expr: Make it hashable (#540)

docs/examples/useful-tips/debugging.py

@@ -5,6 +5,7 @@
 import pandas as pd
 
 from fennel.testing import mock
+from fennel.expr import lit
 
 __owner__ = "aditya@fennel.ai"
 
@@ -41,7 +42,7 @@ def my_pipeline(cls, user: Dataset):
                 schema = ds.schema()
                 print(schema)
                 # docsnip-highlight end
-                return ds.assign("country", str, lambda df: "US")
+                return ds.assign(country=lit("US").astype(str))
 
         # /docsnip
 

fennel/expr/test_expr.py

@@ -4,14 +4,22 @@
 from typing import Dict
 from fennel.datasets import dataset
 
-from fennel.expr import col, when
+from fennel.expr import col, when, lit
 from fennel.expr.visitor import ExprPrinter, FetchReferences
 from fennel.expr.serializer import ExprSerializer
 from google.protobuf.json_format import ParseDict  # type: ignore
 from fennel.gen.expr_pb2 import Expr
 from fennel.testing.test_utils import error_message
 
 
+def test_const_expr():
+    expr = lit(1, int)
+    assert expr.typeof({}) == int
+    df = pd.DataFrame({"a": [1, 2, 3, 4]})
+    df2 = expr.eval(df, {"a": int})
+    assert df2.tolist() == [1, 1, 1, 1]
+
+
 def test_basic_expr1():
     expr = (col("num") + col("d")).isnull()
     df = pd.DataFrame({"num": [1, 2, 3, 4], "d": [5, 6, 7, 8]})

fennel/testing/executor.py

@@ -812,6 +812,7 @@ def visitAssign(self, obj):
         else:
             input_df = copy.deepcopy(input_ret.df)
             df = copy.deepcopy(input_df)
+            df.reset_index(drop=True, inplace=True)
             for col, typed_expr in obj.output_expressions.items():
                 if col in input_ret.df.columns:
                     raise Exception(

fennel/utils.py

@@ -9,7 +9,7 @@
 import sys
 import textwrap
 from typing import Any, cast, Callable, Dict, List, Tuple, Union
-
+from fennel.expr.expr import TypedExpr, Expr
 from pandas import DataFrame
 
 import fennel._vendor.astunparse as astunparse
@@ -163,6 +163,12 @@ def _json_default(item: Any):
     if isinstance(item, datetime.timedelta):
         return str(item.total_seconds())
 
+    if isinstance(item, TypedExpr):
+        return str(item.expr) + str(item.dtype)
+
+    if isinstance(item, Expr):
+        return str(item)
+
     raise TypeError(f"object of type {type(item).__name__} not hashable")
 
 

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "fennel-ai"
-version = "1.5.8"
+version = "1.5.9"
 description = "The modern realtime feature engineering platform"
 authors = ["Fennel AI <developers@fennel.ai>"]
 packages = [{ include = "fennel" }]