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

scraibe/cli.py

@@ -9,9 +9,10 @@ This version is adapted for LocalAI-based transcription and diarization.
 
 import os
 import json
+import logging
 from argparse import ArgumentParser, ArgumentDefaultsHelpFormatter
 from .autotranscript import Scraibe
-from .misc import set_threads
+from .misc import set_threads, setup_logging
 
 
 def cli():
@@ -20,6 +21,11 @@ def cli():
     and diarize audio files via a LocalAI server.
     """
 
+    # Initialize logging (can be overridden via --log-level)
+    setup_logging(level=os.getenv("LOG_LEVEL", "INFO"))
+
+    logger = logging.getLogger("scraibe.cli")
+
     def str2bool(string):
         str2val = {"True": True, "False": False}
         if string in str2val:
@@ -181,18 +187,34 @@ def cli():
         help="Number of speakers in the audio.",
     )
 
+    parser.add_argument(
+        "--log-level",
+        type=str,
+        default=None,
+        choices=["DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"],
+        help="Override LOG_LEVEL env var for logging verbosity.",
+    )
+
     args = parser.parse_args()
 
+    # Apply log-level override if provided
+    log_level = args.log_level or os.getenv("LOG_LEVEL", "INFO")
+    setup_logging(level=log_level)
+    logger.info("CLI starting with log_level=%s", log_level)
+
     arg_dict = vars(args)
 
     # configure output
     out_folder = arg_dict.pop("output_directory")
     os.makedirs(out_folder, exist_ok=True)
+    logger.info("Output directory: %s", out_folder)
 
     out_format = arg_dict.pop("output_format")
-
     task = arg_dict.pop("task")
 
+    logger.info("Task: %s", task)
+    logger.info("Output format: %s", out_format)
+
     set_threads(arg_dict.pop("num_threads"))
 
     # Build kwargs for Scraibe (LocalAI-backed)
@@ -208,13 +230,18 @@ def cli():
         "verbose": arg_dict.pop("verbose_output"),
     }
 
+    logger.info("LocalAI API URL: %s", class_kwargs["api_url"] or os.getenv("LOCALAI_API_URL", "<not set>"))
+    logger.info("LocalAI Model: %s", class_kwargs["model"] or os.getenv("LOCALAI_MODEL", "<not set>"))
+
     model = Scraibe(**class_kwargs)
 
     if arg_dict["audio_files"]:
         audio_files = arg_dict.pop("audio_files")
+        logger.info("Audio files: %s", audio_files)
 
         if task == "transcribe":
             for audio in audio_files:
+                logger.info("Starting 'transcribe' for: %s", audio)
                 out = model.transcribe(
                     audio,
                     language=arg_dict.pop("language"),
@@ -223,12 +250,14 @@ def cli():
                 )
                 basename = audio.split("/")[-1].split(".")[0]
                 path = os.path.join(out_folder, f"{basename}.{out_format}")
-                print(f"Saving {basename}.{out_format} to {out_folder}")
+                logger.info("Saving transcript to: %s", path)
                 with open(path, "w", encoding="utf-8") as f:
                     f.write(out)
+                logger.info("Transcript saved: %s", path)
 
         elif task == "transcript_and_summarize":
             for audio in audio_files:
+                logger.info("Starting 'transcript_and_summarize' for: %s", audio)
                 result = model.transcript_and_summarize(
                     audio,
                     summarizer_api_url=arg_dict.pop("summarizer_api_url"),
@@ -246,7 +275,7 @@ def cli():
 
                 # Always use .md for transcript_and_summarize
                 md_path = os.path.join(out_folder, f"{basename}.md")
-                print(f"Saving {basename}.md (transcript + summary) to {out_folder}")
+                logger.info("Saving transcript + summary to: %s", md_path)
 
                 with open(md_path, "w", encoding="utf-8") as f:
                     f.write("# Transcript\n\n")
@@ -254,5 +283,7 @@ def cli():
                     f.write("\n\n# Summary\n\n")
                     f.write(summary_text)
 
+                logger.info("Transcript + summary saved: %s", md_path)
+
 if __name__ == "__main__":
     cli()