Implement chat interface with message history and system prompt support; update frontend and tests accordingly

Co-authored-by: Copilot <copilot@github.com>

frontend/tests/test_frontend.py

@@ -172,7 +172,7 @@ def test_generate_text_page_renders(client):
     _set_auth(client)
     resp = client.get("/generate/text")
     assert resp.status_code == 200
-    assert b"Text Generation" in resp.data
+    assert b"Text Chat" in resp.data
 
 
 def test_generate_text_requires_login(client):
@@ -183,13 +183,108 @@ def test_generate_text_requires_login(client):
 
 def test_generate_text_success(client):
     _set_auth(client)
-    mock = _mock_response(
+    gen_mock = _mock_response(
         200, {"id": "g1", "model": "openai/gpt-4o", "content": "Hello world", "usage": None})
-    with patch("frontend.app.main.httpx.request", return_value=mock):
+    models_mock = _mock_response(200, [
+        {"id": "openai/gpt-4o", "name": "GPT-4o", "modality": "text"}
+    ])
+    with patch("frontend.app.main.httpx.request", side_effect=[gen_mock, models_mock]):
         resp = client.post(
-            "/generate/text", data={"model": "openai/gpt-4o", "prompt": "Say hello"})
+            "/generate/text",
+            data={"model": "openai/gpt-4o", "prompt": "Say hello", "action": "send"})
     assert resp.status_code == 200
     assert b"Hello world" in resp.data
+    assert b"chat-bubble--assistant" in resp.data
+
+
+def test_generate_text_page_shows_optional_system_prompt(client):
+    _set_auth(client)
+    models_mock = _mock_response(200, [])
+    with patch("frontend.app.main.httpx.request", return_value=models_mock):
+        resp = client.get("/generate/text")
+    assert resp.status_code == 200
+    assert b"System prompt (optional)" in resp.data
+    assert b'name="system_prompt"' in resp.data
+
+
+def test_generate_text_forwards_system_prompt(client):
+    _set_auth(client)
+    gen_mock = _mock_response(
+        200, {"id": "g1", "model": "openai/gpt-4o", "content": "Hello world", "usage": None})
+    models_mock = _mock_response(200, [
+        {"id": "openai/gpt-4o", "name": "GPT-4o", "modality": "text"}
+    ])
+
+    with patch("frontend.app.main.httpx.request", side_effect=[gen_mock, models_mock]) as mock_request:
+        resp = client.post(
+            "/generate/text",
+            data={
+                "model": "openai/gpt-4o",
+                "prompt": "Say hello",
+                "system_prompt": "You are concise.",
+                "action": "send",
+            },
+        )
+
+    assert resp.status_code == 200
+    first_call_kwargs = mock_request.call_args_list[0].kwargs
+    assert first_call_kwargs["json"]["system_prompt"] == "You are concise."
+    # Messages array sent (not bare prompt)
+    assert "messages" in first_call_kwargs["json"]
+
+
+def test_generate_text_chat_history_accumulates(client):
+    """Second message includes prior user+assistant turns in messages array."""
+    _set_auth(client)
+
+    turn1_gen = _mock_response(
+        200, {"id": "g1", "model": "openai/gpt-4o", "content": "Turn 1 reply", "usage": None})
+    turn1_models = _mock_response(
+        200, [{"id": "openai/gpt-4o", "name": "GPT-4o", "modality": "text"}])
+    turn2_gen = _mock_response(
+        200, {"id": "g2", "model": "openai/gpt-4o", "content": "Turn 2 reply", "usage": None})
+    turn2_models = _mock_response(
+        200, [{"id": "openai/gpt-4o", "name": "GPT-4o", "modality": "text"}])
+
+    with patch("frontend.app.main.httpx.request", side_effect=[turn1_gen, turn1_models]):
+        client.post(
+            "/generate/text", data={"model": "openai/gpt-4o", "prompt": "First", "action": "send"})
+
+    with patch("frontend.app.main.httpx.request", side_effect=[turn2_gen, turn2_models]) as mock_req:
+        resp = client.post(
+            "/generate/text", data={"model": "openai/gpt-4o", "prompt": "Second", "action": "send"})
+
+    assert resp.status_code == 200
+    assert b"Turn 1 reply" in resp.data
+    assert b"Turn 2 reply" in resp.data
+    # Backend received 3 messages: First(user), Turn1(assistant), Second(user)
+    sent_messages = mock_req.call_args_list[0].kwargs["json"]["messages"]
+    assert len(sent_messages) == 3
+    assert sent_messages[0]["role"] == "user" and sent_messages[0]["content"] == "First"
+    assert sent_messages[1]["role"] == "assistant"
+    assert sent_messages[2]["role"] == "user" and sent_messages[2]["content"] == "Second"
+
+
+def test_generate_text_clear_resets_history(client):
+    """Clear action removes session history and redirects."""
+    _set_auth(client)
+
+    gen_mock = _mock_response(
+        200, {"id": "g1", "model": "openai/gpt-4o", "content": "Reply", "usage": None})
+    models_mock = _mock_response(
+        200, [{"id": "openai/gpt-4o", "name": "GPT-4o", "modality": "text"}])
+    with patch("frontend.app.main.httpx.request", side_effect=[gen_mock, models_mock]):
+        client.post(
+            "/generate/text", data={"model": "openai/gpt-4o", "prompt": "Hi", "action": "send"})
+
+    clear_resp = client.post("/generate/text", data={"action": "clear"})
+    assert clear_resp.status_code == 302
+
+    models_mock2 = _mock_response(
+        200, [{"id": "openai/gpt-4o", "name": "GPT-4o", "modality": "text"}])
+    with patch("frontend.app.main.httpx.request", return_value=models_mock2):
+        get_resp = client.get("/generate/text")
+    assert b"No messages yet" in get_resp.data
 
 
 def test_generate_image_page_renders(client):
@@ -251,8 +346,11 @@ def test_generate_video_image_mode(client):
 
 def test_generate_upstream_error_shows_message(client):
     _set_auth(client)
-    mock = _mock_response(502, {"detail": "OpenRouter error: timeout"})
-    with patch("frontend.app.main.httpx.request", return_value=mock):
+    gen_mock = _mock_response(502, {"detail": "OpenRouter error: timeout"})
+    models_mock = _mock_response(200, [
+        {"id": "openai/gpt-4o", "name": "GPT-4o", "modality": "text"}
+    ])
+    with patch("frontend.app.main.httpx.request", side_effect=[gen_mock, models_mock]):
         resp = client.post(
             "/generate/text", data={"model": "openai/gpt-4o", "prompt": "Hi"})
     assert resp.status_code == 200