cache improvement (gemini)

Readme.md

@@ -1,6 +1,8 @@
 # Manual Slop
 
-An AI coding assistant interface featuring multi-project management, rich discussion history, dynamic tool utilization, and a robust frontend.
+Vibe coding.. but more manual
+
+![img](./gallery/python_2026-02-21_23-37-29.png)
 
 This tool is designed to work as an auxiliary assistant that natively interacts with your codebase via PowerShell and MCP-like file tools, supporting both Anthropic and Gemini APIs.
 
@@ -25,7 +27,19 @@ Features:
 * [docs/guide_tools.md](docs/guide_tools.md) for information on the AI tooling capabilities
 * [docs/guide_architecture.md](docs/guide_architecture.md) for an in-depth breakdown of the codebase architecture
 
-## Running
+## Instructions
 
-Use `uv run gui.py` or `python gui.py` to launch the frontend.
+1. Make a credentials.toml in the immediate directory of your clone:
-Requires a `credentials.toml` file in the root directory containing `gemini` and `anthropic` API keys.
+
+```toml
+[gemini]
+api_key = "****"
+[anthropic]
+api_key = "****"
+```
+
+2. Have fun. This is experiemntal slop.
+
+```ps1
+uv run .\gui.py
+```

ai_client.py

@@ -430,11 +430,24 @@ def _send_gemini(md_content: str, user_message: str, base_dir: str, file_items:
         sys_instr = f"{_get_combined_system_prompt()}\n\n<context>\n{md_content}\n</context>"
         tools_decl = [_gemini_tool_declaration()]
         
-        global _gemini_cache
+        global _gemini_cache, _gemini_chat
+        
+        # DYNAMIC CONTEXT: Check if files/context changed mid-session
+        current_md_hash = hash(md_content)
+        old_history = None
+        if _gemini_chat and getattr(_gemini_chat, "_last_md_hash", None) != current_md_hash:
+            old_history = list(_gemini_chat.history) if _gemini_chat.history else []
+            if _gemini_cache:
+                try: _gemini_client.caches.delete(name=_gemini_cache.name)
+                except: pass
+            _gemini_chat = None
+            _gemini_cache = None
+            _append_comms("OUT", "request", {"message": "[CONTEXT CHANGED] Rebuilding cache and chat session..."})
+
         if not _gemini_chat:
             chat_config = types.GenerateContentConfig(system_instruction=sys_instr, tools=tools_decl)
             try:
-                # Gemini requires >= 32,768 tokens for caching. We try to cache, and fallback if it fails.
+                # Gemini requires 1024 (Flash) or 4096 (Pro) tokens to cache.
                 _gemini_cache = _gemini_client.caches.create(
                     model=_model,
                     config=types.CreateCachedContentConfig(
@@ -446,10 +459,15 @@ def _send_gemini(md_content: str, user_message: str, base_dir: str, file_items:
                 chat_config = types.GenerateContentConfig(cached_content=_gemini_cache.name)
                 _append_comms("OUT", "request", {"message": f"[CACHE CREATED] {_gemini_cache.name}"})
             except Exception as e:
-                # Fallback to standard request if under 32k tokens or cache creation fails
+                # Fallback if under token limit or API error
                 pass
                 
-            _gemini_chat = _gemini_client.chats.create(model=_model, config=chat_config)
+            kwargs = {"model": _model, "config": chat_config}
+            if old_history:
+                kwargs["history"] = old_history
+            
+            _gemini_chat = _gemini_client.chats.create(**kwargs)
+            _gemini_chat._last_md_hash = current_md_hash
         
         _append_comms("OUT", "request", {"message": f"[ctx {len(md_content)} + msg {len(user_message)}]"})
         payload, all_text = user_message, []