mirror of https://github.com/aliasrobotics/cai.git
Fix stream plus reason
This commit is contained in:
luijait
parent
e740d6499d
commit
423c80585a
|
|
@ -994,12 +994,14 @@ class OpenAIChatCompletionsModel(Model):
|
|||
|
||||
# Handle Claude reasoning content first (before regular content)
|
||||
reasoning_content = None
|
||||
|
||||
# Check for Claude reasoning in different possible formats
|
||||
if hasattr(delta, 'reasoning_content') and delta.reasoning_content is not None:
|
||||
reasoning_content = delta.reasoning_content
|
||||
elif isinstance(delta, dict) and 'reasoning_content' in delta and delta['reasoning_content'] is not None:
|
||||
reasoning_content = delta['reasoning_content']
|
||||
|
||||
# Also check for thinking_blocks structure
|
||||
# Also check for thinking_blocks structure (Claude 4 format)
|
||||
thinking_blocks = None
|
||||
if hasattr(delta, 'thinking_blocks') and delta.thinking_blocks is not None:
|
||||
thinking_blocks = delta.thinking_blocks
|
||||
|
|
@ -1012,11 +1014,29 @@ class OpenAIChatCompletionsModel(Model):
|
|||
if isinstance(block, dict) and block.get('type') == 'thinking':
|
||||
reasoning_content = block.get('thinking', '')
|
||||
break
|
||||
elif isinstance(block, dict) and block.get('type') == 'text' and 'thinking' in str(block):
|
||||
# Sometimes thinking content comes as text blocks
|
||||
reasoning_content = block.get('text', '')
|
||||
break
|
||||
|
||||
if reasoning_content and thinking_context:
|
||||
# Update the thinking display
|
||||
from cai.util import update_claude_thinking_content
|
||||
update_claude_thinking_content(thinking_context, reasoning_content)
|
||||
# Check for direct thinking field (some Claude models)
|
||||
if not reasoning_content:
|
||||
if hasattr(delta, 'thinking') and delta.thinking is not None:
|
||||
reasoning_content = delta.thinking
|
||||
elif isinstance(delta, dict) and 'thinking' in delta and delta['thinking'] is not None:
|
||||
reasoning_content = delta['thinking']
|
||||
|
||||
# Update thinking display if we have reasoning content
|
||||
if reasoning_content:
|
||||
if thinking_context:
|
||||
# Streaming mode: Update the rich thinking display
|
||||
from cai.util import update_claude_thinking_content
|
||||
update_claude_thinking_content(thinking_context, reasoning_content)
|
||||
else:
|
||||
# Non-streaming mode: Use simple text output
|
||||
from cai.util import print_claude_reasoning_simple, detect_claude_thinking_in_stream
|
||||
if detect_claude_thinking_in_stream(str(self.model)):
|
||||
print_claude_reasoning_simple(reasoning_content, self.agent_name, str(self.model))
|
||||
|
||||
|
||||
|
||||
|
|
@ -1028,6 +1048,14 @@ class OpenAIChatCompletionsModel(Model):
|
|||
content = delta['content']
|
||||
|
||||
if content:
|
||||
# IMPORTANT: If we have content and thinking_context is active,
|
||||
# it means thinking is complete and normal content is starting
|
||||
# Close the thinking display automatically
|
||||
if thinking_context:
|
||||
from cai.util import finish_claude_thinking_display
|
||||
finish_claude_thinking_display(thinking_context)
|
||||
thinking_context = None # Clear the context
|
||||
|
||||
# For Ollama, we need to accumulate the full content to check for function calls
|
||||
if is_ollama:
|
||||
ollama_full_content += content
|
||||
|
|
@ -1035,10 +1063,9 @@ class OpenAIChatCompletionsModel(Model):
|
|||
# Add to the streaming text buffer
|
||||
streaming_text_buffer += content
|
||||
|
||||
# Update streaming display if enabled - always do this for text content
|
||||
# BUT: if thinking context is active, don't show text content during streaming
|
||||
# Instead, it will be shown after thinking completes
|
||||
if streaming_context and not thinking_context:
|
||||
# Update streaming display if enabled - ALWAYS respect CAI_STREAM setting
|
||||
# Both thinking and regular content should stream if streaming is enabled
|
||||
if streaming_context:
|
||||
# Check if this is a local model and should have zero cost
|
||||
model_str = str(self.model).lower()
|
||||
is_local_model = (
|
||||
|
|
@ -1749,41 +1776,7 @@ class OpenAIChatCompletionsModel(Model):
|
|||
from cai.util import finish_claude_thinking_display
|
||||
finish_claude_thinking_display(thinking_context)
|
||||
|
||||
# After thinking completes, display the text content if it exists
|
||||
# This ensures that Claude thinking models show their final response
|
||||
if (state.text_content_index_and_output and
|
||||
state.text_content_index_and_output[1].text and
|
||||
not streamed_tool_calls): # Only show if there were no tool calls
|
||||
|
||||
# Create a message-like object for displaying the text response
|
||||
text_msg = type('TextResponse', (), {
|
||||
'content': state.text_content_index_and_output[1].text,
|
||||
'tool_calls': None
|
||||
})
|
||||
|
||||
# Display the text response using the CLI utility
|
||||
cli_print_agent_messages(
|
||||
agent_name=getattr(self, 'agent_name', 'Agent'),
|
||||
message=text_msg,
|
||||
counter=getattr(self, 'interaction_counter', 0),
|
||||
model=str(self.model),
|
||||
debug=False,
|
||||
interaction_input_tokens=interaction_input,
|
||||
interaction_output_tokens=interaction_output,
|
||||
interaction_reasoning_tokens=int(
|
||||
final_response.usage.output_tokens_details.reasoning_tokens
|
||||
if final_response.usage and final_response.usage.output_tokens_details
|
||||
and hasattr(final_response.usage.output_tokens_details, 'reasoning_tokens')
|
||||
else 0
|
||||
),
|
||||
total_input_tokens=total_input,
|
||||
total_output_tokens=total_output,
|
||||
total_reasoning_tokens=getattr(self, 'total_reasoning_tokens', 0),
|
||||
interaction_cost=interaction_cost,
|
||||
total_cost=total_cost,
|
||||
tool_output=None,
|
||||
suppress_empty=True
|
||||
)
|
||||
# Note: Content is now displayed during streaming, no need to show it again here
|
||||
|
||||
if tracing.include_data():
|
||||
span_generation.span_data.output = [final_response.model_dump()]
|
||||
|
|
@ -2065,7 +2058,15 @@ class OpenAIChatCompletionsModel(Model):
|
|||
|
||||
# Add extended reasoning support for Claude models
|
||||
# Supports Claude 3.7, Claude 4, and any model with "thinking" in the name
|
||||
has_reasoning_capability = "thinking" in model_str
|
||||
has_reasoning_capability = (
|
||||
"thinking" in model_str or
|
||||
# Claude 4 models support reasoning
|
||||
"-4-" in model_str or
|
||||
"sonnet-4" in model_str or
|
||||
"haiku-4" in model_str or
|
||||
"opus-4" in model_str or
|
||||
"3.7" in model_str
|
||||
)
|
||||
|
||||
if has_reasoning_capability:
|
||||
# Clean the model name by removing "thinking" before sending to API
|
||||
|
|
|
|||
|
|
@ -341,23 +341,6 @@ class CostTracker:
|
|||
if model_name in self.model_pricing_cache:
|
||||
return self.model_pricing_cache[model_name]
|
||||
|
||||
# Try to load pricing from remote pricing.json first
|
||||
try:
|
||||
import requests
|
||||
pricing_url = "https://raw.githubusercontent.com/aliasrobotics/cai/refs/heads/mcp_stdio/pricing.json"
|
||||
response = requests.get(pricing_url, timeout=5)
|
||||
if response.status_code == 200:
|
||||
remote_pricing = response.json()
|
||||
pricing_info = remote_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 remote 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 remote pricing.json: {str(e)}")
|
||||
|
||||
# Fallback to LiteLLM API if both remote and local pricing not found
|
||||
LITELLM_URL = (
|
||||
"https://raw.githubusercontent.com/BerriAI/litellm/main/"
|
||||
|
|
@ -3276,20 +3259,46 @@ def detect_claude_thinking_in_stream(model_name):
|
|||
model_str = str(model_name).lower()
|
||||
|
||||
# Check for Claude models with reasoning capability
|
||||
# According to LiteLLM docs, Claude 3.7, Claude 4 and models with "thinking" support reasoning
|
||||
# Claude 4 models (like claude-sonnet-4-20250514) support reasoning
|
||||
# Also check for explicit "thinking" in model name
|
||||
has_reasoning = (
|
||||
"claude" in model_str and (
|
||||
# Claude 4 models (sonnet-4, haiku-4, opus-4)
|
||||
"-4-" in model_str or
|
||||
"sonnet-4" in model_str or
|
||||
"haiku-4" in model_str or
|
||||
"opus-4" in model_str or
|
||||
# Legacy support for 3.7 and explicit thinking models
|
||||
"3.7" in model_str or
|
||||
"4" in model_str or
|
||||
"thinking" in model_str
|
||||
)
|
||||
)
|
||||
|
||||
return has_reasoning
|
||||
|
||||
def print_claude_reasoning_simple(reasoning_content, agent_name, model_name):
|
||||
"""
|
||||
Print Claude reasoning content in simple mode (no Rich panels).
|
||||
Used when CAI_STREAM=False.
|
||||
|
||||
Args:
|
||||
reasoning_content: The reasoning/thinking text
|
||||
agent_name: The agent name
|
||||
model_name: The model name
|
||||
"""
|
||||
if not reasoning_content or not reasoning_content.strip():
|
||||
return
|
||||
|
||||
# Simple text output without Rich formatting
|
||||
timestamp = datetime.now().strftime("%H:%M:%S")
|
||||
print(f"\n🧠 Reasoning | {agent_name} | {model_name} | {timestamp}")
|
||||
print("=" * 60)
|
||||
print(reasoning_content)
|
||||
print("=" * 60 + "\n")
|
||||
|
||||
def start_claude_thinking_if_applicable(model_name, agent_name, counter):
|
||||
"""
|
||||
Start Claude thinking display if the model supports it.
|
||||
Start Claude thinking display if the model supports it AND streaming is enabled.
|
||||
|
||||
Args:
|
||||
model_name: The model name
|
||||
|
|
@ -3299,6 +3308,9 @@ def start_claude_thinking_if_applicable(model_name, agent_name, counter):
|
|||
Returns:
|
||||
The thinking context if created, None otherwise
|
||||
"""
|
||||
if detect_claude_thinking_in_stream(model_name):
|
||||
# Only show thinking in streaming mode
|
||||
streaming_enabled = os.getenv('CAI_STREAM', 'false').lower() == 'true'
|
||||
|
||||
if streaming_enabled and detect_claude_thinking_in_stream(model_name):
|
||||
return create_claude_thinking_context(agent_name, counter, model_name)
|
||||
return None
|
||||
|
|
|
|||
Loading…
Reference in New Issue