batch-analytics 0.3.20__tar.gz → 0.3.22__tar.gz

{batch_analytics-0.3.20 → batch_analytics-0.3.22}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: batch-analytics
-Version: 0.3.20
+Version: 0.3.22
 Summary: PySpark batch analytics: Extract, Transform, Stage, and analytical modules (linear regression, correlation, PCA, t-test).
 Author: Litewave Analytics Team
 License: MIT
@@ -24,7 +24,8 @@ Requires-Dist: boto3>=1.28; extra == "output"
 Requires-Dist: clickhouse-connect<0.9,>=0.7; python_version < "3.9" and extra == "output"
 Requires-Dist: clickhouse-connect>=0.7; python_version >= "3.9" and extra == "output"
 Provides-Extra: autogluon
-Requires-Dist: autogluon<2.0,>=1.0; extra == "autogluon"
+Requires-Dist: autogluon-tabular[lightgbm]<2.0,>=1.0; extra == "autogluon"
+Requires-Dist: typing-extensions>=4.8.0; extra == "autogluon"
 Requires-Dist: pandas>=1.3.0; extra == "autogluon"
 Requires-Dist: boto3>=1.28; extra == "autogluon"
 Requires-Dist: clickhouse-connect<0.9,>=0.7; python_version < "3.9" and extra == "autogluon"
@@ -36,7 +37,8 @@ Requires-Dist: scipy>=1.5.0; extra == "full"
 Requires-Dist: boto3>=1.28; extra == "full"
 Requires-Dist: clickhouse-connect<0.9,>=0.7; python_version < "3.9" and extra == "full"
 Requires-Dist: clickhouse-connect>=0.7; python_version >= "3.9" and extra == "full"
-Requires-Dist: autogluon<2.0,>=1.0; extra == "full"
+Requires-Dist: autogluon-tabular[lightgbm]<2.0,>=1.0; extra == "full"
+Requires-Dist: typing-extensions>=4.8.0; extra == "full"
 Requires-Dist: pyarrow>=10.0.0; python_version >= "3.8" and extra == "full"
 
 # Batch Analytics

{batch_analytics-0.3.20 → batch_analytics-0.3.22}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "batch-analytics"
-version = "0.3.20"
+version = "0.3.22"
 description = "PySpark batch analytics: Extract, Transform, Stage, and analytical modules (linear regression, correlation, PCA, t-test)."
 readme = "README.md"
 requires-python = ">=3.8"
@@ -31,8 +31,10 @@ output = [
     "clickhouse-connect>=0.7,<0.9; python_version < '3.9'",
     "clickhouse-connect>=0.7; python_version >= '3.9'",
 ]
+# autogluon-tabular[lightgbm]: tabular AutoML with LightGBM only (no PyTorch from full autogluon metapackage)
 autogluon = [
-    "autogluon>=1.0,<2.0",
+    "autogluon-tabular[lightgbm]>=1.0,<2.0",
+    "typing-extensions>=4.8.0",
     "pandas>=1.3.0",
     "boto3>=1.28",
     "clickhouse-connect>=0.7,<0.9; python_version < '3.9'",
@@ -45,7 +47,8 @@ full = [
     "boto3>=1.28",
     "clickhouse-connect>=0.7,<0.9; python_version < '3.9'",
     "clickhouse-connect>=0.7; python_version >= '3.9'",
-    "autogluon>=1.0,<2.0",
+    "autogluon-tabular[lightgbm]>=1.0,<2.0",
+    "typing-extensions>=4.8.0",
     "pyarrow>=10.0.0; python_version >= '3.8'",
 ]
 

{batch_analytics-0.3.20 → batch_analytics-0.3.22}/src/batch_analytics/analytics/gluon_autogluon_train.py

@@ -4,11 +4,13 @@ Train AutoGluon TabularPredictor from ClickHouse staging data; upload artifacts
 Env (injected by analytics_runner Gluon job):
   CLICKHOUSE_*, BATCH_STAGING_TABLE, MODEL_S3_PREFIX, TASK_ID,
   AUTOGLUON_LABEL, AUTOGLUON_FEATURES, AUTOGLUON_PROBLEM_TYPE, AUTOGLUON_TIME_LIMIT,
-  optional AUTOGLUON_MAX_ROWS
+  optional AUTOGLUON_MAX_ROWS,
+  optional AUTOGLUON_HYPERPARAMETERS — JSON object of AutoGluon model hyperparameters; default {"GBM": {}} (LightGBM only, no PyTorch)
 """
 
 from __future__ import annotations
 
+import json
 import logging
 import os
 import shutil
@@ -44,6 +46,22 @@ def _require(name: str) -> str:
     return v
 
 
+def _hyperparameters_from_env() -> dict:
+    """Default: GBM (LightGBM) only — matches autogluon-tabular[lightgbm] and avoids NN/torch models."""
+    raw = os.environ.get("AUTOGLUON_HYPERPARAMETERS", "").strip()
+    if not raw:
+        return {"GBM": {}}
+    try:
+        parsed = json.loads(raw)
+    except json.JSONDecodeError as e:
+        logger.error("AUTOGLUON_HYPERPARAMETERS must be valid JSON: %s", e)
+        sys.exit(2)
+    if not isinstance(parsed, dict):
+        logger.error("AUTOGLUON_HYPERPARAMETERS must be a JSON object, e.g. {\"GBM\": {}}")
+        sys.exit(2)
+    return parsed
+
+
 def main() -> None:
     model_prefix = _require("MODEL_S3_PREFIX")
     label = _require("AUTOGLUON_LABEL")
@@ -79,8 +97,13 @@ def main() -> None:
 
     try:
         from autogluon.tabular import TabularPredictor
-    except ImportError:
-        logger.exception("autogluon is not installed; use pip install 'batch-analytics[autogluon]'")
+    except ImportError as e:
+        if isinstance(e, ModuleNotFoundError) and e.name == "typing_extensions":
+            logger.exception(
+                "Install typing-extensions (required by AutoGluon): pip install 'typing-extensions>=4.8'"
+            )
+        else:
+            logger.exception("autogluon is not installed; use pip install 'batch-analytics[autogluon]'")
         sys.exit(4)
 
     problem_type = os.environ.get("AUTOGLUON_PROBLEM_TYPE", "binary").strip() or "binary"
@@ -92,10 +115,12 @@ def main() -> None:
     os.makedirs(local_dir, exist_ok=True)
 
     train_df = df[feature_list + [label]]
+    hyper = _hyperparameters_from_env()
     logger.info(
-        "Fitting TabularPredictor problem_type=%s time_limit=%ss rows=%s",
+        "Fitting TabularPredictor problem_type=%s time_limit=%ss hyperparameters=%s rows=%s",
         problem_type,
         time_limit,
+        hyper,
         len(train_df),
     )
     predictor = TabularPredictor(
@@ -103,7 +128,7 @@ def main() -> None:
         problem_type=problem_type,
         path=local_dir,
     )
-    predictor.fit(train_df, time_limit=time_limit)
+    predictor.fit(train_df, time_limit=time_limit, hyperparameters=hyper)
 
     logger.info("Uploading model artifacts to %s", model_prefix)
     upload_directory_to_s3(local_dir, model_prefix)

{batch_analytics-0.3.20 → batch_analytics-0.3.22}/src/batch_analytics/transform.py

@@ -200,20 +200,58 @@ def stage_to_clickhouse(
     """
     Write transformed data to ClickHouse staging table.
     Separate job from transform; must complete before analytics can run.
-    Uses native connector if available, else JDBC.
+
+    Preferred path: Spark SQL **catalog** API (``DataFrame.writeTo``), matching
+    ``job_runner.create_spark_session`` registration of ``ClickHouseCatalog``
+    (``BATCH_CLICKHOUSE_CATALOG``, default ``batch_ch``). The clickhouse-spark-runtime
+    0.8.x connector does **not** register the legacy short name ``format("clickhouse")``
+    / ``clickhouse.DefaultSource``.
+
+    Fallback: legacy ``format("clickhouse")`` (older stacks), then JDBC (may fail on
+    ClickHouse 25+ auto-DDL without ORDER BY).
+
     Write mode from BATCH_STAGING_WRITE_MODE (default overwrite = full replace).
     """
     n = df.count()
     mode = _normalize_staging_write_mode(config.transform.staging_write_mode)
+    ch = config.clickhouse
+    tbl = config.transform.staging_table
+    cat = os.environ.get("BATCH_CLICKHOUSE_CATALOG", "batch_ch").strip()
+
+    if cat:
+        try:
+            full_name = f"{cat}.{ch.database}.{tbl}"
+            logger.info(
+                "Staging to ClickHouse via catalog %s (mode=%s)",
+                full_name,
+                mode,
+            )
+            w2 = df.writeTo(full_name)
+            if mode == "overwrite":
+                w2.createOrReplace()
+            else:
+                w2.append()
+            logger.info(
+                "Staged data to ClickHouse %s.%s (%d rows)",
+                ch.database,
+                tbl,
+                n,
+            )
+            return
+        except Exception as e:
+            logger.warning(
+                "ClickHouse catalog write failed (%s), trying legacy format/jdbc",
+                e,
+            )
+
     try:
-        ch = config.clickhouse
         writer = (
             df.write.format("clickhouse")
             .option("host", ch.host)
             .option("protocol", ch.protocol)
             .option("http_port", str(ch.port))
             .option("database", ch.database)
-            .option("table", config.transform.staging_table)
+            .option("table", tbl)
             .option("user", ch.user)
             .mode(mode)
         )
@@ -221,17 +259,17 @@ def stage_to_clickhouse(
             writer = writer.option("password", ch.password)
         writer.save()
     except Exception as e:
-        logger.warning("ClickHouse connector failed (%s), using JDBC", e)
+        logger.warning("ClickHouse legacy format failed (%s), using JDBC", e)
         df.write.jdbc(
             config.clickhouse.jdbc_url,
-            config.transform.staging_table,
+            tbl,
             mode=mode,
             properties=config.clickhouse.jdbc_properties,
         )
     logger.info(
         "Staged data to ClickHouse %s.%s (%d rows)",
-        config.clickhouse.database,
-        config.transform.staging_table,
+        ch.database,
+        tbl,
         n,
     )
 

{batch_analytics-0.3.20 → batch_analytics-0.3.22}/src/batch_analytics.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: batch-analytics
-Version: 0.3.20
+Version: 0.3.22
 Summary: PySpark batch analytics: Extract, Transform, Stage, and analytical modules (linear regression, correlation, PCA, t-test).
 Author: Litewave Analytics Team
 License: MIT
@@ -24,7 +24,8 @@ Requires-Dist: boto3>=1.28; extra == "output"
 Requires-Dist: clickhouse-connect<0.9,>=0.7; python_version < "3.9" and extra == "output"
 Requires-Dist: clickhouse-connect>=0.7; python_version >= "3.9" and extra == "output"
 Provides-Extra: autogluon
-Requires-Dist: autogluon<2.0,>=1.0; extra == "autogluon"
+Requires-Dist: autogluon-tabular[lightgbm]<2.0,>=1.0; extra == "autogluon"
+Requires-Dist: typing-extensions>=4.8.0; extra == "autogluon"
 Requires-Dist: pandas>=1.3.0; extra == "autogluon"
 Requires-Dist: boto3>=1.28; extra == "autogluon"
 Requires-Dist: clickhouse-connect<0.9,>=0.7; python_version < "3.9" and extra == "autogluon"
@@ -36,7 +37,8 @@ Requires-Dist: scipy>=1.5.0; extra == "full"
 Requires-Dist: boto3>=1.28; extra == "full"
 Requires-Dist: clickhouse-connect<0.9,>=0.7; python_version < "3.9" and extra == "full"
 Requires-Dist: clickhouse-connect>=0.7; python_version >= "3.9" and extra == "full"
-Requires-Dist: autogluon<2.0,>=1.0; extra == "full"
+Requires-Dist: autogluon-tabular[lightgbm]<2.0,>=1.0; extra == "full"
+Requires-Dist: typing-extensions>=4.8.0; extra == "full"
 Requires-Dist: pyarrow>=10.0.0; python_version >= "3.8" and extra == "full"
 
 # Batch Analytics

{batch_analytics-0.3.20 → batch_analytics-0.3.22}/src/batch_analytics.egg-info/requires.txt

@@ -1,7 +1,8 @@
 numpy>=1.19.0
 
 [autogluon]
-autogluon<2.0,>=1.0
+autogluon-tabular[lightgbm]<2.0,>=1.0
+typing-extensions>=4.8.0
 pandas>=1.3.0
 boto3>=1.28
 
@@ -29,7 +30,8 @@ pytest>=7.0
 pyspark<3.6,>=3.4
 scipy>=1.5.0
 boto3>=1.28
-autogluon<2.0,>=1.0
+autogluon-tabular[lightgbm]<2.0,>=1.0
+typing-extensions>=4.8.0
 
 [full:python_version < "3.9"]
 clickhouse-connect<0.9,>=0.7