|
2021-02-26 19:07:26 +00:00
|
|
|
|
2021-02-27 19:30:06 +00:00
|
|
|
assert b"sometimes" not in stripped_content
|
|
|
|
|
assert b"Some content" in stripped_content
|
2021-02-26 19:07:26 +00:00
|
|
|
|
|
|
|
|
|
2021-02-27 08:05:25 +00:00
|
|
|
def set_original_ignore_response():
|
2021-02-26 19:07:26 +00:00
|
|
|
test_return_data = """<html>
|
|
|
|
|
<body>
|
2023-03-12 16:05:34 +00:00
|
|
|
Some initial text<br>
|
2021-02-26 19:07:26 +00:00
|
|
|
<p>Which is across multiple lines</p>
|
2023-03-12 16:05:34 +00:00
|
|
|
<br>
|
|
|
|
|
So let's see what happens. <br>
|
2021-02-26 19:07:26 +00:00
|
|
|
</body>
|
|
|
|
|
</html>
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
|
2021-08-12 10:05:59 +00:00
|
|
|
with open("test-datastore/endpoint-content.txt", "w") as f:
|
2021-02-26 19:07:26 +00:00
|
|
|
f.write(test_return_data)
|
|
|
|
|
|
|
|
|
|
|
2021-02-27 19:30:06 +00:00
|
|
|
def set_modified_original_ignore_response():
|
|
|
|
|
test_return_data = """<html>
|
|
|
|
|
<body>
|
2023-03-12 16:05:34 +00:00
|
|
|
Some NEW nice initial text<br>
|
2021-02-27 19:30:06 +00:00
|
|
|
<p>Which is across multiple lines</p>
|
2023-03-12 16:05:34 +00:00
|
|
|
<br>
|
|
|
|
|
So let's see what happens. <br>
|
2022-02-23 09:49:25 +00:00
|
|
|
<p>new ignore stuff</p>
|
|
|
|
|
<p>blah</p>
|
2021-02-27 19:30:06 +00:00
|
|
|
</body>
|
|
|
|
|
</html>
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
|
2021-08-12 10:05:59 +00:00
|
|
|
with open("test-datastore/endpoint-content.txt", "w") as f:
|
2021-02-27 19:30:06 +00:00
|
|
|
f.write(test_return_data)
|
|
|
|
|
|
|
|
|
|
|
2021-02-26 19:07:26 +00:00
|
|
|
# Is the same but includes ZZZZZ, 'ZZZZZ' is the last line in ignore_text
|
2021-02-27 08:05:25 +00:00
|
|
|
def set_modified_ignore_response():
|
2021-02-26 19:07:26 +00:00
|
|
|
test_return_data = """<html>
|
|
|
|
|
<body>
|
2023-03-12 16:05:34 +00:00
|
|
|
Some initial text<br>
|
2021-02-26 19:07:26 +00:00
|
|
|
<p>Which is across multiple lines</p>
|
2022-02-23 11:01:11 +00:00
|
|
|
<P>ZZZZz</P>
|
2023-03-12 16:05:34 +00:00
|
|
|
<br>
|
|
|
|
|
So let's see what happens. <br>
|
2021-02-26 19:07:26 +00:00
|
|
|
</body>
|
|
|
|
|
</html>
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
|
2021-08-12 10:05:59 +00:00
|
|
|
with open("test-datastore/endpoint-content.txt", "w") as f:
|
2021-02-26 19:07:26 +00:00
|
|
|
f.write(test_return_data)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def test_check_ignore_text_functionality(client, live_server):
|
|
|
|
|
|
2022-02-23 11:01:11 +00:00
|
|
|
# Use a mix of case in ZzZ to prove it works case-insensitive.
|
|
|
|
|
ignore_text = "XXXXX\r\nYYYYY\r\nzZzZZ\r\nnew ignore stuff"
|
2021-02-27 08:05:25 +00:00
|
|
|
set_original_ignore_response()
|
2021-02-26 19:07:26 +00:00
|
|
|
|
|
|
|
|
# Give the endpoint time to spin up
|
|
|
|
|
time.sleep(1)
|
|
|
|
|
|
|
|
|
|
# Add our URL to the import page
|
2021-02-27 08:05:25 +00:00
|
|
|
test_url = url_for('test_endpoint', _external=True)
|
2021-02-26 19:07:26 +00:00
|
|
|
res = client.post(
|
|
|
|
|
url_for("import_page"),
|
|
|
|
|
data={"urls": test_url},
|
|
|
|
|
follow_redirects=True
|
|
|
|
|
)
|
|
|
|
|
assert b"1 Imported" in res.data
|
|
|
|
|
|
2021-02-27 21:09:25 +00:00
|
|
|
# Give the thread time to pick it up
|
2023-06-19 21:29:13 +00:00
|
|
|
wait_for_all_checks(client)
|
2021-02-27 21:09:25 +00:00
|
|
|
|
2021-02-26 19:07:26 +00:00
|
|
|
# Goto the edit page, add our ignore text
|
|
|
|
|
# Add our URL to the import page
|
|
|
|
|
res = client.post(
|
|
|
|
|
url_for("edit_page", uuid="first"),
|
2021-08-12 10:05:59 +00:00
|
|
|
data={"ignore_text": ignore_text, "url": test_url, 'fetch_backend': "html_requests"},
|
2021-02-26 19:07:26 +00:00
|
|
|
follow_redirects=True
|
|
|
|
|
)
|
|
|
|
|
assert b"Updated watch." in res.data
|
|
|
|
|
|
|
|
|
|
# Check it saved
|
|
|
|
|
res = client.get(
|
|
|
|
|
url_for("edit_page", uuid="first"),
|
|
|
|
|
)
|
|
|
|
|
assert bytes(ignore_text.encode('utf-8')) in res.data
|
|
|
|
|
|
|
|
|
|
# Trigger a check
|
2022-05-20 14:27:51 +00:00
|
|
|
client.get(url_for("form_watch_checknow"), follow_redirects=True)
|
2021-02-26 19:07:26 +00:00
|
|
|
|
|
|
|
|
# Give the thread time to pick it up
|
2023-06-19 21:29:13 +00:00
|
|
|
wait_for_all_checks(client)
|
2021-02-26 19:07:26 +00:00
|
|
|
|
|
|
|
|
# It should report nothing found (no new 'unviewed' class)
|
|
|
|
|
res = client.get(url_for("index"))
|
|
|
|
|
assert b'unviewed' not in res.data
|
2021-02-27 08:05:25 +00:00
|
|
|
assert b'/test-endpoint' in res.data
|
2021-02-26 19:07:26 +00:00
|
|
|
|
2021-02-27 19:30:06 +00:00
|
|
|
# Make a change
|
2021-02-27 08:05:25 +00:00
|
|
|
set_modified_ignore_response()
|
2021-02-26 19:07:26 +00:00
|
|
|
|
|
|
|
|
# Trigger a check
|
2022-05-20 14:27:51 +00:00
|
|
|
client.get(url_for("form_watch_checknow"), follow_redirects=True)
|
2021-02-26 19:07:26 +00:00
|
|
|
# Give the thread time to pick it up
|
2023-06-19 21:29:13 +00:00
|
|
|
wait_for_all_checks(client)
|
2021-02-26 19:07:26 +00:00
|
|
|
|
|
|
|
|
# It should report nothing found (no new 'unviewed' class)
|
|
|
|
|
res = client.get(url_for("index"))
|
|
|
|
|
assert b'unviewed' not in res.data
|
2021-02-27 08:05:25 +00:00
|
|
|
assert b'/test-endpoint' in res.data
|
2021-02-26 19:07:26 +00:00
|
|
|
|
2022-02-23 09:49:25 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-02-27 19:30:06 +00:00
|
|
|
# Just to be sure.. set a regular modified change..
|
|
|
|
|
set_modified_original_ignore_response()
|
2022-05-20 14:27:51 +00:00
|
|
|
client.get(url_for("form_watch_checknow"), follow_redirects=True)
|
2023-06-19 21:29:13 +00:00
|
|
|
wait_for_all_checks(client)
|
2022-02-23 09:49:25 +00:00
|
|
|
|
2021-02-27 19:30:06 +00:00
|
|
|
res = client.get(url_for("index"))
|
|
|
|
|
assert b'unviewed' in res.data
|
|
|
|
|
|
2022-02-23 09:49:25 +00:00
|
|
|
# Check the preview/highlighter, we should be able to see what we ignored, but it should be highlighted
|
|
|
|
|
# We only introduce the "modified" content that includes what we ignore so we can prove the newest version also displays
|
|
|
|
|
# at /preview
|
|
|
|
|
res = client.get(url_for("preview_page", uuid="first"))
|
|
|
|
|
# We should be able to see what we ignored
|
|
|
|
|
assert b'<div class="ignored">new ignore stuff' in res.data
|
|
|
|
|
|
2022-05-20 14:27:51 +00:00
|
|
|
res = client.get(url_for("form_delete", uuid="all"), follow_redirects=True)
|
2021-02-27 08:37:40 +00:00
|
|
|
assert b'Deleted' in res.data
|
2022-01-01 13:53:08 +00:00
|
|
|
|
|
|
|
|
def test_check_global_ignore_text_functionality(client, live_server):
|
|
|
|
|
|
2022-04-24 14:56:32 +00:00
|
|
|
# Give the endpoint time to spin up
|
|
|
|
|
time.sleep(1)
|
|
|
|
|
|
2022-01-01 13:53:08 +00:00
|
|
|
ignore_text = "XXXXX\r\nYYYYY\r\nZZZZZ"
|
|
|
|
|
set_original_ignore_response()
|
|
|
|
|
|
2022-04-24 14:56:32 +00:00
|
|
|
# Goto the settings page, add our ignore text
|
|
|
|
|
res = client.post(
|
|
|
|
|
url_for("settings_page"),
|
|
|
|
|
data={
|
|
|
|
|
"requests-time_between_check-minutes": 180,
|
|
|
|
|
"application-global_ignore_text": ignore_text,
|
|
|
|
|
'application-fetch_backend': "html_requests"
|
|
|
|
|
},
|
|
|
|
|
follow_redirects=True
|
|
|
|
|
)
|
|
|
|
|
assert b"Settings updated." in res.data
|
|
|
|
|
|
2022-01-01 13:53:08 +00:00
|
|
|
|
|
|
|
|
# Add our URL to the import page
|
|
|
|
|
test_url = url_for('test_endpoint', _external=True)
|
|
|
|
|
res = client.post(
|
|
|
|
|
url_for("import_page"),
|
|
|
|
|
data={"urls": test_url},
|
|
|
|
|
follow_redirects=True
|
|
|
|
|
)
|
|
|
|
|
assert b"1 Imported" in res.data
|
|
|
|
|
|
|
|
|
|
# Give the thread time to pick it up
|
2023-06-19 21:29:13 +00:00
|
|
|
wait_for_all_checks(client)
|
2022-01-01 13:53:08 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
# Goto the edit page of the item, add our ignore text
|
|
|
|
|
# Add our URL to the import page
|
|
|
|
|
res = client.post(
|
|
|
|
|
url_for("edit_page", uuid="first"),
|
|
|
|
|
data={"ignore_text": "something irrelevent but just to check", "url": test_url, 'fetch_backend': "html_requests"},
|
|
|
|
|
follow_redirects=True
|
|
|
|
|
)
|
|
|
|
|
assert b"Updated watch." in res.data
|
|
|
|
|
|
|
|
|
|
# Check it saved
|
|
|
|
|
res = client.get(
|
|
|
|
|
url_for("settings_page"),
|
|
|
|
|
)
|
|
|
|
|
assert bytes(ignore_text.encode('utf-8')) in res.data
|
|
|
|
|
|
|
|
|
|
# Trigger a check
|
2022-05-20 14:27:51 +00:00
|
|
|
client.get(url_for("form_watch_checknow"), follow_redirects=True)
|
2022-01-01 13:53:08 +00:00
|
|
|
|
|
|
|
|
# Give the thread time to pick it up
|
2023-06-19 21:29:13 +00:00
|
|
|
wait_for_all_checks(client)
|
2022-01-01 13:53:08 +00:00
|
|
|
|
2022-04-24 14:56:32 +00:00
|
|
|
# so that we are sure everything is viewed and in a known 'nothing changed' state
|
|
|
|
|
res = client.get(url_for("diff_history_page", uuid="first"))
|
|
|
|
|
|
2022-01-01 13:53:08 +00:00
|
|
|
# It should report nothing found (no new 'unviewed' class)
|
|
|
|
|
res = client.get(url_for("index"))
|
|
|
|
|
assert b'unviewed' not in res.data
|
|
|
|
|
assert b'/test-endpoint' in res.data
|
|
|
|
|
|
2022-04-24 14:56:32 +00:00
|
|
|
|
|
|
|
|
# Make a change which includes the ignore text
|
2022-01-01 13:53:08 +00:00
|
|
|
set_modified_ignore_response()
|
|
|
|
|
|
|
|
|
|
# Trigger a check
|
2022-05-20 14:27:51 +00:00
|
|
|
client.get(url_for("form_watch_checknow"), follow_redirects=True)
|
2022-01-01 13:53:08 +00:00
|
|
|
# Give the thread time to pick it up
|
2023-06-19 21:29:13 +00:00
|
|
|
wait_for_all_checks(client)
|
2022-01-01 13:53:08 +00:00
|
|
|
|
|
|
|
|
# It should report nothing found (no new 'unviewed' class)
|
|
|
|
|
res = client.get(url_for("index"))
|
|
|
|
|
assert b'unviewed' not in res.data
|
|
|
|
|
assert b'/test-endpoint' in res.data
|
|
|
|
|
|
2022-04-24 14:56:32 +00:00
|
|
|
# Just to be sure.. set a regular modified change that will trigger it
|
2022-01-01 13:53:08 +00:00
|
|
|
set_modified_original_ignore_response()
|
2022-05-20 14:27:51 +00:00
|
|
|
client.get(url_for("form_watch_checknow"), follow_redirects=True)
|
2023-06-19 21:29:13 +00:00
|
|
|
wait_for_all_checks(client)
|
2022-01-01 13:53:08 +00:00
|
|
|
res = client.get(url_for("index"))
|
|
|
|
|
assert b'unviewed' in res.data
|
|
|
|
|
|
2022-05-20 14:27:51 +00:00
|
|
|
res = client.get(url_for("form_delete", uuid="all"), follow_redirects=True)
|
2022-01-02 21:28:34 +00:00
|
|
|
assert b'Deleted' in res.data
|
