databricks-labs-lakebridge 0.10.3__py3-none-any.whl → 0.10.5__py3-none-any.whl

databricks/labs/lakebridge/__about__.py

@@ -1,2 +1,2 @@
 # DO NOT MODIFY THIS FILE
-__version__ = "0.10.3"
+__version__ = "0.10.5"

databricks/labs/lakebridge/cli.py

@@ -132,7 +132,7 @@ def transpile(
     logger.debug(f"Final configuration for transpilation: {config!r}")
 
     assert config.source_dialect is not None, "Source dialect has been validated by this point."
-    with_user_agent_extra("transpiler_source_tech", config.source_dialect)
+    with_user_agent_extra("transpiler_source_tech", make_alphanum_or_semver(config.source_dialect))
     plugin_name = engine.transpiler_name
     plugin_name = re.sub(r"\s+", "_", plugin_name)
     with_user_agent_extra("transpiler_plugin_name", plugin_name)
@@ -639,27 +639,32 @@ def configure_reconcile(w: WorkspaceClient):
     with_user_agent_extra("cmd", "configure-reconcile")
     user = w.current_user
     logger.debug(f"User: {user}")
-    dbsql_id = _create_warehouse(w)
-    w.config.warehouse_id = dbsql_id
+    if not w.config.warehouse_id:
+        dbsql_id = _create_warehouse(w)
+        w.config.warehouse_id = dbsql_id
+    logger.debug(f"Warehouse ID used for configuring reconcile: {w.config.warehouse_id}.")
     installer = _installer(w)
     installer.run(module="reconcile")
-    _remove_warehouse(w, dbsql_id)
 
 
 @lakebridge.command()
-def analyze(w: WorkspaceClient, source_directory: str, report_file: str):
+def analyze(w: WorkspaceClient, source_directory: str, report_file: str, source_tech: str | None = None):
     """Run the Analyzer"""
     with_user_agent_extra("cmd", "analyze")
     ctx = ApplicationContext(w)
     prompts = ctx.prompts
     output_file = report_file
     input_folder = source_directory
-    source_tech = prompts.choice("Select the source technology", Analyzer.supported_source_technologies())
+    if source_tech is None:
+        source_tech = prompts.choice("Select the source technology", Analyzer.supported_source_technologies())
     with_user_agent_extra("analyzer_source_tech", make_alphanum_or_semver(source_tech))
     user = ctx.current_user
     logger.debug(f"User: {user}")
     is_debug = logger.getEffectiveLevel() == logging.DEBUG
     Analyzer.analyze(Path(input_folder), Path(output_file), source_tech, is_debug=is_debug)
+    logger.info(
+        f"Successfully Analyzed files in ${source_directory} for ${source_tech} and saved report to {report_file}"
+    )
 
 
 if __name__ == "__main__":

databricks/labs/lakebridge/install.py

@@ -182,22 +182,6 @@ class WheelInstaller(TranspilerInstaller):
             logger.error(f"Error while fetching PyPI metadata: {product_name}", exc_info=e)
             return None
 
-    @classmethod
-    def download_artifact_from_pypi(cls, product_name: str, version: str, target: Path, extension="whl") -> int:
-        suffix = "-py3-none-any.whl" if extension == "whl" else ".tar.gz" if extension == "tar" else f".{extension}"
-        filename = f"{product_name.replace('-', '_')}-{version}{suffix}"
-        url = f"https://pypi.debian.net/{product_name}/{filename}"
-        try:
-            path, _ = request.urlretrieve(url)
-            logger.info(f"Successfully downloaded {path}")
-            if not target.exists():
-                logger.info(f"Moving {path} to {target!s}")
-                move(path, target)
-            return 0
-        except URLError as e:
-            logger.error("While downloading from pypi", exc_info=e)
-            return -1
-
     def __init__(self, product_name: str, pypi_name: str, artifact: Path | None = None):
         self._product_name = product_name
         self._pypi_name = pypi_name

databricks/labs/lakebridge/transpiler/execute.py

@@ -49,8 +49,9 @@ class TranspilingContext:
 
 
 async def _process_one_file(context: TranspilingContext) -> tuple[int, list[TranspileError]]:
+    input_path = context.input_path
 
-    logger.debug(f"Started processing file: {context.input_path!s}")
+    logger.debug(f"Started processing file: {input_path}")
 
     if not context.config.source_dialect:
         error = TranspileError(
@@ -62,7 +63,11 @@ async def _process_one_file(context: TranspilingContext) -> tuple[int, list[Tran
         )
         return 0, [error]
 
-    source_code = read_text(context.input_path)
+    # Check if it looks like XML, where we need to sniff the encoding instead of relying on a BOM or defaulting to the
+    # local platform encoding.
+    sniff_xml_encoding = input_path.suffix.lower() == ".xml"
+
+    source_code = read_text(input_path, detect_xml=sniff_xml_encoding)
     context = dataclasses.replace(context, source_code=source_code)
 
     transpile_result = await _transpile(

{databricks_labs_lakebridge-0.10.3.dist-info → databricks_labs_lakebridge-0.10.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: databricks-labs-lakebridge
-Version: 0.10.3
+Version: 0.10.5
 Summary: Fast and predictable migrations to Databricks Lakehouse Platform. This tool is designed to help you migrate your data and workloads to the Databricks Lakehouse Platform in a fast, predictable, and reliable way. It provides a set of tools and utilities to help you reconcile your data and workloads, assess your current state, and plan your migration.
 Project-URL: Documentation, https://databrickslabs.github.io/lakebridge
 Project-URL: Issues, https://github.com/databrickslabs/lakebridge/issues
@@ -26,7 +26,7 @@ Classifier: Topic :: Utilities
 Requires-Python: >=3.10
 Requires-Dist: cryptography<45.1.0,>=44.0.2
 Requires-Dist: databricks-bb-analyzer~=0.1.9
-Requires-Dist: databricks-labs-blueprint[yaml]<0.12.0,>=0.11.2
+Requires-Dist: databricks-labs-blueprint[yaml]<0.12.0,>=0.11.3
 Requires-Dist: databricks-labs-lsql==0.16.0
 Requires-Dist: databricks-sdk~=0.51.0
 Requires-Dist: duckdb~=1.2.2

{databricks_labs_lakebridge-0.10.3.dist-info → databricks_labs_lakebridge-0.10.5.dist-info}/RECORD

@@ -8,12 +8,12 @@ docs/lakebridge/src/theme/Footer/index.tsx,sha256=Jj8zY5WDiTLXwF_mAgld8Dh1A3MY1H
 docs/lakebridge/src/theme/Layout/index.tsx,sha256=IkdLr13jKmLxT0jWQqrwqrjVXc8Rwd_kWNpTd1t2sc0,592
 databricks/__init__.py,sha256=YqH8Hy8lHJxd0hLMZF6kWirUDdPiX90LRDX6S6yTMn0,261
 databricks/labs/__init__.py,sha256=YqH8Hy8lHJxd0hLMZF6kWirUDdPiX90LRDX6S6yTMn0,261
-databricks/labs/lakebridge/__about__.py,sha256=LBCN0OI_6vUqxgIo75HVdQc1TP5LOmy5HIs3OKrDIpk,49
+databricks/labs/lakebridge/__about__.py,sha256=I-rxjdEQtJwfk11pg66LiskWUXIpHZGxnU39vNXcHwM,49
 databricks/labs/lakebridge/__init__.py,sha256=nUNECqNvyfpT0aeWwlqG0ADT8U8ScCLb8WWpLydppcA,464
 databricks/labs/lakebridge/base_install.py,sha256=8NxXsNpgqXnuADKXVFh5oQL3osdvygRMY1amJwKfU08,490
-databricks/labs/lakebridge/cli.py,sha256=6exPUJs7c2qVo-X9VXFg5VM3XqOCdlk0_5OXfPw6nbY,31578
+databricks/labs/lakebridge/cli.py,sha256=VJHfy665ln7w-8e4xX18TXemT-XYVVQrXQKA0FkB-VI,31903
 databricks/labs/lakebridge/config.py,sha256=IjxvphM9fRQHQ2FAxwZ23deJGgSemJ3rMV0sp1Ob6e8,5833
-databricks/labs/lakebridge/install.py,sha256=EmtzbC-pOeiK7lqn4wxSRoeODlkqB_lQBJ9Mj4E0kjE,40536
+databricks/labs/lakebridge/install.py,sha256=x8YQwX-EwSKGYKHeotzUwADfzBrvFeplDHbjifLq9mA,39757
 databricks/labs/lakebridge/jvmproxy.py,sha256=F9pXpemzdaJXwpshHxVM9PYU_eNn4zTCUFQ5vc9WIhA,1573
 databricks/labs/lakebridge/lineage.py,sha256=Q2oky4RkODRHWMwIQIwbYXSdZTmRkMWwEh6RssBiQxY,1843
 databricks/labs/lakebridge/uninstall.py,sha256=hf36YgeW9XO2cRvvn6AXUZdihQ1ZMHnR38OVEF5sfRw,759
@@ -147,7 +147,7 @@ databricks/labs/lakebridge/resources/reconcile/queries/installation/details.sql,
 databricks/labs/lakebridge/resources/reconcile/queries/installation/main.sql,sha256=s_A0YyGSX_pCWnQsQnY65VYFcbNvq2qKJvYxU6zam6E,794
 databricks/labs/lakebridge/resources/reconcile/queries/installation/metrics.sql,sha256=FdvjQp7gCwsbcu4UrOuJN-bBLJFpvUIyxH6PQvg04Wo,1006
 databricks/labs/lakebridge/transpiler/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-databricks/labs/lakebridge/transpiler/execute.py,sha256=7DpeIixATOPryyt4TD93-sdwE1C_fIwuo6bKwClaF_s,17007
+databricks/labs/lakebridge/transpiler/execute.py,sha256=AgsOGQtmPfCoWYP9U1cYezY14UjGx6TPR_PU1NAY6sE,17268
 databricks/labs/lakebridge/transpiler/transpile_engine.py,sha256=5zC8fkpBBlt9RjE_BeA_Sd6vaRxA3mBdhTqoRGFTc_Y,1616
 databricks/labs/lakebridge/transpiler/transpile_status.py,sha256=MO-Ju-ki3FCY15WxgwfPV9EC7Ma9q8aIfSTgHAmnkGU,1715
 databricks/labs/lakebridge/transpiler/lsp/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -165,9 +165,9 @@ databricks/labs/lakebridge/transpiler/sqlglot/parsers/presto.py,sha256=bY6Ku8ZPW
 databricks/labs/lakebridge/transpiler/sqlglot/parsers/snowflake.py,sha256=dZ7BdOlBZlkbiN9G9bu4l2c456265Gx9WoWUPRa7Ffg,23203
 databricks/labs/lakebridge/upgrades/v0.4.0_add_main_table_operation_name_column.py,sha256=wMTbj1q5td4fa5DCk0tWFJ-OmhhzsExRLYUe4PKmk0s,3527
 databricks/labs/lakebridge/upgrades/v0.6.0_alter_metrics_datatype.py,sha256=hnTHRtqzwPSF5Judzh6ss-uB5h3IFtm2ylWduwRNq5Y,2424
-databricks_labs_lakebridge-0.10.3.dist-info/METADATA,sha256=e7yr--8po1oLKE8_BRRTFbv_y2fJ0Fw5F95wT2bnn8U,3078
-databricks_labs_lakebridge-0.10.3.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-databricks_labs_lakebridge-0.10.3.dist-info/entry_points.txt,sha256=Idr1CT73b8wShdr287yu1hheGbDbhBvucVUlZcbpiPo,75
-databricks_labs_lakebridge-0.10.3.dist-info/licenses/LICENSE,sha256=1hG0Cvw6mp9nL9qRoHFcCUk9fYqhcnj2vgJ75rt3BxA,3862
-databricks_labs_lakebridge-0.10.3.dist-info/licenses/NOTICE,sha256=wtxMsNvTkw1hAEkkWHz8A8JrYySAUSt1tOTcqddkWEg,1797
-databricks_labs_lakebridge-0.10.3.dist-info/RECORD,,
+databricks_labs_lakebridge-0.10.5.dist-info/METADATA,sha256=fY2pyveC_2Z9FjVEFHYNX23xEc8Fl1dbrhfiFcb886o,3078
+databricks_labs_lakebridge-0.10.5.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+databricks_labs_lakebridge-0.10.5.dist-info/entry_points.txt,sha256=Idr1CT73b8wShdr287yu1hheGbDbhBvucVUlZcbpiPo,75
+databricks_labs_lakebridge-0.10.5.dist-info/licenses/LICENSE,sha256=1hG0Cvw6mp9nL9qRoHFcCUk9fYqhcnj2vgJ75rt3BxA,3862
+databricks_labs_lakebridge-0.10.5.dist-info/licenses/NOTICE,sha256=wtxMsNvTkw1hAEkkWHz8A8JrYySAUSt1tOTcqddkWEg,1797
+databricks_labs_lakebridge-0.10.5.dist-info/RECORD,,