fix: SmartTurnMetricsData was reporting 0 for inference and processing time

CHANGELOG.md

@@ -5,6 +5,13 @@ All notable changes to **Pipecat** will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [Unreleased]
+
+### Fixed
+
+- Fixed an issue where the `SmartTurnMetricsData` was reporting 0ms for
+  inference and processing time when using the `FalSmartTurnAnalyzer`.
+
 ## [0.0.65] - 2025-04-23 "Sant Jordi's release"
 
 https://en.wikipedia.org/wiki/Saint_George%27s_Day_in_Catalonia

src/pipecat/audio/turn/smart_turn/base_smart_turn.py

@@ -149,13 +149,18 @@ class BaseSmartTurn(BaseTurnAnalyzer):
                 # Calculate processing time
                 e2e_processing_time_ms = (end_time - start_time) * 1000
 
+                # Extract metrics from the nested structure
+                metrics = result.get("metrics", {})
+                inference_time = metrics.get("inference_time", 0)
+                total_time = metrics.get("total_time", 0)
+
                 # Prepare the result data
                 result_data = SmartTurnMetricsData(
                     processor="BaseSmartTurn",
                     is_complete=result["prediction"] == 1,
                     probability=result["probability"],
-                    inference_time_ms=result.get("inference_time", 0) * 1000,
+                    inference_time_ms=inference_time * 1000,
-                    server_total_time_ms=result.get("total_time", 0) * 1000,
+                    server_total_time_ms=total_time * 1000,
                     e2e_processing_time_ms=e2e_processing_time_ms,
                 )