Skip to content

glguy/toml-parser

Repository files navigation

TOML Parser

This package implements a validating parser for TOML 1.1.0.

This package uses an alex-generated lexer and happy-generated parser.

It also provides a pair of classes for serializing into and out of TOML.

Package Structure

--- title: Package Structure --- stateDiagram-v2 classDef important font-weight:bold; TOML:::important --> ApplicationTypes:::important : decode ApplicationTypes --> TOML : encode TOML --> [Token]: Lexer [Token] --> [Expr]: Parser [Expr] --> Table : Semantics Table --> ApplicationTypes : FromValue ApplicationTypes --> Table : ToValue Table --> TOML : Pretty 
Loading

Most users will only need to import Toml or Toml.Schema. Other top-level modules are for low-level hacking on the TOML format itself. All modules below these top-level modules are exposed to provide direct access to library implementation details.

  • Toml - Basic encoding and decoding TOML
  • Toml.Schema - TOML schemas for application types
  • Toml.Semantics - Low-level semantic operations on TOML syntax
  • Toml.Syntax - Low-level parsing of text into TOML raw syntax

Examples

This file uses markdown-unlit to ensure that its code typechecks and stays in sync with the rest of the package.

{-# Language OverloadedStrings #-} importData.Text (Text) importGHC.Generics (Generic) importQuoteStr (quoteStr) importTest.Hspec (Spec, hspec, it, shouldBe) importTomlimportToml.Schemamain::IO() main = hspec (parses >> decodes >> encodes >> warns >> errors)

Using the raw parser

Consider this sample TOML text from the TOML specification.

fruitStr::Text fruitStr = [quoteStr|
[[fruits]] name = "apple" [fruits.physical] # subtablecolor = "red"shape = "round" [[fruits.varieties]] # nested array of tablesname = "red delicious" [[fruits.varieties]] name = "granny smith" [[fruits]] name = "banana" [[fruits.varieties]] name = "plantain"
|]

Parsing using this package generates the following unstructured value

parses::Spec parses = it "parses"$ forgetTableAnns <$> parse fruitStr `shouldBe`Right (table [ ("fruits", List [ Table (table [ ("name", Text"apple"), ("physical", Table (table [ ("color", Text"red"), ("shape", Text"round")])), ("varieties", List [ Table (table [("name", Text"red delicious")]), Table (table [("name", Text"granny smith")])])]), Table (table [ ("name", Text"banana"), ("varieties", List [ Table (table [("name", Text"plantain")])])])])])

Defining a schema

We can define a schema for our TOML format in the form of instances of FromValue, ToValue, and ToTable in order to read TOML directly into structured data form. This example manually derives some of the instances as a demonstration.

newtypeFruits=Fruits{fruits:: [Fruit] }deriving (Eq, Show, Generic) deriving (ToTable, ToValue, FromValue) viaGenericTomlTableFruitsdataFruit=Fruit{name::String, physical::MaybePhysical, varieties:: [Variety] }deriving (Eq, Show, Generic) deriving (ToTable, ToValue, FromValue) viaGenericTomlTableFruitdataPhysical=Physical{color::String, shape::String}deriving (Eq, Show, Generic) deriving (ToTable, ToValue, FromValue) viaGenericTomlTablePhysicalnewtypeVariety=VarietyStringderiving (Eq, Show) instanceFromValueVarietywhere fromValue = parseTableFromValue (Variety<$> reqKey "name") instanceToValueVarietywhere toValue = defaultTableToValue instanceToTableVarietywhere toTable (Variety x) = table ["name".= x] 

We can run this example on the original value to deserialize it into domain-specific datatypes.

decodes::Spec decodes = it "decodes"$ decode fruitStr `shouldBe`Success[] (Fruits [ Fruit"apple" (Just (Physical"red""round")) [Variety"red delicious", Variety"granny smith"], Fruit"banana"Nothing [Variety"plantain"]]) encodes::Spec encodes = it "encodes"$show (encode (Fruits [Fruit"apple" (Just (Physical"red""round")) [Variety"red delicious", Variety"granny smith"]])) `shouldBe` [quoteStr| [[fruits]] name = "apple" [fruits.physical] color = "red" shape = "round" [[fruits.varieties]] name = "red delicious" [[fruits.varieties]] name = "granny smith"|]

Useful errors and warnings

This package takes care to preserve source information as much as possible in order to provide useful feedback to users. These examples show a couple of the message that can be generated when things don't go perfectly.

warns::Spec warns = it "warns"$ decode [quoteStr| name = "simulated" typo = 10|] `shouldBe`Success ["2:1: unexpected key: typo in <top-level>"] -- warnings (Variety"simulated") errors::Spec errors = it "errors"$ decode [quoteStr| # Physical characteristics table color = "blue" shape = []|] `shouldBe` (Failure ["3:9: expected string but got array in shape"] ::ResultStringPhysical)

More Examples

A demonstration of using this package at a more realistic scale can be found in HieDemoSpec. The various unit test files demonstrate what you can do with this library and what outputs you can expect.

See the low-level operations used to build a TOML syntax highlighter in TomlHighlighter.

About

Haskell parser and printer for the TOML 1.0.0 file format

Topics

Resources

License

Stars

Watchers

Forks

Contributors 4

  •  
  •  
  •  
  •