fix: resolve bugs 01 through 08 and set Ollama TTL to 1h

tests/test_api.py

@@ -2497,6 +2497,56 @@ def test_lessons_debrief_and_history():
     assert last_session["errors_count"] == 2
 
 
+def test_reverse_dictionary_lookup():
+    """
+    BUG-06 regression test: Searching Russian words should perform reverse lookup,
+    returning the corresponding English term, phonetics, and Russian meaning.
+    """
+    from app.auth import _login_failures
+    _login_failures.clear()
+    login_resp = client.post("/api/login", json={"username": "testadmin", "password": "testpass12345"})
+    assert login_resp.status_code == 200
+    token = login_resp.json()["token"]
+    headers = {"Authorization": f"Bearer {token}"}
+
+    resp = client.get("/api/lookup?q=собака", headers=headers)
+    assert resp.status_code == 200
+    data = resp.json()
+    assert data["is_reverse"] is True
+    assert data["term"].lower() == "dog"
+    assert "собака" in data["translation"].lower()
+    assert data["phonetic"] != ""
+
+
+def test_wikipedia_parsoid_html_cleaning():
+    """
+    BUG-05 regression test: Wikipedia Parsoid HTML containing data-mw attributes with
+    embedded <ref> and literal '>' chars should be cleanly stripped without leaking into text.
+    """
+    from app.reader import clean_wiki_html
+
+    sample_html = (
+        '<body>'
+        '<span class="mw-empty-elt" about="#mwt1" typeof="mw:Transclusion" '
+        'data-mw=\'{"parts":[{"template":{"target":{"wt":"Infobox"},"params":{"title":{"wt":"Test<ref>RefText</ref>"}}}]}\'>'
+        '</span>'
+        '<p>Linux is a Unix-like operating system.</p>'
+        '<table style="font-family:\'Times New Roman\', serif;"><tr><td>Clean Cell</td></tr></table>'
+        '</body>'
+    )
+    chapters, cover, words = clean_wiki_html(sample_html, "Test Article")
+    assert len(chapters) >= 1
+    content = chapters[0]["content"]
+    assert "data-mw" not in content
+    assert "{" not in content
+    assert "Infobox" not in content
+    assert "RefText" not in content
+    assert "Linux is a Unix-like operating system." in content
+    assert "Times New Roman" not in content  # stripped cleanly from table
+    assert "Clean Cell" in content
+
+
+