fix: clean up verbose integration logging

Author: marcellodebernardi

plexe/config.py

@@ -562,6 +562,9 @@ def setup_logging(config: Config) -> logging.Logger:
     # Get package root logger
     package_logger = logging.getLogger("plexe")
     package_logger.setLevel(getattr(logging, config.log_level.upper()))
+    # Avoid duplicate output when external handlers (e.g., pytest live logging)
+    # are attached to the root logger.
+    package_logger.propagate = False
 
     # Clear existing handlers to avoid duplicates
     package_logger.handlers = []

plexe/execution/dataproc/session.py

@@ -109,8 +109,8 @@ def _create_local_spark(config) -> SparkSession:
             logger.info("Using pre-bundled Spark JARs from /opt/spark-jars/")
             builder = builder.config("spark.jars", spark_jars_env)
         else:
-            # Fallback: Download JARs at runtime via Maven (local development)
-            logger.info("Downloading Spark JARs from Maven Central (first run may take ~40s)")
+            # Fallback: Resolve JARs via Maven (download occurs only on cache miss)
+            logger.info("Resolving Spark JARs via Maven Central (download only on first run/cache miss)")
             builder = builder.config(
                 "spark.jars.packages",
                 "org.apache.hadoop:hadoop-aws:3.3.6,com.amazonaws:aws-java-sdk-bundle:1.12.367",

scripts/tests/run_integration_staged.sh

@@ -28,14 +28,24 @@ else
   WORKERS="auto"
 fi
 PYTEST_PARALLEL_ARGS=(-n "$WORKERS")
+PYTEST_LOG_DISABLE_ARGS=(
+  --log-disable=LiteLLM
+  --log-disable=litellm
+  --log-disable=httpx
+  --log-disable=httpcore
+  --log-disable=urllib3
+  --log-disable=py4j
+  --log-disable=py4j.clientserver
+  --log-disable=py4j.java_gateway
+)
 
 run_stage() {
   local stage_name="$1"
   local marker="$2"
   local cmd=(poetry run pytest tests/integration -m "$marker" "${PYTEST_PARALLEL_ARGS[@]}" --maxfail=1)
 
   if [[ "${PLEXE_IT_VERBOSE:-0}" == "1" ]]; then
-    cmd+=(-s -vv -o log_cli=true -o log_cli_level=INFO --capture=tee-sys)
+    cmd+=(-s -vv -o log_cli=true -o log_cli_level=INFO --capture=tee-sys "${PYTEST_LOG_DISABLE_ARGS[@]}")
   fi
 
   PLEXE_IT_STAGE="$stage_name" "${cmd[@]}"

tests/unit/test_config.py

@@ -1,9 +1,11 @@
 """Unit tests for config helpers."""
 
+import logging
+
 import pytest
 import yaml
 
-from plexe.config import Config, RoutingConfig, RoutingProviderConfig, get_routing_for_model
+from plexe.config import Config, RoutingConfig, RoutingProviderConfig, get_routing_for_model, setup_logging
 
 
 def test_get_routing_for_model_mapping_and_default():
@@ -72,3 +74,12 @@ def test_get_temperature_resolves_override_and_default():
 
     assert config.get_temperature("hypothesiser") == pytest.approx(0.7)
     assert config.get_temperature("layout_detector") == pytest.approx(0.2)
+
+
+def test_setup_logging_disables_propagation():
+    """Plexe logger should not propagate to root to avoid duplicate log lines."""
+    logger = setup_logging(Config(log_level="INFO"))
+
+    assert logger.name == "plexe"
+    assert logger.propagate is False
+    assert any(isinstance(h, logging.StreamHandler) for h in logger.handlers)