Merge branch 'pricing' into '0.4.0'

Pricing

See merge request aliasrobotics/alias_research/cai!177

pricing.json

@@ -0,0 +1,27 @@
+{
+    "alias0": {
+        "max_tokens": 128000,
+        "max_input_tokens": 200000,
+        "max_output_tokens": 128000,
+        "input_cost_per_token": 0.000005,
+        "output_cost_per_token": 0.00005,
+        "cache_creation_input_token_cost": 0.000005,
+        "cache_read_input_token_cost": 0.0000005,
+        "search_context_cost_per_query": {
+            "search_context_size_low": 1e-2,
+            "search_context_size_medium": 1e-2,
+            "search_context_size_high": 1e-2
+        },
+        "litellm_provider": "openai",
+        "mode": "chat",
+        "supports_function_calling": true,
+        "supports_vision": true,
+        "tool_use_system_prompt_tokens": 159,
+        "supports_assistant_prefill": true,
+        "supports_pdf_input": true,
+        "supports_prompt_caching": true,
+        "supports_response_schema": true,
+        "deprecation_date": "2026-02-01",
+        "supports_tool_choice": true
+    }
+}

src/cai/sdk/agents/models/openai_chatcompletions.py

@@ -14,7 +14,7 @@ import asyncio
 from collections.abc import AsyncIterator, Iterable
 from dataclasses import dataclass, field
 from typing import TYPE_CHECKING, Any, Literal, cast, overload
-from cai.util import get_ollama_api_base, fix_message_list, cli_print_agent_messages, create_agent_streaming_context, update_agent_streaming_content, finish_agent_streaming, calculate_model_cost
+from cai.util import get_ollama_api_base, fix_message_list, cli_print_agent_messages, create_agent_streaming_context, update_agent_streaming_content, finish_agent_streaming, calculate_model_cost, COST_TRACKER
 from cai.util import start_idle_timer, stop_idle_timer, start_active_timer, stop_active_timer
 from wasabi import color
 from cai.sdk.agents.run_to_jsonl import get_session_recorder
@@ -1288,8 +1288,12 @@ class OpenAIChatCompletionsModel(Model):
                 total_cost = calculate_model_cost(model_name, total_input, total_output)
                 
                 # Explicit conversion to float with fallback to ensure they're never None or 0
-                interaction_cost = max(float(interaction_cost if interaction_cost is not None else 0.0), 0.00001)
-                total_cost = max(float(total_cost if total_cost is not None else 0.0), 0.00001)
+                interaction_cost = float(interaction_cost if interaction_cost is not None else 0.0)
+                total_cost = float(total_cost if total_cost is not None else 0.0)
+                
+                # Update the global COST_TRACKER with the cost of this specific interaction
+                if hasattr(COST_TRACKER, "add_interaction_cost") and interaction_cost > 0.0:
+                    COST_TRACKER.add_interaction_cost(interaction_cost)
                 
                 # Store the total cost for future recording
                 self.total_cost = total_cost

src/cai/tools/common.py

@@ -1249,7 +1249,7 @@ def run_command(command, ctf=None, stdout=False,  # pylint: disable=too-many-arg
             command, 
             stdout, 
             timeout, 
-            stream=stream, 
+            stream=True, 
             call_id=call_id,
             tool_name=tool_name,
             workspace_dir=_get_workspace_dir(),

src/cai/util.py

@@ -247,7 +247,6 @@ class CostTracker:
         if os.environ.get("CAI_COST_DISPLAYED", "").lower() == "true":
             return
         print(f"\nTotal CAI Session Cost: ${self.session_total_cost:.6f}")
-    
     def get_model_pricing(self, model_name: str) -> tuple:
         """Get and cache pricing information for a model"""
         # Use the centralized function to standardize model names
@@ -256,8 +255,24 @@ class CostTracker:
         # Check cache first
         if model_name in self.model_pricing_cache:
             return self.model_pricing_cache[model_name]
+
+        # Try to load pricing from local pricing.json first
+        try:
+            pricing_path = pathlib.Path("pricing.json")
+            if pricing_path.exists():
+                with open(pricing_path, "r", encoding="utf-8") as f:
+                    local_pricing = json.load(f)
+                    pricing_info = local_pricing.get("alias0", {})
+                    input_cost = pricing_info.get("input_cost_per_token", 0)
+                    output_cost = pricing_info.get("output_cost_per_token", 0)
+                    
+                    # Cache and return local pricing
+                    self.model_pricing_cache[model_name] = (input_cost, output_cost)
+                    return input_cost, output_cost
+        except Exception as e:
+            print(f"  WARNING: Error loading local pricing.json: {str(e)}")
         
-        # Fetch from LiteLLM API
+        # Fallback to LiteLLM API if local pricing not found
         LITELLM_URL = (
             "https://raw.githubusercontent.com/BerriAI/litellm/main/"
             "model_prices_and_context_window.json"
@@ -280,7 +295,7 @@ class CostTracker:
         except Exception as e:
             print(f"  WARNING: Error fetching model pricing: {str(e)}")
         
-        # Default values if pricing not found
+        # Default values if no pricing found
         default_pricing = (0, 0)
         self.model_pricing_cache[model_name] = default_pricing
         return default_pricing