Quickstart

This page takes five minutes. It assumes that you installed Astrolabe and set up the project.

Declare a schema

A schema is a class that inherits from Columns. Each field is a column, with a Polars data type. Use the schema in the type TypedDataFrame:

quickstart.pymypy: 1 error

1import polars as pl

2

3from astrolabe import Columns, TypedDataFrame

4

5

6class Trades(Columns):

7 symbol: pl.String

8 price: pl.Float64

9

10

11def cheapest(trades: TypedDataFrame[Trades]) -> TypedDataFrame[Trades]:

12 return trades.sort("prices").head(1)

↑error: Unknown column "prices" [unknown-column]

Run mypy. The function sorts by prices, and the frame has price, so the checker reports the line.

Fix the fault, and read a type

Correct the name. Then ask mypy for the type of a new frame with reveal_type:

quickstart.pymypy: no errors

1class Trades(Columns):

2 symbol: pl.String

3 price: pl.Float64

4

5

6def cheapest(trades: TypedDataFrame[Trades]) -> TypedDataFrame[Trades]:

7 return trades.sort("price").head(1)

8

9

10trades: TypedDataFrame[Trades]

Resulttrades.with_columns(doubled=pl.col("price") * 2)declaredadded

symbolStringpriceFloat64doubledFloat64

Revealed type

TypedDataFrame[TypedDict({'symbol': String, 'price': Float64, 'doubled': Float64}, closed=True), None]

The new column doubled has the data type Float64, because a Float64 times an integer gives a Float64 in Polars. The checker computes that type with the rules of Polars, so each later step knows the column and its data type.

Next steps

• Follow the Tutorial to type a whole pipeline, one idea at a time.
• Read the Guides for one topic at a time.
• Look up an error in Error codes.