npcpy 1.2.36__py3-none-any.whl → 1.3.1__py3-none-any.whl

npcpy/__init__.py

@@ -1,6 +1,14 @@
 from . import npc_compiler
 from . import npc_sysenv
 from . import llm_funcs
+from . import ml_funcs
+from . import npc_array
 from . import sql
-from . import work 
-from . import gen
+from . import work
+from . import gen
+
+# Expose key classes at package level
+from .npc_array import NPCArray, ResponseTensor, LazyResult, infer_matrix, ensemble_vote
+from .npc_compiler import NPC, Team, Jinx
+from .llm_funcs import get_llm_response, check_llm_command, execute_llm_command
+from .ml_funcs import fit_model, predict_model, score_model, ensemble_predict

npcpy/gen/image_gen.py

@@ -363,7 +363,8 @@ def generate_image(
     api_url: Optional[str] = None,
     attachments: Union[List[Union[str, bytes, Image.Image]], None] = None,
     save_path: Optional[str] = None,
-    custom_model_path: Optional[str] = None, # <--- NEW: Accept custom_model_path
+    custom_model_path: Optional[str] = None, # <--- NEW: Accept custom_model_path,
+    
 ):
     """
     Unified function to generate or edit images using various providers.
@@ -429,7 +430,9 @@ def generate_image(
             attachments=attachments,
             height=height,
             width=width,
-            n_images=n_images
+            n_images=n_images,
+            api_key=api_key
+            
         )
         all_generated_pil_images.extend(images)
 

npcpy/gen/response.py

@@ -19,9 +19,76 @@ try:
 except ImportError:
     pass
 except OSError:
-    
     pass
 
+# Token costs per 1M tokens (input, output)
+TOKEN_COSTS = {
+    # OpenAI
+    "gpt-4o": (2.50, 10.00),
+    "gpt-4o-mini": (0.15, 0.60),
+    "gpt-4-turbo": (10.00, 30.00),
+    "gpt-3.5-turbo": (0.50, 1.50),
+    "gpt-5": (1.25, 10.00),
+    "gpt-5-mini": (0.25, 2.00),
+    "o1": (15.00, 60.00),
+    "o1-mini": (3.00, 12.00),
+    "o3": (10.00, 40.00),
+    "o3-mini": (1.10, 4.40),
+    "o4-mini": (1.10, 4.40),
+    # Anthropic
+    "claude-3-5-sonnet": (3.00, 15.00),
+    "claude-3-opus": (15.00, 75.00),
+    "claude-3-haiku": (0.25, 1.25),
+    "claude-sonnet-4": (3.00, 15.00),
+    "claude-opus-4": (15.00, 75.00),
+    "claude-opus-4-5": (5.00, 25.00),
+    "claude-sonnet-4-5": (3.00, 15.00),
+    "claude-haiku-4": (0.80, 4.00),
+    # Google
+    "gemini-1.5-pro": (1.25, 5.00),
+    "gemini-1.5-flash": (0.075, 0.30),
+    "gemini-2.0-flash": (0.10, 0.40),
+    "gemini-2.5-pro": (1.25, 10.00),
+    "gemini-2.5-flash": (0.15, 0.60),
+    "gemini-3-pro": (2.00, 12.00),
+    # Groq (free tier limits, paid is cheap)
+    "llama-3": (0.05, 0.08),
+    "llama-3.1": (0.05, 0.08),
+    "llama-3.2": (0.05, 0.08),
+    "llama-4": (0.05, 0.10),
+    "mixtral": (0.24, 0.24),
+    # DeepSeek
+    "deepseek-v3": (0.27, 1.10),
+    "deepseek-r1": (0.55, 2.19),
+    # Mistral
+    "mistral-large": (2.00, 6.00),
+    "mistral-small": (0.20, 0.60),
+    # xAI
+    "grok-2": (2.00, 10.00),
+    "grok-3": (3.00, 15.00),
+}
+
+def calculate_cost(model: str, input_tokens: int, output_tokens: int) -> float:
+    """Calculate cost in USD for a response."""
+    if not model:
+        return 0.0
+
+    # Normalize model name - strip provider prefix and lowercase
+    model_key = model.split("/")[-1].lower()
+
+    # Check for exact or partial match
+    costs = None
+    for key, cost in TOKEN_COSTS.items():
+        if key in model_key or model_key in key:
+            costs = cost
+            break
+
+    if not costs:
+        return 0.0  # Unknown/local model, assume free
+
+    input_cost, output_cost = costs
+    return (input_tokens * input_cost / 1_000_000) + (output_tokens * output_cost / 1_000_000)
+
 def handle_streaming_json(api_params):
     """
     Handles streaming responses when JSON format is requested from LiteLLM.
@@ -117,7 +184,7 @@ Do not include any additional markdown formatting or leading ```json tags in you
        
        if detected_tools:
            result["tool_calls"] = detected_tools
-           result = process_tool_calls(result, tool_map, "local", "transformers", result["messages"])
+           result = process_tool_calls(result, tool_map, "local", "transformers", result["messages"], tools=tools)
    
    if format == "json":
        try:
@@ -278,12 +345,21 @@ def get_ollama_response(
     if not auto_process_tool_calls or not (tools and tool_map):
         res = ollama.chat(**api_params, options=options)
         result["raw_response"] = res
-        
+
+        # Extract usage from ollama response
+        if hasattr(res, 'prompt_eval_count') or 'prompt_eval_count' in res:
+            input_tokens = getattr(res, 'prompt_eval_count', None) or res.get('prompt_eval_count', 0) or 0
+            output_tokens = getattr(res, 'eval_count', None) or res.get('eval_count', 0) or 0
+            result["usage"] = {
+                "input_tokens": input_tokens,
+                "output_tokens": output_tokens,
+            }
+
         if stream:
-            result["response"] = res  
+            result["response"] = res
             return result
         else:
-            
+
             message = res.get("message", {})
             response_content = message.get("content", "")
             result["response"] = response_content
@@ -333,11 +409,12 @@ def get_ollama_response(
         }
         
         
-        processed_result = process_tool_calls(response_for_processing, 
-                                              tool_map, model, 
-                                              'ollama', 
-                                              messages, 
-                                              stream=False)
+        processed_result = process_tool_calls(response_for_processing,
+                                              tool_map, model,
+                                              'ollama',
+                                              messages,
+                                              stream=False,
+                                              tools=tools)
         
         
         if stream:
@@ -624,7 +701,20 @@ def get_litellm_response(
         api_params["stream"] = stream
         resp = completion(**api_params)
         result["raw_response"] = resp
-        
+
+        # Extract usage if available (handles both standard litellm and ollama formats)
+        if hasattr(resp, 'usage') and resp.usage:
+            result["usage"] = {
+                "input_tokens": getattr(resp.usage, 'prompt_tokens', 0) or 0,
+                "output_tokens": getattr(resp.usage, 'completion_tokens', 0) or 0,
+            }
+        elif hasattr(resp, 'prompt_eval_count'):
+            # Ollama format
+            result["usage"] = {
+                "input_tokens": getattr(resp, 'prompt_eval_count', 0) or 0,
+                "output_tokens": getattr(resp, 'eval_count', 0) or 0,
+            }
+
         if stream:
             result["response"] = resp  
             return result
@@ -687,59 +777,99 @@ def get_litellm_response(
     
     initial_api_params = api_params.copy()
     initial_api_params["stream"] = False
-    
-    
-    resp = completion(**initial_api_params)
+
+    try:
+        resp = completion(**initial_api_params)
+    except Exception as e:
+        from termcolor import colored
+        print(colored(f"[litellm ERROR] completion() failed: {type(e).__name__}: {e}", "red"))
+        result["error"] = str(e)
+        result["response"] = f"LLM call failed: {e}"
+        return result
+
     result["raw_response"] = resp
-    
-    
+
+    # Extract usage if available
+    if hasattr(resp, 'usage') and resp.usage:
+        result["usage"] = {
+            "input_tokens": getattr(resp.usage, 'prompt_tokens', 0) or 0,
+            "output_tokens": getattr(resp.usage, 'completion_tokens', 0) or 0,
+        }
+
+    if not resp.choices:
+        result["response"] = "No response from model"
+        return result
+
     has_tool_calls = hasattr(resp.choices[0].message, 'tool_calls') and resp.choices[0].message.tool_calls
     
     if has_tool_calls:
-
-        
         result["tool_calls"] = resp.choices[0].message.tool_calls
-        
-        
-        processed_result = process_tool_calls(result, 
-                                              tool_map, 
-                                              model, 
-                                              provider, 
-                                              result["messages"], 
-                                              stream=False)
-        
-        
-        if stream:
 
-            
+        processed_result = process_tool_calls(result,
+                                              tool_map,
+                                              model,
+                                              provider,
+                                              result["messages"],
+                                              stream=False,
+                                              tools=tools)
 
-            clean_messages = []
-            for msg in processed_result["messages"]:
-                if msg.get('role') == 'assistant' and 'tool_calls' in msg:
-                    continue  
-                                
-                else:
-                    clean_messages.append(msg)
-            
-            final_api_params = api_params.copy()
-            final_api_params["messages"] = clean_messages
-            final_api_params["stream"] = True
+        # Always do a follow-up call to get a proper response after tool execution
+        # Convert tool interactions to a clean format for the follow-up call
+        clean_messages = []
+        tool_results_summary = []
+
+        for msg in processed_result["messages"]:
+            role = msg.get('role', '')
+            if role == 'assistant' and 'tool_calls' in msg:
+                # Skip the tool_calls message - we'll summarize results instead
+                continue
+            elif role == 'tool':
+                # Collect tool results for summary
+                content = msg.get('content', '')
+                # Truncate very long results
+                if len(content) > 2000:
+                    content = content[:2000] + "... (truncated)"
+                tool_results_summary.append(content)
+            else:
+                clean_messages.append(msg)
 
+        # Add tool results as an assistant message summarizing what was done
+        if tool_results_summary:
+            clean_messages.append({
+                "role": "assistant",
+                "content": "I executed the requested tools. Here are the results:\n\n" + "\n\n".join(tool_results_summary)
+            })
 
-            final_api_params = api_params.copy()
-            final_api_params["messages"] = clean_messages
-            final_api_params["stream"] = True
-            if "tools" in final_api_params:
-                del final_api_params["tools"]
-            if "tool_choice" in final_api_params:
-                del final_api_params["tool_choice"]
+        # Add instruction for the LLM to provide a helpful response
+        clean_messages.append({
+            "role": "user",
+            "content": "Based on the tool results above, provide a brief summary of what happened. Do NOT output any code - the tool has already executed. Just describe the results concisely."
+        })
 
-            final_stream = completion(**final_api_params)
+        final_api_params = api_params.copy()
+        final_api_params["messages"] = clean_messages
+        final_api_params["stream"] = stream
+        if "tools" in final_api_params:
+            del final_api_params["tools"]
+        if "tool_choice" in final_api_params:
+            del final_api_params["tool_choice"]
+
+        final_resp = completion(**final_api_params)
+
+        if stream:
+            processed_result["response"] = final_resp
+        else:
+            if final_resp.choices:
+                final_content = final_resp.choices[0].message.content
+                processed_result["response"] = final_content
+                processed_result["messages"].append({"role": "assistant", "content": final_content})
+            else:
+                # No choices returned, use the tool results summary directly
+                if tool_results_summary:
+                    processed_result["response"] = "\n\n".join(tool_results_summary)
+                else:
+                    processed_result["response"] = "Tool executed successfully."
 
-            
-            final_stream = completion(**final_api_params)
-            processed_result["response"] = final_stream
-            
         return processed_result
         
         
@@ -770,18 +900,41 @@ def get_litellm_response(
         else:
             result["response"] = llm_response
     return result            
-def process_tool_calls(response_dict, tool_map, model, provider, messages, stream=False):
+def process_tool_calls(response_dict, tool_map, model, provider, messages, stream=False, tools=None):
     result = response_dict.copy()
     result["tool_results"] = []
-    
+
     if "messages" not in result:
         result["messages"] = messages if messages else []
-    
+
     tool_calls = result.get("tool_calls", [])
-    
+
     if not tool_calls:
         return result
-    
+
+    # First, add the assistant message with tool_calls (required by Gemini and other providers)
+    # This must come BEFORE the tool results
+    tool_calls_for_message = []
+    for tc in tool_calls:
+        if isinstance(tc, dict):
+            tool_calls_for_message.append(tc)
+        else:
+            # Convert object to dict format
+            tool_calls_for_message.append({
+                "id": getattr(tc, "id", str(uuid.uuid4())),
+                "type": "function",
+                "function": {
+                    "name": getattr(tc.function, "name", "") if hasattr(tc, "function") else "",
+                    "arguments": getattr(tc.function, "arguments", "{}") if hasattr(tc, "function") else "{}"
+                }
+            })
+
+    result["messages"].append({
+        "role": "assistant",
+        "content": None,
+        "tool_calls": tool_calls_for_message
+    })
+
     for tool_call in tool_calls:
         tool_id = str(uuid.uuid4())
         tool_name = None
@@ -812,10 +965,54 @@ def process_tool_calls(response_dict, tool_map, model, provider, messages, strea
             tool_result_str = ""
             serializable_result = None
 
+            # Show tool execution indicator with truncated args
+            # Store full args for Ctrl+O expansion
+            _last_tool_call = {"name": tool_name, "arguments": arguments}
+            try:
+                import builtins
+                builtins._npcsh_last_tool_call = _last_tool_call
+            except:
+                pass
+
+            try:
+                from termcolor import colored
+                # Format arguments nicely - show key=value pairs
+                is_truncated = False
+                if arguments:
+                    arg_parts = []
+                    for k, v in arguments.items():
+                        v_str = str(v)
+                        if len(v_str) > 40:
+                            v_str = v_str[:40] + "…"
+                            is_truncated = True
+                        arg_parts.append(f"{v_str}")
+                    args_display = " ".join(arg_parts)
+                    if len(args_display) > 60:
+                        args_display = args_display[:60] + "…"
+                        is_truncated = True
+                else:
+                    args_display = ""
+
+                if args_display:
+                    hint = colored(" [^O]", "white", attrs=["dark"]) if is_truncated else ""
+                    print(colored(f"  ⚡ {tool_name}", "cyan") + colored(f" {args_display}", "white", attrs=["dark"]) + hint, end="", flush=True)
+                else:
+                    print(colored(f"  ⚡ {tool_name}", "cyan"), end="", flush=True)
+            except:
+                pass
+
             try:
                 tool_result = tool_map[tool_name](**arguments)
+                try:
+                    print(colored(" ✓", "green"), flush=True)
+                except:
+                    pass
             except Exception as e:
-                tool_result = f"Error executing tool '{tool_name}': {str(e)}. Tool map is : {tool_map}"
+                tool_result = f"Error executing tool '{tool_name}': {str(e)}"
+                try:
+                    print(colored(f" ✗ {str(e)[:50]}", "red"), flush=True)
+                except:
+                    pass
 
             try:
                 tool_result_str = json.dumps(tool_result, default=str)
@@ -826,18 +1023,19 @@ def process_tool_calls(response_dict, tool_map, model, provider, messages, strea
             except Exception as e_serialize:
                 tool_result_str = f"Error serializing result for {tool_name}: {str(e_serialize)}"
                 serializable_result = {"error": tool_result_str}
-            
+
             result["tool_results"].append({
                 "tool_call_id": tool_id,
                 "tool_name": tool_name,
                 "arguments": arguments,
                 "result": serializable_result
             })
-            
-            
+
+            # Add tool result as a tool message (proper format for multi-turn)
             result["messages"].append({
-                "role": "assistant",
-                "content": f'The results of the tool call for {tool_name} with {arguments} are as follows:' +tool_result_str
+                "role": "tool",
+                "tool_call_id": tool_id,
+                "content": tool_result_str
             })
     
     return result