ayechat-dev 0.36.9.20260204011001__py3-none-any.whl → 0.36.9.20260205232002__py3-none-any.whl

aye/controller/llm_invoker.py

@@ -13,12 +13,16 @@ from aye.model.source_collector import collect_sources
 from aye.model.auth import get_user_config
 from aye.model.offline_llm_manager import is_offline_model
 from aye.controller.util import is_truncated_json
-from aye.model.config import SYSTEM_PROMPT, MODELS, DEFAULT_MAX_OUTPUT_TOKENS, DEFAULT_CONTEXT_TARGET_KB
+from aye.model.config import SYSTEM_PROMPT, MODELS, DEFAULT_MAX_OUTPUT_TOKENS, DEFAULT_CONTEXT_TARGET_KB, CONTEXT_HARD_LIMIT_KB
 from aye.model import telemetry
 
 import os
 
 
+def _is_verbose():
+    return get_user_config("verbose", "off").lower() == "on"
+
+
 def _is_debug():
     return get_user_config("debug", "off").lower() == "on"
 
@@ -77,7 +81,7 @@ def _get_context_hard_limit(model_id: str) -> int:
     model_config = _get_model_config(model_id)
     if model_config and "max_prompt_kb" in model_config:
         return model_config["max_prompt_kb"] * 1024
-    return 170 * 1024
+    return CONTEXT_HARD_LIMIT_KB * 1024
 
 
 def _filter_ground_truth(files: Dict[str, str], conf: Any, verbose: bool) -> Dict[str, str]:
@@ -134,6 +138,8 @@ def _get_rag_context_files(prompt: str, conf: Any, verbose: bool) -> Dict[str, s
 
     context_target_size = _get_context_target_size(conf.selected_model)
     context_hard_limit = _get_context_hard_limit(conf.selected_model)
+    #context_target_size = DEFAULT_CONTEXT_TARGET_KB #_get_context_target_size(conf.selected_model)
+    #context_hard_limit = CONTEXT_HARD_LIMIT_KB # _get_context_hard_limit(conf.selected_model)
 
     if _is_debug():
         rprint(f"[yellow]Context target: {context_target_size / 1024:.1f}KB, hard limit: {context_hard_limit / 1024:.1f}KB[/]")
@@ -298,50 +304,54 @@ def invoke_llm(
     model_config = _get_model_config(conf.selected_model)
     max_output_tokens = model_config.get("max_output_tokens", DEFAULT_MAX_OUTPUT_TOKENS) if model_config else DEFAULT_MAX_OUTPUT_TOKENS
 
-    # 1. Try local/offline model plugins first (no streaming UI for local models)
-    local_response = plugin_manager.handle_command("local_model_invoke", {
-        "prompt": prompt,
-        "model_id": conf.selected_model,
-        "source_files": source_files,
-        "chat_id": chat_id,
-        "root": conf.root,
-        "system_prompt": system_prompt,
-        "max_output_tokens": max_output_tokens
-    })
-
-    if local_response is not None:
-        return LLMResponse(
-            summary=local_response.get("summary", ""),
-            updated_files=local_response.get("updated_files", []),
-            chat_id=None,
-            source=LLMSource.LOCAL
-        )
-
-    # 2. API call with spinner + streaming display
-    if _is_debug():
-        print(f"[DEBUG] Processing chat message with chat_id={chat_id or -1}, model={conf.selected_model}")
-
-    telemetry_payload = telemetry.build_payload(top_n=20) if telemetry.is_enabled() else None
-
-    # Create spinner - will be stopped when streaming starts
+    # Create spinner - will be shown for ALL model types (local, databricks, API)
     spinner = StoppableSpinner(
         console,
         messages=DEFAULT_THINKING_MESSAGES,
         interval=15.0
     )
     
+    # For API calls, we also have streaming display
+    streaming_display: Optional[StreamingResponseDisplay] = None
+    
     def stop_spinner():
-        """Callback to stop spinner when first content arrives."""
+        """Callback to stop spinner when first content arrives (for streaming API)."""
         spinner.stop()
-    
-    # Create streaming display with callback to stop spinner on first content
-    streaming_display = StreamingResponseDisplay(on_first_content=stop_spinner)
-    stream_callback = create_streaming_callback(streaming_display)
 
     try:
-        # Start the spinner before the API call
+        # Start the spinner before ANY LLM call (local or API)
         spinner.start()
         
+        # 1. Try local/offline model plugins first
+        local_response = plugin_manager.handle_command("local_model_invoke", {
+            "prompt": prompt,
+            "model_id": conf.selected_model,
+            "source_files": source_files,
+            "chat_id": chat_id,
+            "root": conf.root,
+            "system_prompt": system_prompt,
+            "max_output_tokens": max_output_tokens
+        })
+
+        if local_response is not None:
+            # Local model handled the request - spinner will be stopped in finally block
+            return LLMResponse(
+                summary=local_response.get("summary", ""),
+                updated_files=local_response.get("updated_files", []),
+                chat_id=None,
+                source=LLMSource.LOCAL
+            )
+
+        # 2. API call with streaming display
+        if _is_debug():
+            print(f"[DEBUG] Processing chat message with chat_id={chat_id or -1}, model={conf.selected_model}")
+
+        telemetry_payload = telemetry.build_payload(top_n=20) if telemetry.is_enabled() else None
+
+        # Create streaming display with callback to stop spinner on first content
+        streaming_display = StreamingResponseDisplay(on_first_content=stop_spinner)
+        stream_callback = create_streaming_callback(streaming_display)
+
         api_resp = cli_invoke(
             message=prompt,
             chat_id=chat_id or -1,
@@ -352,29 +362,29 @@ def invoke_llm(
             telemetry=telemetry_payload,
             on_stream_update=stream_callback
         )
-    finally:
-        # Ensure spinner is stopped (in case no streaming content was received)
-        spinner.stop()
-        
-        # Always stop the streaming display when done
-        if streaming_display.is_active():
-            streaming_display.stop()
 
-    if telemetry_payload is not None:
-        telemetry.reset()
+        if telemetry_payload is not None:
+            telemetry.reset()
 
-    if _is_debug():
-        print(f"[DEBUG] Chat message processed, response keys: {api_resp.keys() if api_resp else 'None'}")
+        if _is_debug():
+            print(f"[DEBUG] Chat message processed, response keys: {api_resp.keys() if api_resp else 'None'}")
 
-    # Check if we already displayed the response via streaming
-    streamed_summary = bool(api_resp.get("_streamed_summary")) if isinstance(api_resp, dict) else False
+        # Check if we already displayed the response via streaming
+        streamed_summary = bool(api_resp.get("_streamed_summary")) if isinstance(api_resp, dict) else False
 
-    # 3. Parse API response
-    assistant_resp, new_chat_id = _parse_api_response(api_resp)
+        # 3. Parse API response
+        assistant_resp, new_chat_id = _parse_api_response(api_resp)
 
-    return LLMResponse(
-        summary="" if streamed_summary else assistant_resp.get("answer_summary", ""),
-        updated_files=assistant_resp.get("source_files", []),
-        chat_id=new_chat_id,
-        source=LLMSource.API
-    )
+        return LLMResponse(
+            summary="" if streamed_summary else assistant_resp.get("answer_summary", ""),
+            updated_files=assistant_resp.get("source_files", []),
+            chat_id=new_chat_id,
+            source=LLMSource.API
+        )
+    finally:
+        # Ensure spinner is stopped for ALL code paths (local model, API, or error)
+        spinner.stop()
+        
+        # Stop the streaming display if it was created and is active
+        if streaming_display is not None and streaming_display.is_active():
+            streaming_display.stop()

aye/model/config.py

@@ -17,12 +17,17 @@ SMALL_PROJECT_FILE_LIMIT = 200
 # Projects smaller than this will skip RAG and include all files directly.
 # Set to match default max_prompt_kb (170KB) so all files can fit in context.
 SMALL_PROJECT_TOTAL_SIZE_LIMIT = 170 * 1024  # 170KB
+#SMALL_PROJECT_TOTAL_SIZE_LIMIT = 100 * 1024  # 170KB
 
 # Default maximum output tokens for LLM responses
 DEFAULT_MAX_OUTPUT_TOKENS = 32000
+#DEFAULT_MAX_OUTPUT_TOKENS = 16000
 
 # Default context target size in KB (used when model doesn't specify one)
 DEFAULT_CONTEXT_TARGET_KB = 150
+#DEFAULT_CONTEXT_TARGET_KB = 20
+
+CONTEXT_HARD_LIMIT_KB = 170
 
 # Shared system prompt for all LLM interactions
 SYSTEM_PROMPT = (

aye/plugins/databricks_model.py

@@ -0,0 +1,312 @@
+import os
+import json
+from typing import Dict, Any, Optional
+import httpx
+from pathlib import Path
+import traceback
+
+from rich import print as rprint
+
+from .plugin_base import Plugin
+from aye.model.config import SYSTEM_PROMPT, MODELS, DEFAULT_MAX_OUTPUT_TOKENS
+from aye.model.auth import get_user_config
+from aye.controller.util import is_truncated_json
+
+LLM_TIMEOUT = 600.0
+
+
+# Message shown when LLM response is truncated due to output token limits
+TRUNCATED_RESPONSE_MESSAGE = (
+    "It looks like my response was cut off because it exceeded the output limit. "
+    "This usually happens when you ask me to generate or modify many files at once.\n\n"
+    "**To fix this, please try:**\n"
+    "1. Break your request into smaller parts (e.g., one file at a time)\n"
+    "2. Use the `with` command to focus on specific files: `with file1.py, file2.py: your request`\n"
+    "3. Ask me to work on fewer files or smaller changes in each request\n\n"
+    "For example, instead of 'update all files to add logging', try:\n"
+    "  `with src/main.py: add logging to this file`"
+)
+
+
+def _get_model_config(model_id: str) -> Optional[Dict[str, Any]]:
+    """Get configuration for a specific model."""
+    for model in MODELS:
+        if model["id"] == model_id:
+            return model
+    return None
+
+
+def _extract_json_object(raw_response: str, prefer_last: bool = True, require_keys=None):
+    """
+    Best-effort extraction of a JSON object (dict) from a raw LLM response.
+
+    Handles common failure modes where the model returns:
+    - extra commentary before/after JSON
+    - multiple JSON objects (e.g., an invalid attempt + a corrected attempt)
+
+    Args:
+        raw_response: raw LLM response string
+        prefer_last: when multiple JSON objects exist, return the last parsed object
+        require_keys: optional iterable of keys; if provided, only consider objects
+                      that contain all of these keys
+
+    Returns:
+        dict or None
+    """
+    # 1) Direct JSON parse
+    try:
+        obj = json.loads(raw_response)
+        if isinstance(obj, dict):
+            if require_keys and not all(k in obj for k in require_keys):
+                return None
+            return obj
+    except Exception:
+        pass
+
+    text = str(raw_response)
+
+    # 2) Scan for balanced JSON object candidates (string/escape-aware)
+    candidates = []
+    depth = 0
+    start = None
+    in_str = False
+    escape = False
+
+    for i, ch in enumerate(text):
+        if in_str:
+            if escape:
+                escape = False
+            elif ch == '\\':
+                escape = True
+            elif ch == '"':
+                in_str = False
+            continue
+
+        # not in string
+        if ch == '"':
+            in_str = True
+            continue
+
+        if ch == '{':
+            if depth == 0:
+                start = i
+            depth += 1
+        elif ch == '}' and depth > 0:
+            depth -= 1
+            if depth == 0 and start is not None:
+                candidates.append(text[start : i + 1])
+                start = None
+
+    parsed = []
+    for cand in candidates:
+        try:
+            obj = json.loads(cand)
+            if not isinstance(obj, dict):
+                continue
+            if require_keys and not all(k in obj for k in require_keys):
+                continue
+            parsed.append(obj)
+        except Exception:
+            continue
+
+    if not parsed:
+        return None
+
+    return parsed[-1] if prefer_last else parsed[0]
+
+
+class DatabricksModelPlugin(Plugin):
+    name = "databricks_model"
+    version = "1.0.0"
+    premium = "free"
+
+    def __init__(self):
+        super().__init__()
+        self.chat_history: Dict[str, list] = {}
+        self.history_file: Optional[Path] = None
+
+    def init(self, cfg: Dict[str, Any]) -> None:
+        """Initialize the local model plugin."""
+        super().init(cfg)
+        if self.debug:
+            rprint(f"[bold yellow]Initializing {self.name} v{self.version}[/]")
+        
+    def _load_history(self) -> None:
+        """Load chat history from disk."""
+        if not self.history_file:
+            if self.verbose:
+                rprint("[yellow]History file path not set for local model. Skipping load.[/]")
+            self.chat_history = {}
+            return
+
+        if self.history_file.exists():
+            try:
+                data = json.loads(self.history_file.read_text(encoding="utf-8"))
+                self.chat_history = data.get("conversations", {})
+            except Exception as e:
+                if self.verbose:
+                    rprint(f"[yellow]Could not load chat history: {e}[/]")
+                self.chat_history = {}
+        else:
+            self.chat_history = {}
+
+    def _save_history(self) -> None:
+        """Save chat history to disk."""
+        if not self.history_file:
+            if self.verbose:
+                rprint("[yellow]History file path not set for local model. Skipping save.[/]")
+            return
+
+        try:
+            self.history_file.parent.mkdir(parents=True, exist_ok=True)
+            data = {"conversations": self.chat_history}
+            self.history_file.write_text(json.dumps(data, indent=2), encoding="utf-8")
+        except Exception as e:
+            if self.verbose:
+                rprint(f"[yellow]Could not save chat history: {e}[/]")
+
+    def _get_conversation_id(self, chat_id: Optional[int] = None) -> str:
+        """Get conversation ID for history tracking."""
+        return str(chat_id) if chat_id and chat_id > 0 else "default"
+
+    def _build_user_message(self, prompt: str, source_files: Dict[str, str]) -> str:
+        """Build the user message with optional source files appended."""
+        user_message = prompt
+        if source_files:
+            user_message += "\n\n--- Source files are below. ---\n"
+            for file_name, content in source_files.items():
+                user_message += f"\n** {file_name} **\n```\n{content}\n```\n"
+        return user_message
+
+    def _parse_llm_response(self, generated_text: str) -> Dict[str, Any]:
+        """Parse LLM response text and convert to expected format."""
+        try:
+            llm_response = json.loads(generated_text)
+        except json.JSONDecodeError:
+            # Check if this looks like a truncated response
+            #if is_truncated_json(generated_text):
+            #    return {
+            #        "summary": TRUNCATED_RESPONSE_MESSAGE,
+            #        "updated_files": []
+            #    }
+            
+            # Not truncated, just malformed - return as plain text
+            llm_response = {
+                "answer_summary": generated_text,
+                "source_files": []
+            }
+        
+        return {
+            "summary": llm_response.get("answer_summary", ""),
+            "updated_files": [
+                {
+                    "file_name": f.get("file_name"),
+                    "file_content": f.get("file_content")
+                }
+                for f in llm_response.get("source_files", [])
+            ]
+        }
+
+    def _create_error_response(self, error_msg: str) -> Dict[str, Any]:
+        """Create a standardized error response."""
+        if self.verbose:
+            rprint(f"[red]{error_msg}[/]")
+        return {
+            "summary": error_msg,
+            "updated_files": []
+        }
+
+    def _handle_databricks(self, prompt: str, source_files: Dict[str, str], chat_id: Optional[int] = None, system_prompt: Optional[str] = None, max_output_tokens: int = DEFAULT_MAX_OUTPUT_TOKENS) -> Optional[Dict[str, Any]]:
+        api_url = os.environ.get("AYE_DBX_API_URL")
+        api_key = os.environ.get("AYE_DBX_API_KEY")
+        model_name = os.environ.get("AYE_DBX_MODEL", "gpt-3.5-turbo")
+        
+        if not api_url or not api_key:
+            return None
+        
+        conv_id = self._get_conversation_id(chat_id)
+        if conv_id not in self.chat_history:
+            self.chat_history[conv_id] = []
+        
+        user_message = self._build_user_message(prompt, source_files)
+        effective_system_prompt = system_prompt if system_prompt else SYSTEM_PROMPT
+        
+        messages_json = [{"role": "system", "content": effective_system_prompt}] + self.chat_history[conv_id] + [{"role": "user", "content": user_message}]
+        messages = messages_json # json.dumps(messages_json)
+        if self.debug:
+            print(">>>>>>>>>>>>>>>>")
+            print(self.chat_history[conv_id])
+            print(">>>>>>>>>>>>>>>>")
+        
+        headers = {"Content-Type": "application/json", "Authorization": f"Bearer {api_key}"}
+        #payload = {"model": model_name, "messages": messages, "temperature": 0.7, "max_tokens": max_output_tokens, "response_format": {"type": "json_object"}}
+        payload = {"model": model_name, "messages": messages, "temperature": 0.7, "max_tokens": max_output_tokens}
+        
+        try:
+            with httpx.Client(timeout=LLM_TIMEOUT) as client:
+                response = client.post(api_url, json=payload, headers=headers)
+                if self.verbose and response.status_code != 200:
+                    print(f"Status code: {response.status_code}")
+                    print("-----------------")
+                    print(response.text)
+                    print("-----------------")
+                response.raise_for_status()
+                result = response.json()
+                if result.get("choices") and result["choices"][0].get("message"):
+                    raw_response = result["choices"][0]["message"]["content"]
+                    generated_json = _extract_json_object(raw_response)
+                    generated_text = json.dumps(generated_json)
+                    if self.debug:
+                        print("-----------------")
+                        print(response.text)
+                        print("-----------------")
+                        print(generated_text)
+                        print("-----------------")
+                    #generated_text = result["choices"][0]["message"]["content"][1]["text"]
+                    self.chat_history[conv_id].append({"role": "user", "content": user_message})
+                    self.chat_history[conv_id].append({"role": "assistant", "content": generated_text})
+                    self._save_history()
+                    #return json.dumps(generated_text)
+                    return self._parse_llm_response(generated_text)
+                return self._create_error_response("Failed to get a valid response from the Databricks API")
+        except httpx.HTTPStatusError as e:
+            traceback.print_exc()
+            error_msg = f"DBX API error: {e.response.status_code}"
+            try:
+                error_detail = e.response.json()
+                if "error" in error_detail:
+                    error_msg += f" - {error_detail['error'].get('message', str(error_detail['error']))}"
+            except: error_msg += f" - {e.response.text[:200]}"
+            return self._create_error_response(error_msg)
+        except Exception as e:
+            traceback.print_exc()
+            return self._create_error_response(f"Error calling Databricks API: {str(e)}")
+
+
+    def on_command(self, command_name: str, params: Dict[str, Any]) -> Optional[Dict[str, Any]]:
+        if command_name == "new_chat":
+            root = params.get("root")
+            history_file = Path(root) / ".aye" / "chat_history.json" if root else Path.cwd() / ".aye" / "chat_history.json"
+            history_file.unlink(missing_ok=True)
+            self.chat_history = {}
+            if self.verbose: rprint("[yellow]Local model chat history cleared.[/]")
+            return {"status": "local_history_cleared"}
+
+        if command_name == "local_model_invoke":
+            prompt = params.get("prompt", "").strip()
+            model_id = params.get("model_id", "")
+            source_files = params.get("source_files", {})
+            chat_id = params.get("chat_id")
+            root = params.get("root")
+            system_prompt = params.get("system_prompt")
+            max_output_tokens = params.get("max_output_tokens", DEFAULT_MAX_OUTPUT_TOKENS)
+
+            self.history_file = Path(root) / ".aye" / "chat_history.json" if root else Path.cwd() / ".aye" / "chat_history.json"
+            self._load_history()
+
+            result = self._handle_databricks(prompt, source_files, chat_id, system_prompt, max_output_tokens)
+            if result is not None: return result
+
+            return None
+
+        return None

aye/plugins/local_model.py

@@ -136,48 +136,6 @@ class LocalModelPlugin(Plugin):
             "updated_files": []
         }
 
-    def _handle_databricks(self, prompt: str, source_files: Dict[str, str], chat_id: Optional[int] = None, system_prompt: Optional[str] = None, max_output_tokens: int = DEFAULT_MAX_OUTPUT_TOKENS) -> Optional[Dict[str, Any]]:
-        api_url = os.environ.get("AYE_DBX_API_URL")
-        api_key = os.environ.get("AYE_DBX_API_KEY")
-        model_name = os.environ.get("AYE_DBX_MODEL", "gpt-3.5-turbo")
-        
-        if not api_url or not api_key:
-            return None
-        
-        conv_id = self._get_conversation_id(chat_id)
-        if conv_id not in self.chat_history:
-            self.chat_history[conv_id] = []
-        
-        user_message = self._build_user_message(prompt, source_files)
-        effective_system_prompt = system_prompt if system_prompt else SYSTEM_PROMPT
-        
-        messages = [{"role": "system", "content": effective_system_prompt}] + self.chat_history[conv_id] + [{"role": "user", "content": user_message}]
-        
-        headers = {"Content-Type": "application/json", "Authorization": f"Bearer {api_key}"}
-        payload = {"model": model_name, "messages": messages, "temperature": 0.7, "max_tokens": max_output_tokens, "response_format": {"type": "json_object"}}
-        
-        try:
-            with httpx.Client(timeout=LLM_TIMEOUT) as client:
-                response = client.post(api_url, json=payload, headers=headers)
-                response.raise_for_status()
-                result = response.json()
-                if result.get("choices") and result["choices"][0].get("message"):
-                    generated_text = result["choices"][0]["message"]["content"][0]["text"]
-                    self.chat_history[conv_id].append({"role": "user", "content": user_message})
-                    self.chat_history[conv_id].append({"role": "assistant", "content": generated_text})
-                    self._save_history()
-                    return self._parse_llm_response(generated_text)
-                return self._create_error_response("Failed to get a valid response from the Databricks API")
-        except httpx.HTTPStatusError as e:
-            error_msg = f"DBX API error: {e.response.status_code}"
-            try:
-                error_detail = e.response.json()
-                if "error" in error_detail:
-                    error_msg += f" - {error_detail['error'].get('message', str(error_detail['error']))}"
-            except: error_msg += f" - {e.response.text[:200]}"
-            return self._create_error_response(error_msg)
-        except Exception as e:
-            return self._create_error_response(f"Error calling Databricks API: {str(e)}")
 
     def _handle_openai_compatible(self, prompt: str, source_files: Dict[str, str], chat_id: Optional[int] = None, system_prompt: Optional[str] = None, max_output_tokens: int = DEFAULT_MAX_OUTPUT_TOKENS) -> Optional[Dict[str, Any]]:
         """Handle OpenAI-compatible API endpoints.
@@ -288,9 +246,6 @@ class LocalModelPlugin(Plugin):
             result = self._handle_openai_compatible(prompt, source_files, chat_id, system_prompt, max_output_tokens)
             if result is not None: return result
 
-            result = self._handle_databricks(prompt, source_files, chat_id, system_prompt, max_output_tokens)
-            if result is not None: return result
-
             if model_id == "google/gemini-2.5-pro":
                 return self._handle_gemini_pro_25(prompt, source_files, chat_id, system_prompt, max_output_tokens)
             

{ayechat_dev-0.36.9.20260204011001.dist-info → ayechat_dev-0.36.9.20260205232002.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ayechat-dev
-Version: 0.36.9.20260204011001
+Version: 0.36.9.20260205232002
 Summary: Aye Chat: Terminal-first AI Code Generator
 Author-email: "Acrotron, Inc." <info@acrotron.com>
 License: MIT

{ayechat_dev-0.36.9.20260204011001.dist-info → ayechat_dev-0.36.9.20260205232002.dist-info}/RECORD

@@ -6,7 +6,7 @@ aye/controller/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 aye/controller/command_handlers.py,sha256=S_HOYY91lbLkGtCkby28T8qUFPNtfy-CbOGn9VGZtTo,17446
 aye/controller/commands.py,sha256=sXmK_sgNBrw9Fs7mKcr93-wsu740ZlvWSisQfS-1EUE,12278
 aye/controller/llm_handler.py,sha256=gY3X2rHcvhPp8iqp_Vor7QpbwHltmcm_Uu9MWF1Z598,7120
-aye/controller/llm_invoker.py,sha256=p_Vk2a3YrWKwDupLfSVRinR5llDfq1Fb_f7WrYozK6M,14127
+aye/controller/llm_invoker.py,sha256=1TMsF5q9HT5r8F4zZ5HQo_g_eYyt7uBX4d46QsUHY4k,14868
 aye/controller/plugin_manager.py,sha256=9ZuITyA5sQJJJU-IntLQ1SsxXsDnbgZKPOF4e9VmsEU,3018
 aye/controller/repl.py,sha256=nSlzAHD8MQlQ7vnpBTLVzA1jQ8jkNjxUnUoqE31chxY,27028
 aye/controller/tutorial.py,sha256=lc92jOcJOYCVrrjTEF0Suk4-8jn-ku98kTJEIL8taUA,7254
@@ -16,7 +16,7 @@ aye/model/api.py,sha256=HhSMQQ_szdC2ZPOSfNsJRbs1FRwb6WyYIeLejB2ScbA,13272
 aye/model/ast_chunker.py,sha256=rVcDdynVUXXyxWVgtUcsee_STqB7SAwP776ktWTlYig,4462
 aye/model/auth.py,sha256=ozV_uQxdqXtUoWO3nZwpzVnDOIfnRAmSMC6W0N724vE,4800
 aye/model/autodiff_config.py,sha256=b8pyudkJFYXF5JWPxft0bH5uazeCab9i-A11dNt1d7U,931
-aye/model/config.py,sha256=o6bQhj5gqhSqtWD6DLow7NNy6Hdaede02h_xb7uPLXo,9280
+aye/model/config.py,sha256=s9c5w4bmOegpNu2CM1z8ITWUbsEzMaX9_BliqHr7KLA,9430
 aye/model/download_plugins.py,sha256=6omyFGdxlEIb7tKPLq4rRVrRYeCPUUCE8aZHvJAKGSc,4442
 aye/model/file_processor.py,sha256=b7YGvHAmhGto9JbtzcfrsdkFtksHbosYt-42EnR22Uo,2131
 aye/model/ignore_patterns.py,sha256=AhcnZuU9_a77Q4yKFRTG6yIrDG4HcECwdY7D1NQkBDY,2767
@@ -43,7 +43,8 @@ aye/plugins/__init__.py,sha256=dSTxs461ICx0O1tbCBCca0W_7QIAa0Yt9PQhHiT5uZQ,173
 aye/plugins/at_file_completer.py,sha256=uNS4gWpfKvn9_nGxZbhQVjVg_S82g977gfBR-pL3XrQ,19582
 aye/plugins/auto_detect_mask.py,sha256=gZKH4qkR-A73uKpMkPXhlgI452Ae_2YG1nHtaIkOvwM,6864
 aye/plugins/completer.py,sha256=qhxke5Q76P2u0LojSIL3V48RTNG5tWL-5-TK5tNutrE,13893
-aye/plugins/local_model.py,sha256=q0RjSjLhEQcDMOCLAK6k1YCW5ECrvdT_g0lKRHMX-AE,14810
+aye/plugins/databricks_model.py,sha256=TpERRIZKh_vJBVu630ModfT2geMXwEEX5XDqzlLt7dI,12494
+aye/plugins/local_model.py,sha256=u3cVLkAD2XGaEvXEFKnmZcDhmqFOi1aqD0ocgysxwCc,12036
 aye/plugins/offline_llm.py,sha256=qFmd1e8Lbl7yiMgXpXjOQkQTNxOk0_WXU7km2DTKXGY,13357
 aye/plugins/plugin_base.py,sha256=t5hTOnA0dZC237BnseAgdXbOqErlSCNLUo_Uul09TSw,1673
 aye/plugins/shell_executor.py,sha256=a0mlZnQeURONdtPM7iageTcQ8PiNLQbjxoY54EsS32o,7502
@@ -54,9 +55,9 @@ aye/presenter/diff_presenter.py,sha256=cbxfOEqGomPTDvQpKdybfYeNUD2DYVAl85j1uy5--
 aye/presenter/repl_ui.py,sha256=PVENlAQM_tm_k2dANsmQH6I8ATMVXhrdj_hNzc38pSw,8156
 aye/presenter/streaming_ui.py,sha256=_3tBEuNH9UQ9Gyq2yuvRfX4SWVkcGMYirEUGj-MXVJ0,12768
 aye/presenter/ui_utils.py,sha256=6KXR4_ZZZUdF5pCHrPqO8yywlQk7AOzWe-2B4Wj_-ZQ,5441
-ayechat_dev-0.36.9.20260204011001.dist-info/licenses/LICENSE,sha256=U1ou6lkMKmPo16-E9YowIu3goU7sOWKUprGo0AOA72s,1065
-ayechat_dev-0.36.9.20260204011001.dist-info/METADATA,sha256=c6g026ZNTxo1IUHsgXnRCWdyJZnX_ddUTwwo4i4mvtc,7718
-ayechat_dev-0.36.9.20260204011001.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
-ayechat_dev-0.36.9.20260204011001.dist-info/entry_points.txt,sha256=KGsOma6szoefNN6vHozg3Pbf1fjZ7ZbmwrOiVwBd0Ik,41
-ayechat_dev-0.36.9.20260204011001.dist-info/top_level.txt,sha256=7WZL0LOx4-GKKvgU1mtI5s4Dhk2OdieVZZvVnxFJHr8,4
-ayechat_dev-0.36.9.20260204011001.dist-info/RECORD,,
+ayechat_dev-0.36.9.20260205232002.dist-info/licenses/LICENSE,sha256=U1ou6lkMKmPo16-E9YowIu3goU7sOWKUprGo0AOA72s,1065
+ayechat_dev-0.36.9.20260205232002.dist-info/METADATA,sha256=pqjcZNYIDp3AmFU5M6Ia7EmlAoIBsNlWbq1Ay2pGLBo,7718
+ayechat_dev-0.36.9.20260205232002.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+ayechat_dev-0.36.9.20260205232002.dist-info/entry_points.txt,sha256=KGsOma6szoefNN6vHozg3Pbf1fjZ7ZbmwrOiVwBd0Ik,41
+ayechat_dev-0.36.9.20260205232002.dist-info/top_level.txt,sha256=7WZL0LOx4-GKKvgU1mtI5s4Dhk2OdieVZZvVnxFJHr8,4
+ayechat_dev-0.36.9.20260205232002.dist-info/RECORD,,