Add structured logging for Docker; support LOG_LEVEL env and --log-level

scraibe/localai_client.py

@@ -19,10 +19,13 @@ Environment Variables:
 import os
 import io
 import json
+import logging
 from typing import Dict, List, Any, Optional
 
 import httpx
 
+logger = logging.getLogger("scraibe.localai_client")
+
 
 class LocalAIError(Exception):
     """Raised when the LocalAI API returns an error or unexpected response."""
@@ -67,6 +70,12 @@ class LocalAIClient:
                 "Provide the LocalAI server URL via environment or constructor."
             )
 
+        logger.info(
+            "Initializing LocalAIClient: url=%s model=%s",
+            self.api_url,
+            self.model,
+        )
+
         self._client = httpx.Client(
             base_url=self.api_url,
             timeout=self.timeout,
@@ -130,7 +139,8 @@ class LocalAIClient:
         if verbose:
             print("Starting diarization and transcription via LocalAI.")
 
-        # Defaults: use verbose_json + include_text to get both diarization and transcription.
+        logger.info("diarize_and_transcribe requested for: %s", audio_path)
+
         if response_format is None:
             response_format = "verbose_json"
         if include_text is None:
@@ -158,6 +168,8 @@ class LocalAIClient:
         if min_duration_off is not None:
             data["min_duration_off"] = str(min_duration_off)
 
+        logger.debug("LocalAI request params: %s", data)
+
         # Open file
         if not os.path.exists(audio_path):
             raise LocalAIError(f"Audio file not found: {audio_path}")
@@ -172,6 +184,7 @@ class LocalAIClient:
                 headers["Authorization"] = f"Bearer {self.api_key}"
 
             # POST /v1/audio/diarization
+            logger.info("Sending request to LocalAI: /v1/audio/diarization")
             resp = self._client.post(
                 "/v1/audio/diarization",
                 data=data,
@@ -179,8 +192,11 @@ class LocalAIClient:
                 headers=headers,
             )
 
+        logger.info("LocalAI response status: %d", resp.status_code)
+
         if resp.status_code >= 400:
             body = resp.text
+            logger.error("LocalAI error response: %s", body)
             raise LocalAIError(
                 f"LocalAI request failed with status {resp.status_code}: {body}"
             )
@@ -188,6 +204,7 @@ class LocalAIClient:
         try:
             result = resp.json()
         except json.JSONDecodeError:
+            logger.error("Failed to parse LocalAI response as JSON.")
             raise LocalAIError(
                 "Failed to parse LocalAI response as JSON."
             )
@@ -209,7 +226,7 @@ class LocalAIClient:
         segments = result.get("segments", [])
 
         if not segments:
-            # If no segments, return empty but valid structure
+            logger.warning("LocalAI returned no segments.")
             return {
                 "segments": [],
                 "speakers": [],
@@ -230,6 +247,11 @@ class LocalAIClient:
             out_speakers.append(speaker)
             out_transcripts.append(text)
 
+        logger.info(
+            "Parsed %d segments from LocalAI.",
+            len(out_segments),
+        )
+
         return {
             "segments": out_segments,
             "speakers": out_speakers,