class SampleDataSchema(pa.SchemaModel):
"""pandera schema of the parquet test dataset"""
registration_dttm: pa.typing.Series[pa.typing.DateTime]
id: pa.typing.Series[pd.Int64Dtype] = pa.Field(nullable=True, coerce=True)
first_name: pa.typing.Series[pa.typing.String]
last_name: pa.typing.Series[pa.typing.String]
email: pa.typing.Series[pa.typing.String]
gender: pa.typing.Series[pa.typing.String] = pa.Field(coerce=True)
ip_address: pa.typing.Series[pa.typing.String]
cc: pa.typing.Series[pa.typing.String]
country: pa.typing.Series[pa.typing.String]
birthdate: pa.typing.Series[pa.typing.String]
salary: pa.typing.Series[pa.typing.Float64] = pa.Field(nullable=True)
title: pa.typing.Series[pa.typing.String]
comments: pa.typing.Series[pa.typing.String] = pa.Field(nullable=True)
@staticmethod
def length():
"""Known length of the data"""
return 5000
@staticmethod
def n_salary_over_150000():
"""Number of rows with salary > 150000"""
return 2384
class Base(pa.SchemaModel):
a: Series[int]
b: Series[int] = pa.Field()
c: Series[int] = pa.Field(alias="_c")
d: Series[int] = pa.Field(alias=123)
i1: Index[int]
i2: Index[int] = pa.Field()
class BaseSchema(pa.SchemaModel):
square_footage: Series[int] = pa.Field(in_range={"min_value": 0, "max_value": 3000})
n_bedrooms: Series[int] = pa.Field(in_range={"min_value": 0, "max_value": 10})
price: Series[int] = pa.Field(in_range={"min_value": 0, "max_value": 1000000})
class Config:
coerce = True
class BollingerBands(pa.SchemaModel):
"""Bollinger bands for a set of stock prices."""
name: Series[str] = pa.Field(description="Ticker symbol of stock")
date: Series[pd.Timestamp] = pa.Field(description="Date of prices")
upper: Series[float] = pa.Field(ge=0, description="Upper band")
lower: Series[float] = pa.Field(description="Lower band")
class StockPrices(pa.SchemaModel):
"""Open/close prices for one or more stocks by day."""
name: Series[str] = pa.Field(description="Ticker symbol of stock")
date: Series[str] = pa.Field(description="Date of prices")
open: Series[float] = pa.Field(ge=0, description="Price at market open")
close: Series[float] = pa.Field(ge=0, description="Price at market close")
class StatsBombGameSchema(GameSchema):
"""Definition of a dataframe containing a list of games."""
competition_stage: Series[str]
home_score: Series[int]
away_score: Series[int]
venue: Series[str] = pa.Field(nullable=True)
referee_id: Series[int] = pa.Field(nullable=True)
class Schema(pa.SchemaModel):
col_2020: Series[int] = pa.Field(alias=2020)
idx: Index[int] = pa.Field(alias="_idx", check_name=True)
@pa.check(2020)
@classmethod
def int_column_lt_100(cls, series: pd.Series) -> Iterable[bool]:
return series < 100
class AnomalousEvents(pa.SchemaModel):
"""Anomalous price events, defined by a day on which a stock's closing price strayed above or
below its Bollinger bands."""
date: Series[pd.Timestamp] = pa.Field(description="Date of price event")
name: Series[str] = pa.Field(description="Ticker symbol of stock")
event: Series[pd.CategoricalDtype] = pa.Field(
description="Type of event: 'high' or low'")
class Child(Base):
b: Series[str] = pa.Field(alias="_b")
c: Series[str]
d: Series[str] = pa.Field()
extra1: Series[int]
extra2: Series[int] = pa.Field()
extra3: Series[int] = pa.Field(alias="_extra3")
i1: Index[str]
i3: Index[int] = pa.Field(alias="_i3")
class CovidPatientDataframe(pa.SchemaModel):
txn_date: Series[DateTime] = pa.Field(ge=datetime(2020, 1, 1), coerce=True)
gender: Series[Category]
age_number: Series[Int] = pa.Field(coerce=True)
age_range: Series[String]
job: Series[String]
risk: Series[String]
patient_type: Series[Category]
province: Series[String]
update_date: Series[DateTime] = pa.Field(ge=datetime(2020, 1, 1),
coerce=True)
class InSchema(pa.SchemaModel):
hourly_pay: Series[float] = pa.Field(ge=7)
hours_worked: Series[float] = pa.Field(ge=10)
@pa.check("hourly_pay", "hours_worked")
def check_numbers_are_positive(cls, series: Series) -> Series[bool]:
"""Defines a column-level custom check."""
return series > 0
class Config:
coerce = True
class Schema(pa.SchemaModel):
"""Schema that uses registered checks in Field."""
col1: pa.typing.Series[int] = pa.Field(custom_gt=100)
col2: pa.typing.Series[float] = pa.Field(custom_in_range={
"min_value": -10,
"max_value": 10
})
class Config:
coerce = True
class EventSchema(pa.SchemaModel):
"""Definition of a dataframe containing event stream data of a game."""
game_id: Series[int]
event_id: Series[int]
period_id: Series[int]
team_id: Series[int] = pa.Field(nullable=True)
player_id: Series[int] = pa.Field(nullable=True)
type_id: Series[int]
type_name: Series[str]
class Config: # noqa: D106
strict = True
class Schema(pa.SchemaModel):
a: Series[int] = pa.Field(eq=1)
@pa.check("a")
@classmethod
def int_column_lt_100(cls, series: pd.Series) -> Iterable[bool]:
return series < 100
def test_field_to_index():
"""Test that Field outputs the correct index options."""
for flag in ["nullable", "allow_duplicates"]:
for value in [True, False]:
index = pa.Field(**{flag: value}).to_index(pa.DateTime)
assert isinstance(index, pa.Index)
assert index.dtype == pa.DateTime.value
assert getattr(index, flag) == value
class Schema(pa.SchemaModel):
a: Series[int] = pa.Field(alias=2020, gt=50)
@pa.check(a)
def int_column_lt_100(cls, series: pd.Series) -> Iterable[bool]:
# pylint:disable=no-self-argument
assert cls is Schema
return series < 100
class SampleSchemaModel(pa.SchemaModel):
a: pa.typing.Series[int] = pa.Field(le=10, description="a desc")
b: pa.typing.Series[float] = pa.Field(lt=-1.2, description="b desc")
c: pa.typing.Series[str] = pa.Field(str_startswith="value_",
description="c desc")
@pa.check("c")
def c_check( # pylint: disable=no-self-argument
cls, series: pa.typing.Series[str]) -> pa.typing.Series[bool]:
"""Two words separated by underscore"""
return series.str.split("_", expand=True).shape[1] == 2
@pa.dataframe_check
def a_gt_b(cls, df):
"""sum(a) > sum(b)"""
return df["a"].sum() > df["b"].sum()
Config = make_schema_model_config(**config_attrs)
def test_field_to_index() -> None:
"""Test that Field outputs the correct index options."""
for flag in ["nullable", "unique"]:
for value in [True, False]:
index = pa.Field(**{
flag: value
}).to_index(pa.DateTime) # type: ignore[arg-type]
assert isinstance(index, pa.Index)
assert index.dtype == Engine.dtype(pa.DateTime)
assert getattr(index, flag) == value
def test_field_to_column():
"""Test that Field outputs the correct column options."""
for flag in ["nullable", "allow_duplicates", "coerce", "regex"]:
for value in [True, False]:
col = pa.Field(**{
flag: value
}).to_column(pa.DateTime, required=value)
assert isinstance(col, pa.Column)
assert col.dtype == pa.DateTime.value
assert col.properties[flag] == value
assert col.required == value
def test_field_to_column() -> None:
"""Test that Field outputs the correct column options."""
for flag in ["nullable", "unique", "coerce", "regex"]:
for value in [True, False]:
col = pa.Field(**{
flag: value
}).to_column( # type: ignore[arg-type]
pa.DateTime, required=value)
assert isinstance(col, pa.Column)
assert col.dtype == Engine.dtype(pa.DateTime)
assert col.properties[flag] == value
assert col.required == value
class StockPrices(pa.SchemaModel):
"""Open/high/low/close prices for a set of stocks by day."""
name: Series[str] = pa.Field(description="Ticker symbol of stock")
date: Series[pd.Timestamp] = pa.Field(description="Date of prices")
open: Series[float] = pa.Field(ge=0, description="Price at market open")
high: Series[float] = pa.Field(ge=0,
description="Highest price of the day")
low: Series[float] = pa.Field(ge=0, description="Lowest price of the day")
close: Series[float] = pa.Field(ge=0, description="Price at market close")
volume: Series[int] = pa.Field(
ge=0, description="Number of shares traded for day")
class SampleDataSchema(pa.SchemaModel):
"""pandera schema of the test dataset"""
col_timedelta: pa.typing.Series[pa.typing.Timedelta] = pa.Field(nullable=True)
col_datetime: pa.typing.Series[pa.typing.DateTime] = pa.Field(nullable=True)
col_bool: pa.typing.Series[pa.typing.Bool] = pa.Field(nullable=True)
col_int: pa.typing.Series[pa.typing.Int] = pa.Field(nullable=True)
col_string: pa.typing.Series[pa.typing.String] = pa.Field(nullable=True)
col_float: pa.typing.Series[pa.typing.Float64] = pa.Field(nullable=True)
class CovidSitRepDataFrame(pa.SchemaModel):
txn_date: Series[DateTime] = pa.Field(ge=datetime(2020, 1, 1), coerce=True)
province: Optional[Series[String]]
new_case: Series[Int] = pa.Field(ge=0, coerce=True)
total_case: Series[Int] = pa.Field(ge=0, coerce=True)
new_case_excludeabroad: Series[Int]
total_case_excludeabroad: Series[Int]
new_death: Series[Int] = pa.Field(ge=0, coerce=True)
total_death: Series[Int] = pa.Field(ge=0, coerce=True)
new_recovered: Optional[Series[Int]]
total_recovered: Optional[Series[Int]]
update_date: Optional[Series[DateTime]] = pa.Field(ge=datetime(2020, 1, 1),
coerce=True)
class StatsBombEventSchema(EventSchema):
"""Definition of a dataframe containing event stream data of a game."""
event_id: Series[object]
index: Series[int]
timestamp: Series[DateTime]
minute: Series[int]
second: Series[int] = pa.Field(ge=0, le=59)
possession: Series[int]
possession_team_id: Series[int]
possession_team_name: Series[str]
play_pattern_id: Series[int]
play_pattern_name: Series[str]
team_name: Series[str]
duration: Series[float] = pa.Field(nullable=True)
extra: Series[object]
related_events: Series[object]
player_name: Series[str] = pa.Field(nullable=True)
position_id: Series[int] = pa.Field(nullable=True)
position_name: Series[str] = pa.Field(nullable=True)
location: Series[object] = pa.Field(nullable=True)
under_pressure: Series[bool] = pa.Field(nullable=True)
counterpress: Series[bool] = pa.Field(nullable=True)
class ChildField(Base):
a: Series[str] = pa.Field()
b: Series[str] = pa.Field()
c: Series[str] = pa.Field()
class MISchema(pa.SchemaModel):
idx1: Index[int] = pa.Field(alias="index0")
idx2: Index[int] = pa.Field(alias="index1")
class Schema(pa.SchemaModel):
a: Optional[Series[str]]
b: Optional[Series[str]] = pa.Field(eq="b")
c: Optional[Series[String]] # test pandera.typing alias
class Schema(pa.SchemaModel):
col_regex: Series[str] = pa.Field(alias="column_([0-9])+", regex=True)
class Base(pa.SchemaModel):
a: Series[int]
b: Series[int] = pa.Field()
c: Series[int] = pa.Field(alias="_c")
class ChildAlias(Base):
a: Series[str] = pa.Field(alias="_a")
b: Series[str] = pa.Field(alias="_b")
c: Series[str] = pa.Field(alias="_c")