#!/usr/bin/env python3
import time
from flask import url_for
from ..html_tools import *
from .util import live_server_setup, wait_for_all_checks
def test_setup(live_server):
live_server_setup(live_server)
def set_original_response():
test_return_data = """
Header
Some initial text
Which is across multiple lines
So let's see what happens.
Some text that will change
"""
with open("test-datastore/endpoint-content.txt", "w") as f:
f.write(test_return_data)
def set_modified_response():
test_return_data = """
Header changed
Some initial text
Which is across multiple lines
So let's see what happens.
Some text that changes
"""
with open("test-datastore/endpoint-content.txt", "w") as f:
f.write(test_return_data)
def test_element_removal_output():
from inscriptis import get_text
# Check text with sub-parts renders correctly
content = """
Header
Some initial text
across multiple lines
Some text that changes
Some text should be matched by xPath // selector
Some text should be matched by xPath selector
Some text should be matched by xPath1 selector
"""
html_blob = element_removal(
[
"header",
"footer",
"nav",
"#changetext",
"//*[contains(text(), 'xPath // selector')]",
"xpath://*[contains(text(), 'xPath selector')]",
"xpath1://*[contains(text(), 'xPath1 selector')]"
],
html_content=content
)
text = get_text(html_blob)
assert (
text
== """Some initial text
across multiple lines
"""
)
def test_element_removal_full(client, live_server, measure_memory_usage):
#live_server_setup(live_server)
set_original_response()
# 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
wait_for_all_checks(client)
# Goto the edit page, add the filter data
# Not sure why \r needs to be added - absent of the #changetext this is not necessary
subtractive_selectors_data = "header\r\nfooter\r\nnav\r\n#changetext"
res = client.post(
url_for("edit_page", uuid="first"),
data={
"subtractive_selectors": subtractive_selectors_data,
"url": test_url,
"tags": "",
"headers": "",
"fetch_backend": "html_requests",
},
follow_redirects=True,
)
assert b"Updated watch." in res.data
wait_for_all_checks(client)
# Check it saved
res = client.get(
url_for("edit_page", uuid="first"),
)
assert bytes(subtractive_selectors_data.encode("utf-8")) in res.data
# Trigger a check
res = client.get(url_for("form_watch_checknow"), follow_redirects=True)
assert b'1 watches queued for rechecking.' in res.data
wait_for_all_checks(client)
# so that we set the state to 'unviewed' after all the edits
client.get(url_for("diff_history_page", uuid="first"))
# Make a change to header/footer/nav
set_modified_response()
# Trigger a check
res = client.get(url_for("form_watch_checknow"), follow_redirects=True)
assert b'1 watches queued for rechecking.' in res.data
# Give the thread time to pick it up
wait_for_all_checks(client)
# There should not be an unviewed change, as changes should be removed
res = client.get(url_for("index"))
assert b"unviewed" not in res.data