feat(stapp): restore chat history bubbles after /continue N (#138)

When the user runs '/continue N' in stapp, the agent's in-memory context
is restored, but the UI previously showed only a single '✅ restored' line
— all prior chat bubbles were missing.

This change parses the target session log and reconstructs the
user/assistant message pairs into st.session_state.messages, so reopening
a session feels like the conversation was never interrupted.

* continue_cmd.py: add extract_ui_messages(path)
  - parses model_responses log into [{role, content}, ...]
  - groups multi-turn LLM calls (prompts whose text starts with the
    '### [WORKING MEMORY]' header) into a single assistant bubble,
    inserting the existing '**LLM Running (Turn N) ...**' marker so
    fold_turns() renders them as collapsible segments.
  - two small helpers (_user_text / _assistant_text) keep parsing local.

* stapp.py: in the /continue branch, resolve the target log path BEFORE
  calling handle_frontend_command (which snapshots the current log and
  would otherwise shift list_sessions indices), then replace
  session_state.messages with the reconstructed history on success.
  Falls back to the previous behavior for bare /continue or failure.

Co-authored-by: wjl2023 <wjl2023@users.noreply.github.com>

frontends/continue_cmd.py

@@ -214,6 +214,57 @@ def handle(agent, query, display_queue):
     return query
 
 
+def _user_text(prompt_body):
+    """User-typed text from a prompt JSON; '' if this is an agent auto-continuation."""
+    try: msg = json.loads(prompt_body)
+    except Exception: return ''
+    if not isinstance(msg, dict): return ''
+    for blk in msg.get('content', []) or []:
+        if isinstance(blk, dict) and blk.get('type') == 'text':
+            t = (blk.get('text') or '').strip()
+            if t and not t.startswith('### [WORKING MEMORY]'): return t
+    return ''
+
+
+def _assistant_text(response_body):
+    """Joined text from a response blocks repr; '' on parse failure."""
+    try: blocks = ast.literal_eval(response_body)
+    except Exception: return ''
+    if not isinstance(blocks, list): return ''
+    return '\n'.join(b['text'] for b in blocks
+                     if isinstance(b, dict) and b.get('type') == 'text'
+                     and isinstance(b.get('text'), str) and b['text'].strip())
+
+
+_TURN_MARK = '**LLM Running (Turn {}) ...**\n\n'
+
+
+def extract_ui_messages(path):
+    """Parse a model_responses log into [{role, content}, ...] for UI replay.
+
+    Auto-continuation turns are folded into one assistant bubble with Turn markers,
+    matching live chat rendering via fold_turns().
+    """
+    try:
+        with open(path, encoding='utf-8', errors='replace') as f: content = f.read()
+    except Exception: return []
+
+    rounds = []  # [(user_text, [turn_text, ...]), ...]
+    for prompt, response in _pairs(content):
+        user = _user_text(prompt)
+        if user or not rounds: rounds.append((user, []))
+        rounds[-1][1].append(_assistant_text(response))
+
+    out = []
+    for user, turns in rounds:
+        if not user or not any(turns): continue
+        body = '\n\n'.join(t if i == 0 else _TURN_MARK.format(i + 1) + t
+                           for i, t in enumerate(turns))
+        out += [{'role': 'user', 'content': user},
+                {'role': 'assistant', 'content': body}]
+    return out
+
+
 def handle_frontend_command(agent, query, exclude_pid=None):
     """Frontend-friendly /continue entry that returns text directly."""
     s = (query or '').strip()

frontends/stapp.py

@@ -15,7 +15,7 @@ import streamlit as st
 import time, json, re, threading, queue
 from agentmain import GeneraticAgent
 import chatapp_common  # activate /continue command (monkey patches GeneraticAgent)
-from continue_cmd import handle_frontend_command, reset_conversation
+from continue_cmd import handle_frontend_command, reset_conversation, list_sessions, extract_ui_messages
 
 st.set_page_config(page_title="Cowork", layout="wide")
 
@@ -190,10 +190,19 @@ if prompt := st.chat_input("any task?"):
         st.session_state.messages = [{"role": "assistant", "content": reset_conversation(agent), "time": ts}]
         _reset_and_rerun()
     if cmd.startswith("/continue"):
-        st.session_state.messages = list(st.session_state.messages) + [
+        m = re.match(r'/continue\s+(\d+)\s*$', cmd.strip())
-            {"role": "user", "content": cmd, "time": ts},
+        sessions = list_sessions(exclude_pid=os.getpid()) if m else []
-            {"role": "assistant", "content": handle_frontend_command(agent, cmd), "time": ts},
+        idx = int(m.group(1)) - 1 if m else -1
-        ]
+        # Resolve target path BEFORE handle (which snapshots current log, shifting indices).
+        target = sessions[idx][0] if 0 <= idx < len(sessions) else None
+        result = handle_frontend_command(agent, cmd)
+        history = extract_ui_messages(target) if target and result.startswith('✅') else None
+        tail = [{"role": "assistant", "content": result, "time": ts}]
+        if history:
+            st.session_state.messages = history + tail
+        else:
+            st.session_state.messages = list(st.session_state.messages) + \
+                [{"role": "user", "content": cmd, "time": ts}] + tail
         _reset_and_rerun()
     st.session_state.messages.append({"role": "user", "content": prompt})
     if hasattr(agent, '_pet_req') and not prompt.startswith('/'): agent._pet_req('state=walk')