code-puppy 0.0.318__py3-none-any.whl → 0.0.320__py3-none-any.whl

code_puppy/agents/base_agent.py

@@ -1267,13 +1267,9 @@ class BaseAgent(ABC):
             ctx: The run context.
             events: Async iterable of streaming events (PartStartEvent, PartDeltaEvent, etc.).
         """
-        import os
-        import time as time_module
-
         from pydantic_ai import PartDeltaEvent, PartStartEvent
         from pydantic_ai.messages import TextPartDelta, ThinkingPartDelta
         from rich.console import Console
-        from rich.live import Live
         from rich.markdown import Markdown
         from rich.markup import escape
 
@@ -1288,14 +1284,6 @@ class BaseAgent(ABC):
             # Fallback if console not set (shouldn't happen in normal use)
             console = Console()
 
-        # Disable Live display in test mode or non-interactive environments
-        # This fixes issues with pexpect PTY where Live() hangs
-        use_live_display = (
-            console.is_terminal
-            and os.environ.get("CODE_PUPPY_TEST_FAST", "").lower() not in ("1", "true")
-            and os.environ.get("CI", "").lower() not in ("1", "true")
-        )
-
         # Track which part indices we're currently streaming (for Text/Thinking parts)
         streaming_parts: set[int] = set()
         thinking_parts: set[int] = (
@@ -1303,11 +1291,9 @@ class BaseAgent(ABC):
         )  # Track which parts are thinking (for dim style)
         text_parts: set[int] = set()  # Track which parts are text
         banner_printed: set[int] = set()  # Track if banner was already printed
-        text_buffer: dict[int, list[str]] = {}  # Buffer text for markdown
-        live_displays: dict[int, Live] = {}  # Live displays for streaming markdown
+        text_buffer: dict[int, list[str]] = {}  # Buffer text for final markdown render
+        token_count: dict[int, int] = {}  # Track token count per text part
         did_stream_anything = False  # Track if we streamed any content
-        last_render_time: dict[int, float] = {}  # Track last render time per part
-        render_interval = 0.1  # Only re-render markdown every 100ms (throttle)
 
         def _print_thinking_banner() -> None:
             """Print the THINKING banner with spinner pause and line clear."""
@@ -1372,9 +1358,11 @@ class BaseAgent(ABC):
                     streaming_parts.add(event.index)
                     text_parts.add(event.index)
                     text_buffer[event.index] = []  # Initialize buffer
+                    token_count[event.index] = 0  # Initialize token counter
                     # Buffer initial content if present
                     if part.content and part.content.strip():
                         text_buffer[event.index].append(part.content)
+                        token_count[event.index] += 1
 
             # PartDeltaEvent - stream the content as it arrives
             elif isinstance(event, PartDeltaEvent):
@@ -1382,43 +1370,23 @@ class BaseAgent(ABC):
                     delta = event.delta
                     if isinstance(delta, (TextPartDelta, ThinkingPartDelta)):
                         if delta.content_delta:
-                            # For text parts, stream markdown with Live display
+                            # For text parts, show token counter then render at end
                             if event.index in text_parts:
-                                # Print banner and start Live on first content
+                                import sys
+
+                                # Print banner on first content
                                 if event.index not in banner_printed:
                                     _print_response_banner()
                                     banner_printed.add(event.index)
-                                    # Only use Live display if enabled (disabled in test/CI)
-                                    if use_live_display:
-                                        live = Live(
-                                            Markdown(""),
-                                            console=console,
-                                            refresh_per_second=10,
-                                            vertical_overflow="visible",  # Allow scrolling for long content
-                                        )
-                                        live.start()
-                                        live_displays[event.index] = live
-                                # Accumulate text and throttle markdown rendering
-                                # (Markdown parsing is O(n), doing it on every token = O(n²) death)
+                                # Accumulate text for final markdown render
                                 text_buffer[event.index].append(delta.content_delta)
-                                now = time_module.monotonic()
-                                last_render = last_render_time.get(event.index, 0)
-
-                                # Only re-render if enough time has passed (throttle)
-                                # Skip Live updates when not using live display
-                                if (
-                                    use_live_display
-                                    and now - last_render >= render_interval
-                                ):
-                                    content = "".join(text_buffer[event.index])
-                                    if event.index in live_displays:
-                                        try:
-                                            live_displays[event.index].update(
-                                                Markdown(content)
-                                            )
-                                            last_render_time[event.index] = now
-                                        except Exception:
-                                            pass
+                                token_count[event.index] += 1
+                                # Update token counter in place (single line)
+                                count = token_count[event.index]
+                                sys.stdout.write(
+                                    f"\r\x1b[K  ⏳ Receiving... {count} tokens"
+                                )
+                                sys.stdout.flush()
                             else:
                                 # For thinking parts, stream immediately (dim)
                                 if event.index not in banner_printed:
@@ -1430,36 +1398,24 @@ class BaseAgent(ABC):
             # PartEndEvent - finish the streaming with a newline
             elif isinstance(event, PartEndEvent):
                 if event.index in streaming_parts:
-                    # For text parts, do final render then stop the Live display
+                    # For text parts, clear counter line and render markdown
                     if event.index in text_parts:
-                        # Final render to ensure we show complete content
-                        # (throttling may have skipped the last few tokens)
-                        if event.index in live_displays and event.index in text_buffer:
-                            try:
-                                final_content = "".join(text_buffer[event.index])
-                                live_displays[event.index].update(
-                                    Markdown(final_content)
-                                )
-                            except Exception:
-                                pass
-                        if event.index in live_displays:
-                            try:
-                                live_displays[event.index].stop()
-                            except Exception:
-                                pass
-                            del live_displays[event.index]
-                        # When not using Live display, print the final content as markdown
-                        elif event.index in text_buffer:
+                        import sys
+
+                        # Clear the token counter line
+                        sys.stdout.write("\r\x1b[K")
+                        sys.stdout.flush()
+                        # Render the final markdown nicely
+                        if event.index in text_buffer:
                             try:
                                 final_content = "".join(text_buffer[event.index])
                                 if final_content.strip():
                                     console.print(Markdown(final_content))
                             except Exception:
                                 pass
-                        if event.index in text_buffer:
                             del text_buffer[event.index]
-                        # Clean up render time tracking
-                        last_render_time.pop(event.index, None)
+                        # Clean up token count
+                        token_count.pop(event.index, None)
                     # For thinking parts, just print newline
                     elif event.index in banner_printed:
                         console.print()  # Final newline after streaming

code_puppy/command_line/model_settings_menu.py

@@ -84,6 +84,12 @@ SETTING_DEFINITIONS: Dict[str, Dict] = {
         "default": 10000,
         "format": "{:.0f}",
     },
+    "interleaved_thinking": {
+        "name": "Interleaved Thinking",
+        "description": "Enable thinking between tool calls (Claude 4 only: Opus 4.5, Opus 4.1, Opus 4, Sonnet 4). Adds beta header. WARNING: On Vertex/Bedrock, this FAILS for non-Claude 4 models!",
+        "type": "boolean",
+        "default": False,
+    },
 }
 
 

code_puppy/model_factory.py

@@ -319,9 +319,21 @@ class ModelFactory:
                 http2=http2_enabled,
             )
 
+            # Check if interleaved thinking is enabled for this model
+            # Only applies to Claude 4 models (Opus 4.5, Opus 4.1, Opus 4, Sonnet 4)
+            from code_puppy.config import get_effective_model_settings
+
+            effective_settings = get_effective_model_settings(model_name)
+            interleaved_thinking = effective_settings.get("interleaved_thinking", False)
+
+            default_headers = {}
+            if interleaved_thinking:
+                default_headers["anthropic-beta"] = "interleaved-thinking-2025-05-14"
+
             anthropic_client = AsyncAnthropic(
                 api_key=api_key,
                 http_client=client,
+                default_headers=default_headers if default_headers else None,
             )
 
             # Ensure cache_control is injected at the Anthropic SDK layer
@@ -351,10 +363,21 @@ class ModelFactory:
                 http2=http2_enabled,
             )
 
+            # Check if interleaved thinking is enabled for this model
+            from code_puppy.config import get_effective_model_settings
+
+            effective_settings = get_effective_model_settings(model_name)
+            interleaved_thinking = effective_settings.get("interleaved_thinking", False)
+
+            default_headers = {}
+            if interleaved_thinking:
+                default_headers["anthropic-beta"] = "interleaved-thinking-2025-05-14"
+
             anthropic_client = AsyncAnthropic(
                 base_url=url,
                 http_client=client,
                 api_key=api_key,
+                default_headers=default_headers if default_headers else None,
             )
 
             # Ensure cache_control is injected at the Anthropic SDK layer
@@ -370,6 +393,31 @@ class ModelFactory:
                 )
                 return None
 
+            # Check if interleaved thinking is enabled (defaults to True for OAuth models)
+            from code_puppy.config import get_effective_model_settings
+
+            effective_settings = get_effective_model_settings(model_name)
+            interleaved_thinking = effective_settings.get("interleaved_thinking", True)
+
+            # Handle anthropic-beta header based on interleaved_thinking setting
+            if "anthropic-beta" in headers:
+                beta_parts = [p.strip() for p in headers["anthropic-beta"].split(",")]
+                if interleaved_thinking:
+                    # Ensure interleaved-thinking is in the header
+                    if "interleaved-thinking-2025-05-14" not in beta_parts:
+                        beta_parts.append("interleaved-thinking-2025-05-14")
+                else:
+                    # Remove interleaved-thinking from the header
+                    beta_parts = [
+                        p for p in beta_parts if "interleaved-thinking" not in p
+                    ]
+                headers["anthropic-beta"] = ",".join(beta_parts) if beta_parts else None
+                if headers.get("anthropic-beta") is None:
+                    del headers["anthropic-beta"]
+            elif interleaved_thinking:
+                # No existing beta header, add one for interleaved thinking
+                headers["anthropic-beta"] = "interleaved-thinking-2025-05-14"
+
             # Use a dedicated client wrapper that injects cache_control on /v1/messages
             if verify is None:
                 verify = get_cert_bundle_path()

code_puppy/models.json

@@ -81,7 +81,7 @@
     "type": "anthropic",
     "name": "claude-opus-4-5",
     "context_length": 200000,
-    "supported_settings": ["temperature", "extended_thinking", "budget_tokens"]
+    "supported_settings": ["temperature", "extended_thinking", "budget_tokens", "interleaved_thinking"]
   },
   "zai-glm-4.6-coding": {
     "type": "zai_coding",

code_puppy/plugins/claude_code_oauth/utils.py

@@ -368,6 +368,7 @@ def add_models_to_extra_config(models: List[str]) -> bool:
                     "temperature",
                     "extended_thinking",
                     "budget_tokens",
+                    "interleaved_thinking",
                 ],
             }
             added += 1

{code_puppy-0.0.318.data → code_puppy-0.0.320.data}/data/code_puppy/models.json

@@ -81,7 +81,7 @@
     "type": "anthropic",
     "name": "claude-opus-4-5",
     "context_length": 200000,
-    "supported_settings": ["temperature", "extended_thinking", "budget_tokens"]
+    "supported_settings": ["temperature", "extended_thinking", "budget_tokens", "interleaved_thinking"]
   },
   "zai-glm-4.6-coding": {
     "type": "zai_coding",

{code_puppy-0.0.318.dist-info → code_puppy-0.0.320.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code-puppy
-Version: 0.0.318
+Version: 0.0.320
 Summary: Code generation agent
 Project-URL: repository, https://github.com/mpfaffenberger/code_puppy
 Project-URL: HomePage, https://github.com/mpfaffenberger/code_puppy

{code_puppy-0.0.318.dist-info → code_puppy-0.0.320.dist-info}/RECORD

@@ -10,9 +10,9 @@ code_puppy/gemini_code_assist.py,sha256=KGS7sO5OLc83nDF3xxS-QiU6vxW9vcm6hmzilu79
 code_puppy/http_utils.py,sha256=w5mWYIGIWJZJvgvMahXs9BmdidoJvGn4CASDRY88a8o,13414
 code_puppy/keymap.py,sha256=Uzvq7HB-6inTjKox-90JWzuijztRdWqhJpfTDZVy5no,3235
 code_puppy/main.py,sha256=82r3vZy_XcyEsenLn82BnUusaoyL3Bpm_Th_jKgqecE,273
-code_puppy/model_factory.py,sha256=2jXTpi3BuFO8RtbhKKRvb5EN02w_OKzQSdsTdSy78X0,31608
+code_puppy/model_factory.py,sha256=H_a5nX462Q-dhX3g3ZY7dmBCIAUOd1aOSZa4HMxF1o4,34191
 code_puppy/model_utils.py,sha256=NU8W8NW5F7QS_PXHaLeh55Air1koUV7IVYFP7Rz3XpY,3615
-code_puppy/models.json,sha256=nbMCW9SJxWptVEOeMrn_gfyDp-lsbM7Chczjl944GKA,3077
+code_puppy/models.json,sha256=mTpmJH0UJlmX8M2KVPbxMWb99de3IxKXCWO-B23b6xo,3101
 code_puppy/models_dev_api.json,sha256=wHjkj-IM_fx1oHki6-GqtOoCrRMR0ScK0f-Iz0UEcy8,548187
 code_puppy/models_dev_parser.py,sha256=8ndmWrsSyKbXXpRZPXc0w6TfWMuCcgaHiMifmlaBaPc,20611
 code_puppy/pydantic_patches.py,sha256=YecAEeCOjSIwIBu2O5vEw72atMSL37cXGrbEuukI07o,4582
@@ -39,7 +39,7 @@ code_puppy/agents/agent_qa_expert.py,sha256=5Ikb4U3SZQknUEfwlHZiyZXKqnffnOTQagr_
 code_puppy/agents/agent_qa_kitten.py,sha256=5PeFFSwCFlTUvP6h5bGntx0xv5NmRwBiw0HnMqY8nLI,9107
 code_puppy/agents/agent_security_auditor.py,sha256=SpiYNA0XAsIwBj7S2_EQPRslRUmF_-b89pIJyW7DYtY,12022
 code_puppy/agents/agent_typescript_reviewer.py,sha256=vsnpp98xg6cIoFAEJrRTUM_i4wLEWGm5nJxs6fhHobM,10275
-code_puppy/agents/base_agent.py,sha256=62TcOKownogs8KVcFGuhuxp0Xvj9VrsaqLNwz1t96zs,80331
+code_puppy/agents/base_agent.py,sha256=sT5bNdF2c-SnSdy1pwfZKnv2P-jwB7i67h353UqUN20,77624
 code_puppy/agents/json_agent.py,sha256=lhopDJDoiSGHvD8A6t50hi9ZBoNRKgUywfxd0Po_Dzc,4886
 code_puppy/agents/prompt_reviewer.py,sha256=JJrJ0m5q0Puxl8vFsyhAbY9ftU9n6c6UxEVdNct1E-Q,5558
 code_puppy/command_line/__init__.py,sha256=y7WeRemfYppk8KVbCGeAIiTuiOszIURCDjOMZv_YRmU,45
@@ -56,7 +56,7 @@ code_puppy/command_line/file_path_completion.py,sha256=gw8NpIxa6GOpczUJRyh7VNZwo
 code_puppy/command_line/load_context_completion.py,sha256=a3JvLDeLLSYxVgTjAdqWzS4spjv6ccCrK2LKZgVJ1IM,2202
 code_puppy/command_line/mcp_completion.py,sha256=eKzW2O7gun7HoHekOW0XVXhNS5J2xCtK7aaWyA8bkZk,6952
 code_puppy/command_line/model_picker_completion.py,sha256=nDnlf0qFCG2zAm_mWW2eMYwVC7eROVQrFe92hZqOKa8,6810
-code_puppy/command_line/model_settings_menu.py,sha256=-GDJyvkFeEIs-fYTvkBPxyVjmMo9CJRrwUq_IuiEbqY,31935
+code_puppy/command_line/model_settings_menu.py,sha256=O5nPp_OyShFcXzpSmsCeYsnnVNrSwcTBFY9bzcayvj0,32263
 code_puppy/command_line/motd.py,sha256=OoNxwewsckexSgJ5H5y40IawP-TzqlqY-rqFUdRbIhs,2186
 code_puppy/command_line/pin_command_completion.py,sha256=juSvdqRpk7AdfkPy1DJx5NzfEUU5KYGlChvP0hisM18,11667
 code_puppy/command_line/prompt_toolkit_completion.py,sha256=x4Of32g8oH9ckhx-P6BigV7HUUhhjL8xkvK03uq9HRw,27308
@@ -130,7 +130,7 @@ code_puppy/plugins/claude_code_oauth/__init__.py,sha256=mCcOU-wM7LNCDjr-w-WLPzom
 code_puppy/plugins/claude_code_oauth/config.py,sha256=DjGySCkvjSGZds6DYErLMAi3TItt8iSLGvyJN98nSEM,2013
 code_puppy/plugins/claude_code_oauth/register_callbacks.py,sha256=0NeX1hhkYIlVfPmjZ1xmcf1yueDAJh_FMUmvJlxSO-E,10057
 code_puppy/plugins/claude_code_oauth/test_plugin.py,sha256=yQy4EeZl4bjrcog1d8BjknoDTRK75mRXXvkSQJYSSEM,9286
-code_puppy/plugins/claude_code_oauth/utils.py,sha256=uxNRrvtmyG_zZxcvCyZIU1fib8wV5KeorHgVv0RWS9s,13394
+code_puppy/plugins/claude_code_oauth/utils.py,sha256=wDaOU21zB3y6PWkuMXwE4mFjQuffyDae-vXysPTS-w8,13438
 code_puppy/plugins/customizable_commands/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 code_puppy/plugins/customizable_commands/register_callbacks.py,sha256=zVMfIzr--hVn0IOXxIicbmgj2s-HZUgtrOc0NCDOnDw,5183
 code_puppy/plugins/example_custom_command/README.md,sha256=5c5Zkm7CW6BDSfe3WoLU7GW6t5mjjYAbu9-_pu-b3p4,8244
@@ -159,10 +159,10 @@ code_puppy/tools/browser/browser_scripts.py,sha256=sNb8eLEyzhasy5hV4B9OjM8yIVMLV
 code_puppy/tools/browser/browser_workflows.py,sha256=nitW42vCf0ieTX1gLabozTugNQ8phtoFzZbiAhw1V90,6491
 code_puppy/tools/browser/camoufox_manager.py,sha256=RZjGOEftE5sI_tsercUyXFSZI2wpStXf-q0PdYh2G3I,8680
 code_puppy/tools/browser/vqa_agent.py,sha256=DBn9HKloILqJSTSdNZzH_PYWT0B2h9VwmY6akFQI_uU,2913
-code_puppy-0.0.318.data/data/code_puppy/models.json,sha256=nbMCW9SJxWptVEOeMrn_gfyDp-lsbM7Chczjl944GKA,3077
-code_puppy-0.0.318.data/data/code_puppy/models_dev_api.json,sha256=wHjkj-IM_fx1oHki6-GqtOoCrRMR0ScK0f-Iz0UEcy8,548187
-code_puppy-0.0.318.dist-info/METADATA,sha256=_7lhagjL8fQpd7zBbl8DQ1Z9qmHGRY2gbQ7E9jwNp54,28030
-code_puppy-0.0.318.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-code_puppy-0.0.318.dist-info/entry_points.txt,sha256=Tp4eQC99WY3HOKd3sdvb22vZODRq0XkZVNpXOag_KdI,91
-code_puppy-0.0.318.dist-info/licenses/LICENSE,sha256=31u8x0SPgdOq3izJX41kgFazWsM43zPEF9eskzqbJMY,1075
-code_puppy-0.0.318.dist-info/RECORD,,
+code_puppy-0.0.320.data/data/code_puppy/models.json,sha256=mTpmJH0UJlmX8M2KVPbxMWb99de3IxKXCWO-B23b6xo,3101
+code_puppy-0.0.320.data/data/code_puppy/models_dev_api.json,sha256=wHjkj-IM_fx1oHki6-GqtOoCrRMR0ScK0f-Iz0UEcy8,548187
+code_puppy-0.0.320.dist-info/METADATA,sha256=Ah_-U-dxX8DnEHjQ0G8--g_1lDfyI3WMGcjnJOEfiBk,28030
+code_puppy-0.0.320.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+code_puppy-0.0.320.dist-info/entry_points.txt,sha256=Tp4eQC99WY3HOKd3sdvb22vZODRq0XkZVNpXOag_KdI,91
+code_puppy-0.0.320.dist-info/licenses/LICENSE,sha256=31u8x0SPgdOq3izJX41kgFazWsM43zPEF9eskzqbJMY,1075
+code_puppy-0.0.320.dist-info/RECORD,,