ygg 0.1.50__py3-none-any.whl → 0.1.52__py3-none-any.whl

yggdrasil/databricks/workspaces/io.py

@@ -9,13 +9,18 @@ from typing import TYPE_CHECKING, Optional, IO, AnyStr, Union
 import pyarrow as pa
 import pyarrow.csv as pcsv
 import pyarrow.parquet as pq
-from pyarrow.dataset import FileFormat, ParquetFileFormat, CsvFileFormat
+from Lib.threading import Thread
+from pyarrow.dataset import (
+    FileFormat,
+    ParquetFileFormat, CsvFileFormat,
+)
 
 from .path_kind import DatabricksPathKind
 from ...libs.databrickslib import databricks
-from ...libs.pandaslib import pandas, PandasDataFrame
+from ...libs.pandaslib import PandasDataFrame
 from ...libs.polarslib import polars, PolarsDataFrame
 from ...types.cast.registry import convert
+from ...types.file_format import ExcelFileFormat
 
 if databricks is not None:
     from databricks.sdk.service.workspace import ImportFormat, ExportFormat
@@ -66,7 +71,10 @@ class DatabricksIO(ABC, IO):
         self.close()
 
     def __del__(self):
-        self.close()
+        try:
+            Thread(target=self.close).start()
+        except BaseException:
+            pass
 
     def __next__(self):
         """Iterate over lines in the file."""
@@ -75,8 +83,11 @@ class DatabricksIO(ABC, IO):
             raise StopIteration
         return line
 
+    def __len__(self):
+        return self.content_length or 0
+
     def __iter__(self):
-        return self
+        return self.read_all_bytes().__iter__()
 
     def __hash__(self):
         return self.path.__hash__()
@@ -87,7 +98,6 @@ class DatabricksIO(ABC, IO):
         path: "DatabricksPath",
         mode: str,
         encoding: Optional[str] = None,
-        compression: Optional[str] = "detect",
         position: int = 0,
         buffer: Optional[io.BytesIO] = None,
     ) -> "DatabricksIO":
@@ -97,7 +107,6 @@ class DatabricksIO(ABC, IO):
             path: DatabricksPath to open.
             mode: File mode string.
             encoding: Optional text encoding for text mode.
-            compression: Optional compression mode.
             position: Initial file cursor position.
             buffer: Optional pre-seeded buffer.
 
@@ -170,6 +179,10 @@ class DatabricksIO(ABC, IO):
     def content_length(self) -> int:
         return self.path.content_length
 
+    @content_length.setter
+    def content_length(self, value: int):
+        self.path.content_length = value
+
     def size(self):
         """Return the size of the file in bytes.
 
@@ -178,10 +191,6 @@ class DatabricksIO(ABC, IO):
         """
         return self.content_length
 
-    @content_length.setter
-    def content_length(self, value: int):
-        self.path.content_length = value
-
     @property
     def buffer(self):
         """Return the in-memory buffer, creating it if necessary.
@@ -204,9 +213,7 @@ class DatabricksIO(ABC, IO):
         Returns:
             None.
         """
-        if self._buffer is not None:
-            self._buffer.close()
-            self._buffer = None
+        self._buffer = None
 
     def clone_instance(self, **kwargs):
         """Clone this IO instance with optional overrides.
@@ -251,15 +258,23 @@ class DatabricksIO(ABC, IO):
         self.path = path
         return self
 
-    def close(self):
+    def close(self, flush: bool = True):
         """Flush pending writes and close the buffer.
 
+        Args:
+            flush: Checks flush data to commit to remote location
+
         Returns:
             None.
         """
-        self.flush()
+        if flush:
+            self.flush()
         self.clear_buffer()
 
+    @property
+    def closed(self):
+        return False
+
     def fileno(self):
         """Return a pseudo file descriptor based on object hash.
 
@@ -508,14 +523,14 @@ class DatabricksIO(ABC, IO):
         if size is None:
             size = self.position
 
-        if self._buffer is not None:
-            self._buffer.truncate(size)
-        else:
-            data = b"\x00" * size
-            self.write_all_bytes(data=data)
+        if self._buffer is None:
+            return self.write_all_bytes(data=b"\x00" * size)
+
+        self._buffer.truncate(size)
 
         self.content_length = size
         self._write_flag = True
+
         return size
 
     def flush(self):
@@ -588,7 +603,13 @@ class DatabricksIO(ABC, IO):
         Returns:
             None.
         """
-        if not isinstance(dest, DatabricksIO):
+        data = self.read_all_bytes(use_cache=False)
+
+        if isinstance(dest, DatabricksIO):
+            dest.write_all_bytes(data=data)
+        elif hasattr(dest, "write"):
+            dest.write(data)
+        else:
             from .path import DatabricksPath
 
             dest_path = DatabricksPath.parse(dest, workspace=self.workspace)
@@ -596,8 +617,6 @@ class DatabricksIO(ABC, IO):
             with dest_path.open(mode="wb") as d:
                 return self.copy_to(dest=d)
 
-        dest.write_all_bytes(data=self.read_all_bytes(use_cache=False))
-
     # ---- format helpers ----
 
     def _reset_for_write(self):
@@ -616,7 +635,6 @@ class DatabricksIO(ABC, IO):
         table: Union[pa.Table, pa.RecordBatch, PolarsDataFrame, PandasDataFrame],
         file_format: Optional[FileFormat] = None,
         batch_size: Optional[int] = None,
-        **kwargs
     ):
         """Write a table-like object to the path using an inferred format.
 
@@ -624,19 +642,18 @@ class DatabricksIO(ABC, IO):
             table: Table-like object to write.
             file_format: Optional file format override.
             batch_size: Optional batch size for writes.
-            **kwargs: Format-specific options.
 
         Returns:
             The result of the specific write implementation.
         """
         if isinstance(table, pa.Table):
-            return self.write_arrow_table(table, file_format=file_format, batch_size=batch_size, **kwargs)
+            return self.write_arrow_table(table, file_format=file_format, batch_size=batch_size)
         elif isinstance(table, pa.RecordBatch):
-            return self.write_arrow_batch(table, file_format=file_format, batch_size=batch_size, **kwargs)
+            return self.write_arrow_batch(table, file_format=file_format, batch_size=batch_size)
         elif isinstance(table, PolarsDataFrame):
-            return self.write_polars(table, file_format=file_format, batch_size=batch_size, **kwargs)
+            return self.write_polars(table, file_format=file_format, batch_size=batch_size)
         elif isinstance(table, PandasDataFrame):
-            return self.write_pandas(table, file_format=file_format, batch_size=batch_size, **kwargs)
+            return self.write_pandas(table, file_format=file_format, batch_size=batch_size)
 
         return self.write_arrow(
             table=table,
@@ -666,10 +683,10 @@ class DatabricksIO(ABC, IO):
         self.seek(0)
 
         if isinstance(file_format, ParquetFileFormat):
-            pq.read_table(self, **kwargs)
+            return pq.read_table(self, **kwargs)
 
         elif isinstance(file_format, CsvFileFormat):
-            pcsv.read_csv(self, parse_options=file_format.parse_options)
+            return pcsv.read_csv(self, parse_options=file_format.parse_options)
 
         else:
             ValueError(f"Unsupported file format for Arrow table: {file_format}")
@@ -679,7 +696,6 @@ class DatabricksIO(ABC, IO):
         table: Union[pa.Table, pa.RecordBatch],
         file_format: Optional[FileFormat] = None,
         batch_size: Optional[int] = None,
-        **kwargs
     ):
         """Write an Arrow table or record batch to the path.
 
@@ -687,7 +703,6 @@ class DatabricksIO(ABC, IO):
             table: Arrow table or batch to write.
             file_format: Optional file format override.
             batch_size: Optional batch size for writes.
-            **kwargs: Format-specific options.
 
         Returns:
             None.
@@ -699,7 +714,6 @@ class DatabricksIO(ABC, IO):
             table=table,
             file_format=file_format,
             batch_size=batch_size,
-            **kwargs
         )
 
     def write_arrow_table(
@@ -707,7 +721,6 @@ class DatabricksIO(ABC, IO):
         table: pa.Table,
         file_format: Optional[FileFormat] = None,
         batch_size: Optional[int] = None,
-        **kwargs
     ):
         """Write an Arrow table using the selected file format.
 
@@ -715,7 +728,6 @@ class DatabricksIO(ABC, IO):
             table: Arrow table to write.
             file_format: Optional file format override.
             batch_size: Optional batch size for writes.
-            **kwargs: Format-specific options.
 
         Returns:
             None.
@@ -724,13 +736,20 @@ class DatabricksIO(ABC, IO):
         buffer = io.BytesIO()
 
         if isinstance(file_format, ParquetFileFormat):
-            pq.write_table(table, buffer, write_batch_size=batch_size, **kwargs)
+            pq.write_table(
+                table, buffer,
+                write_batch_size=batch_size
+            )
 
         elif isinstance(file_format, CsvFileFormat):
-            pcsv.write_csv(table, buffer, **kwargs)
+            pcsv.write_csv(table, buffer)
 
         else:
-            raise ValueError(f"Unsupported file format for Arrow table: {file_format}")
+            return self.write_polars(
+                df=polars.from_arrow(table),
+                file_format=file_format,
+                batch_size=batch_size
+            )
 
         self.write_all_bytes(data=buffer.getvalue())
 
@@ -739,7 +758,6 @@ class DatabricksIO(ABC, IO):
         batch: pa.RecordBatch,
         file_format: Optional[FileFormat] = None,
         batch_size: Optional[int] = None,
-        **kwargs
     ):
         """Write a single Arrow record batch.
 
@@ -747,26 +765,23 @@ class DatabricksIO(ABC, IO):
             batch: RecordBatch to write.
             file_format: Optional file format override.
             batch_size: Optional batch size for writes.
-            **kwargs: Format-specific options.
 
         Returns:
             None.
         """
         table = pa.Table.from_batches([batch])
-        self.write_arrow_table(table, file_format=file_format, batch_size=batch_size, **kwargs)
+        self.write_arrow_table(table, file_format=file_format, batch_size=batch_size)
 
     def read_arrow_batches(
         self,
         file_format: Optional[FileFormat] = None,
         batch_size: Optional[int] = None,
-        **kwargs
     ):
         """Yield Arrow record batches from the file.
 
         Args:
             file_format: Optional file format override.
             batch_size: Optional batch size for reads.
-            **kwargs: Format-specific options.
 
         Returns:
             An iterator over Arrow RecordBatch objects.
@@ -776,7 +791,6 @@ class DatabricksIO(ABC, IO):
             .read_arrow_table(
                 file_format=file_format,
                 batch_size=batch_size,
-                **kwargs
             )
             .to_batches(max_chunksize=batch_size)
         )
@@ -787,36 +801,26 @@ class DatabricksIO(ABC, IO):
         self,
         file_format: Optional[FileFormat] = None,
         batch_size: Optional[int] = None,
-        **kwargs
     ):
         """Read the file into a pandas DataFrame.
 
         Args:
             file_format: Optional file format override.
             batch_size: Optional batch size for reads.
-            **kwargs: Format-specific options.
 
         Returns:
             A pandas DataFrame with the file contents.
         """
-        file_format = self.path.file_format if file_format is None else file_format
-        self.seek(0)
-
-        if isinstance(file_format, ParquetFileFormat):
-            pandas.read_parquet(self, **kwargs)
-
-        elif isinstance(file_format, CsvFileFormat):
-            pandas.read_csv(self, **kwargs)
-
-        else:
-            raise ValueError(f"Unsupported file format for Pandas DataFrame: {file_format}")
+        return self.read_arrow_table(
+            file_format=file_format,
+            batch_size=batch_size
+        ).to_pandas()
 
     def write_pandas(
         self,
         df: PandasDataFrame,
         file_format: Optional[FileFormat] = None,
         batch_size: Optional[int] = None,
-        **kwargs
     ):
         """Write a pandas DataFrame to the file.
 
@@ -824,7 +828,6 @@ class DatabricksIO(ABC, IO):
             df: pandas DataFrame to write.
             file_format: Optional file format override.
             batch_size: Optional batch size for writes.
-            **kwargs: Format-specific options.
 
         Returns:
             None.
@@ -833,13 +836,17 @@ class DatabricksIO(ABC, IO):
         buffer = io.BytesIO()
 
         if isinstance(file_format, ParquetFileFormat):
-            df.to_parquet(buffer, **kwargs)
+            df.to_parquet(buffer)
 
         elif isinstance(file_format, CsvFileFormat):
-            df.to_csv(buffer, **kwargs)
+            df.to_csv(buffer)
 
         else:
-            raise ValueError(f"Unsupported file format for Pandas DataFrame: {file_format}")
+            return self.write_polars(
+                df=polars.from_pandas(df),
+                file_format=file_format,
+                batch_size=batch_size,
+            )
 
         self.write_all_bytes(data=buffer.getvalue())
 
@@ -849,14 +856,12 @@ class DatabricksIO(ABC, IO):
         self,
         file_format: Optional[FileFormat] = None,
         batch_size: Optional[int] = None,
-        **kwargs
     ):
         """Read the file into a polars DataFrame.
 
         Args:
             file_format: Optional file format override.
             batch_size: Optional batch size for reads.
-            **kwargs: Format-specific options.
 
         Returns:
             A polars DataFrame with the file contents.
@@ -865,10 +870,13 @@ class DatabricksIO(ABC, IO):
         self.seek(0)
 
         if isinstance(file_format, ParquetFileFormat):
-            polars.read_parquet(self, **kwargs)
+            return polars.read_parquet(self)
 
         elif isinstance(file_format, CsvFileFormat):
-            polars.read_csv(self, **kwargs)
+            return polars.read_csv(self)
+
+        elif isinstance(file_format, ExcelFileFormat):
+            return polars.read_excel(self)
 
         else:
             raise ValueError(f"Unsupported file format for Polars DataFrame: {file_format}")
@@ -878,7 +886,6 @@ class DatabricksIO(ABC, IO):
         df: PolarsDataFrame,
         file_format: Optional[FileFormat] = None,
         batch_size: Optional[int] = None,
-        **kwargs
     ):
         """Write a polars DataFrame to the file.
 
@@ -886,7 +893,6 @@ class DatabricksIO(ABC, IO):
             df: polars DataFrame to write.
             file_format: Optional file format override.
             batch_size: Optional batch size for writes.
-            **kwargs: Format-specific options.
 
         Returns:
             None.
@@ -895,10 +901,13 @@ class DatabricksIO(ABC, IO):
         buffer = io.BytesIO()
 
         if isinstance(file_format, ParquetFileFormat):
-            df.write_parquet(buffer, **kwargs)
+            df.write_parquet(buffer)
 
         elif isinstance(file_format, CsvFileFormat):
-            df.write_csv(buffer, **kwargs)
+            df.write_csv(buffer)
+
+        elif isinstance(file_format, ExcelFileFormat):
+            df.write_excel(buffer)
 
         else:
             raise ValueError(f"Unsupported file format for Polars DataFrame: {file_format}")