a51c509a87
На Python 3.9 `<![foo]>` вызывал NotImplementedError из ParserBase.error(), а адрес вида `http://[::1` — ValueError в urljoin/urlsplit; одна такая конструкция роняла wiki_sync целиком. Теперь нестандартная секция пропускается до `>`, битая ссылка — пропускается, остальные ссылки страницы сохраняются. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
107 lines
5.4 KiB
Python
107 lines
5.4 KiB
Python
"""Разбор ссылок страницы Eva: вложения и ссылки на страницы (design 6–7, 12)."""
|
||
import os
|
||
import unittest
|
||
|
||
from tests import _common # noqa: F401
|
||
import agent
|
||
|
||
BASE = "https://eva.test"
|
||
OBJ = "/files/obj/CmfDocument/CmfDocument%3Aa08/CmfDocument%3A0b1c2d3e-0000-4000-8000-000000000001"
|
||
HERE = os.path.dirname(os.path.abspath(__file__))
|
||
|
||
|
||
def page(name):
|
||
with open(os.path.join(HERE, "fixtures", "pages", name), encoding="utf-8") as fh:
|
||
return fh.read()
|
||
|
||
|
||
class ClassifyLinksTest(unittest.TestCase):
|
||
def test_image_relative_files(self):
|
||
att, links = agent.classify_links(page("image.html"), BASE)
|
||
self.assertEqual(att, [{"ref": OBJ + "/img.png?version=1", "url": BASE + OBJ + "/img.png", "name": "Схема"}])
|
||
self.assertEqual(links, [])
|
||
|
||
def test_thumbnail_maps_to_file(self):
|
||
att, _ = agent.classify_links(page("thumb_and_link.html"), BASE)
|
||
self.assertEqual(att[0]["url"], BASE + OBJ + "/model.xlsx")
|
||
self.assertEqual(att[0]["name"], "model.xlsx")
|
||
self.assertEqual(att[0]["ref"], OBJ + "/model.xlsx.meta/thumbnail.jpg?version=1")
|
||
self.assertEqual(att[2]["url"], BASE + OBJ + "/data.dbml")
|
||
self.assertEqual(att[2]["name"], "data.dbml")
|
||
|
||
def test_file_link_same_url(self):
|
||
att, _ = agent.classify_links(page("thumb_and_link.html"), BASE)
|
||
self.assertEqual(len(att), 3) # повторы не склеиваются
|
||
self.assertEqual(att[1]["url"], att[0]["url"])
|
||
self.assertEqual(att[1]["ref"], OBJ + "/model.xlsx?version=1")
|
||
self.assertEqual(att[1]["name"], "model.xlsx")
|
||
|
||
def test_static_ignored(self):
|
||
att, links = agent.classify_links('<img src="static/attach-file.png">', BASE)
|
||
self.assertEqual((att, links), ([], []))
|
||
|
||
def test_page_link_relative(self):
|
||
_, links = agent.classify_links(page("links.html"), BASE)
|
||
self.assertEqual(links[0], {"ref": "project/Document/DOC-019834", "code": "DOC-019834"})
|
||
self.assertEqual(links[1], {"ref": "https://eva.test/project/Document/doc-5", "code": "DOC-5"})
|
||
|
||
def test_foreign_host(self):
|
||
att, links = agent.classify_links(page("links.html"), BASE)
|
||
self.assertEqual(att, [])
|
||
self.assertEqual([l["code"] for l in links], ["DOC-019834", "DOC-5"])
|
||
att, links = agent.classify_links(
|
||
'<a href="https://other.test/files/x.png">x</a><a href="https://other.test/project/Document/DOC-1">y</a>',
|
||
BASE)
|
||
self.assertEqual((att, links), ([], []))
|
||
|
||
def test_name_unquoted(self):
|
||
att, _ = agent.classify_links(page("name_quoted.html"), BASE)
|
||
self.assertEqual(att[0]["name"], "ПМП.xlsx")
|
||
self.assertEqual(att[0]["url"], BASE + OBJ + "/%D0%9F%D0%9C%D0%9F.xlsx")
|
||
|
||
def test_ref_exact(self):
|
||
att, links = agent.classify_links(page("amp.html"), BASE)
|
||
self.assertEqual(att[0]["ref"], OBJ + "/a.png?version=1&x=2") # после unescape, как ищет хаб
|
||
self.assertEqual(links[0]["ref"], "project/Document/DOC-7?a=1&b=2")
|
||
self.assertEqual(att[0]["url"], BASE + OBJ + "/a.png")
|
||
|
||
def test_base_with_trailing_slash_and_title(self):
|
||
att, _ = agent.classify_links('<a href="%s/f.bin" title="Файл">текст</a>' % OBJ, BASE + "/")
|
||
self.assertEqual(att[0]["name"], "Файл")
|
||
|
||
def test_name_from_link_text(self):
|
||
att, _ = agent.classify_links('<a href="%s/f.bin"> мой <b>файл</b> </a>' % OBJ, BASE)
|
||
self.assertEqual(att[0]["name"], "мой файл")
|
||
|
||
def test_order_nested_img_in_link(self):
|
||
att, _ = agent.classify_links(
|
||
'<a href="%s/a.bin"><img src="%s/b.png"></a><img src="%s/c.png">' % (OBJ, OBJ, OBJ), BASE)
|
||
self.assertEqual([a["name"] for a in att], ["a.bin", "b.png", "c.png"])
|
||
|
||
def test_unknown_marked_section_does_not_crash(self):
|
||
"""I-1: на 3.9 `<![foo]>` уходит в ParserBase.error() → NotImplementedError; ссылки до и после сохраняются."""
|
||
html = ('<img src="%s/a.png"><![foo]><p>текст</p><![if !supportLists]>x<![endif]>'
|
||
'<a href="/project/Document/DOC-3">c</a><img src="%s/b.png">' % (OBJ, OBJ))
|
||
att, links = agent.classify_links(html, BASE)
|
||
self.assertEqual([a["name"] for a in att], ["a.png", "b.png"])
|
||
self.assertEqual(links, [{"ref": "/project/Document/DOC-3", "code": "DOC-3"}])
|
||
|
||
def test_marked_section_bad_name_and_unclosed(self):
|
||
for html in ('<img src="%s/a.png"><![ ]><img src="%s/b.png">' % (OBJ, OBJ),
|
||
'<img src="%s/a.png"><img src="%s/b.png"><![foo' % (OBJ, OBJ)):
|
||
with self.subTest(html=html):
|
||
att, _ = agent.classify_links(html, BASE)
|
||
self.assertEqual([a["name"] for a in att], ["a.png", "b.png"])
|
||
|
||
def test_bad_url_skipped(self):
|
||
"""I-1: `http://[::1` и порт не числом дают ValueError в urljoin/urlsplit — ссылка пропускается, остальные целы."""
|
||
html = ('<img src="http://[::1"><a href="https://[zz/files/x">x</a><img src="https://eva.test:abc/files/y.png">'
|
||
'<img src="%s/b.png"><a href="/project/Document/DOC-4">d</a>' % OBJ)
|
||
att, links = agent.classify_links(html, BASE)
|
||
self.assertEqual([a["name"] for a in att], ["b.png"])
|
||
self.assertEqual([l["code"] for l in links], ["DOC-4"])
|
||
|
||
|
||
if __name__ == "__main__":
|
||
unittest.main()
|