Repository navigation
Expand file tree
/
Copy pathdataframe.py
More file actions
96 lines (79 loc) · 2.95 KB
/
Copy pathdataframe.py
File metadata and controls
96 lines (79 loc) · 2.95 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
"""FeastDataFrame: A lightweight container for DataFrame-like objects in Feast."""
from enum import Enum
from typing import Any, Dict, Optional
import pandas as pd
import pyarrow as pa
class DataFrameEngine(str, Enum):
"""Supported DataFrame engines."""
PANDAS = "pandas"
SPARK = "spark"
DASK = "dask"
RAY = "ray"
ARROW = "arrow"
POLARS = "polars"
UNKNOWN = "unknown"
class FeastDataFrame:
"""
A lightweight container for DataFrame-like objects in Feast.
This class wraps any DataFrame implementation and provides metadata
about the engine type for proper routing in Feast's processing pipeline.
"""
def __init__(
self,
data: Any,
engine: Optional[DataFrameEngine] = None,
metadata: Optional[Dict[str, Any]] = None,
):
"""
Initialize a FeastDataFrame.
Args:
data: The wrapped DataFrame object (pandas, Spark, Dask, etc.)
engine: Explicitly specify the engine type (auto-detected if None)
metadata: Additional metadata (schema hints, etc.)
"""
self.data = data
self.metadata = metadata or {}
# Detect the actual engine from the data
detected_engine = self._detect_engine()
if engine is not None:
# Validate that the provided engine matches the detected engine
if engine != detected_engine:
raise ValueError(
f"Provided engine '{engine.value}' does not match detected engine '{detected_engine.value}' "
f"for data type {type(data).__name__}"
)
self._engine = engine
else:
self._engine = detected_engine
def _detect_engine(self) -> DataFrameEngine:
"""Auto-detect the DataFrame engine based on type."""
if isinstance(self.data, pd.DataFrame):
return DataFrameEngine.PANDAS
elif isinstance(self.data, pa.Table):
return DataFrameEngine.ARROW
# For optional dependencies, check module name to avoid import errors
module = type(self.data).__module__
if "pyspark" in module:
return DataFrameEngine.SPARK
elif "dask" in module:
return DataFrameEngine.DASK
elif "ray" in module:
return DataFrameEngine.RAY
elif "polars" in module:
return DataFrameEngine.POLARS
else:
return DataFrameEngine.UNKNOWN
@property
def engine(self) -> DataFrameEngine:
"""Get the detected or specified engine type."""
return self._engine
def __repr__(self):
return f"FeastDataFrame(engine={self.engine.value}, type={type(self.data).__name__})"
@property
def is_lazy(self) -> bool:
"""Check if the underlying DataFrame is lazy (Spark, Dask, Ray)."""
return self.engine in [
DataFrameEngine.SPARK,
DataFrameEngine.DASK,
DataFrameEngine.RAY,
]