The plugin parses Wolfram Language through a bundled tree-sitter WebAssembly grammar and prints the result with Prettier. It supports .wl, .wls, .wlt, .mt, and .m files. No Wolfram Engine or local kernel installation is required.

Not affiliated with the original prettier extension.

At A Glance

| Need | What this package provides | | ------------------- | -------------------------------------------------------------------------------- | | Prettier formatting | A Wolfram parser and printer for Prettier 3 projects. | | Editor support | A bundled VS Code extension with document, range, and format-on-save workflows. | | Diagnostics | Formatter-backed rule findings, line-width hints, and fixable style diffs. | | Zero dependencies | Parsing via a bundled tree-sitter WASM grammar; no Wolfram Engine required. |

Requirements

• Node.js
• Prettier 3.x

Install

Install Prettier and the plugin in the project that contains your Wolfram source:

npm install --save-dev prettier prettier-plugin-wolfram

Add the plugin to your Prettier configuration for CLI use:

{
	"plugins": ["prettier-plugin-wolfram"]
}

CLI Usage

Format a file to stdout:

npx prettier --plugin prettier-plugin-wolfram --parser wolfram file.wl

Write changes back:

npx prettier --plugin prettier-plugin-wolfram --parser wolfram --write file.wl

Check whether files would change:

npx prettier --plugin prettier-plugin-wolfram --parser wolfram --check "src/**/*.wl"

Format a byte range with Prettier's standard range flags:

npx prettier --plugin prettier-plugin-wolfram --parser wolfram --range-start 0 --range-end 200 file.wl

Range formatting expands to the top-level Wolfram expressions touched by the range. This avoids formatting incomplete fragments that the parser cannot parse on their own.

Formatting Features

• Formats Wolfram calls, lists, associations, prefix, postfix, infix, binary, ternary, compound, and leaf nodes produced by the tree-sitter grammar.
• Applies specialized layouts to configurable block-style forms such as Module, With, Block, and DynamicModule.
• Applies condition-first layout to configurable forms such as If and Switch.
• Applies alternating condition/body layout to configurable case forms such as Which.
• Wraps long string literals and flattens nested StringJoin[...] calls into a stable multiline StringJoin[...] layout when needed.
• Preserves leading comments, trailing comments, multiline comments, and top-level comment blocks.
• Aligns trailing documentation comments either automatically or at a configured column.
• Preserves ordinary blank lines up to a configured cap and controls spacing between adjacent definitions, including separate Set and SetDelayed definition groups.
• Optionally aligns Rule and RuleDelayed values in multiline argument, list, and association layouts.
• Normalizes general infix x ~ f ~ y to f[x, y] unless the function head is listed in wolfram.preserveTildeInfixFunctions.
• Leaves parse-error files and unsupported CST nodes as original source instead of printing internal parser details.
• Produces idempotent output: a second format pass should not change the first pass result.

Prettier Configuration

Typical .prettierrc example:

{
	"plugins": ["prettier-plugin-wolfram"],
	"printWidth": 80,
	"tabWidth": 2,
	"wolfram": {
		"newlinesBetweenDefinitions": 1,
		"newlinesBetweenSameNameDefinitions": 0,
		"maxBlankLinesBetweenCode": 1,
		"trailingNewline": false,
		"spaceAfterComma": true,
		"spaceAroundOperators": true,
		"alignRuleValues": false,
		"documentationCommentColumn": 0,
		"documentationCommentPadding": 2,
		"documentationCommentMarkers": false,
		"topLevelSpacingMode": "declarations",
		"preserveTildeInfixFunctions": "",
		"moduleVarsBreakThreshold": 40,
		"conditionFirstFunctions": "If,Switch",
		"blockStructureFunctions": "Module,With,Block,DynamicModule",
		"caseStructureFunctions": "Which",
		"lintRules": "{}"
	}
}

Use Prettier's standard printWidth, tabWidth, useTabs, endOfLine, and related options for normal Prettier behavior. The options below are specific to this plugin.

| Option | Type | Default | Description | | ----------------------------------------------------- | ----------- | ----------------------------------- | ----------- | | wolfram.newlinesBetweenDefinitions | integer | 1 | Blank lines inserted between adjacent top-level definitions such as Set, SetDelayed, TagSet, TagSetDelayed, UpSet, and UpSetDelayed. | | wolfram.newlinesBetweenSetDefinitions | integer | inherit | Blank lines inserted between adjacent Set-family definitions. | | wolfram.newlinesBetweenSetDelayedDefinitions | integer | inherit | Blank lines inserted between adjacent SetDelayed-family definitions. | | wolfram.newlinesBetweenSetAndSetDelayedDefinitions | integer | inherit | Blank lines inserted between mixed Set-family and SetDelayed-family definitions. | | wolfram.newlinesBetweenSameNameDefinitions | integer | 0 | Blank lines inserted between adjacent definitions that belong to the same symbol. | | wolfram.maxBlankLinesBetweenCode | integer | 1 | Maximum source blank lines preserved between non-definition code statements. | | wolfram.trailingNewline | boolean | false | Emits one trailing newline at the end of non-empty formatted files. | | wolfram.spaceAfterComma | boolean | true | Inserts a space after commas in argument lists, lists, and associations. | | wolfram.spaceAroundOperators | boolean | true | Inserts spaces around most infix, binary, and ternary operators. Operators that are normally tight, such as ::, ?, and ;;, stay tight. | | wolfram.alignRuleValues | boolean | false | Vertically aligns Rule and RuleDelayed values in multiline argument, list, and association layouts. | | wolfram.documentationCommentColumn | integer | 0 | Column for trailing documentation comments. 0 computes a column per contiguous block. | | wolfram.documentationCommentPadding | integer | 2 | Minimum spaces between code and an aligned trailing documentation comment when the column is computed automatically. | | wolfram.documentationCommentMarkers | boolean | false | Treats trailing comments beginning with < as documentation comments aligned at printWidth. | | wolfram.topLevelSpacingMode | string | "declarations" | Top-level blank-line policy. Allowed values are declarations, all, and none. | | wolfram.preserveTildeInfixFunctions | string | "" | Comma-separated function names that stay in x ~ f ~ y form instead of normalizing to f[x, y]. | | wolfram.moduleVarsBreakThreshold | integer | 40 | Character count at which block-structure variable lists break across lines. | | wolfram.conditionFirstFunctions | string | "If,Switch" | Comma-separated heads whose first argument stays on the same line as the head when it fits. | | wolfram.blockStructureFunctions | string | "Module,With,Block,DynamicModule" | Comma-separated heads formatted with block-structure argument layout. | | wolfram.caseStructureFunctions | string | "Which" | Comma-separated heads formatted with alternating condition/body indentation. | | wolfram.lintRules | string | "{}" | JSON object string for rule-level overrides used by lint integrations, for example {"prefer-rule-delayed":"error"}. |

Top-Level Spacing

wolfram.topLevelSpacingMode controls how the blank-line options are applied:

| Value | Behavior | | -------------- | ----------- | | declarations | Adjacent definitions use wolfram.newlinesBetweenDefinitions plus any configured Set/SetDelayed and same-name overrides; other top-level code preserves source blank lines up to wolfram.maxBlankLinesBetweenCode. | | all | All top-level statements require at least one blank line when wolfram.maxBlankLinesBetweenCode allows it, while still preserving no more than the configured maximum. | | none | Removes top-level blank lines. |

Set wolfram.maxBlankLinesBetweenCode to 0 to remove preserved ordinary code gaps. Set wolfram.newlinesBetweenDefinitions to 0 to keep adjacent definitions together.

The Set/SetDelayed override options inherit wolfram.newlinesBetweenDefinitions when omitted. Same-name definition groups, such as usage messages, options, attributes, and overloads for the same symbol, use wolfram.newlinesBetweenSameNameDefinitions.

Custom Form Layout

The layout categories are driven by comma-separated symbol lists:

{
	"wolfram": {
		"conditionFirstFunctions": "If,Switch,MyConditionForm",
		"blockStructureFunctions": "Module,With,Block,DynamicModule,MyBlock",
		"caseStructureFunctions": "Which,MyCaseForm"
	}
}

Direct comments inside a specialized call cause the formatter to fall back to the generic call layout so comments stay attached to the original arguments.

Lint CLI

The package also exposes a small rule runner:

npx prettier-wolfram lint "src/**/*.wl"

Matched directories and non-Wolfram file extensions are skipped.

It prints diagnostics as:

path/to/file.wl:line:column: WARN [rule-name] message

Override rule levels with WOLFRAM_LINT_RULES:

WOLFRAM_LINT_RULES='{"no-bare-symbol-set":"off","prefer-rule-delayed":"error"}' \
  npx prettier-wolfram lint "src/**/*.wl"

Rule levels are off, warn, and error.

| Rule | Default | Formatter-fixable | Description | | ------------------------------ | ------- | ----------------- | ----------- | | line-width | warn | yes | Reports lines exceeding printWidth, ignoring comment-only overflow. | | newlines-between-definitions | warn | yes | Reports top-level blank-line spacing that differs from the configured policy. | | spacing-operators | warn | yes | Reports operator spacing inconsistent with wolfram.spaceAroundOperators. | | spacing-commas | warn | yes | Reports comma spacing inconsistent with wolfram.spaceAfterComma. | | no-general-infix-function | warn | yes | Reports general infix x ~ f ~ y forms unless f is preserved. | | prefer-rule-delayed | warn | no | Reports definitions where SetDelayed is safer because the right-hand side references pattern variables. | | no-bare-symbol-set | warn | no | Reports top-level global symbol assignments such as x = value. | | no-dynamic-module-leak | warn | no | Reports assignments inside Module, Block, or DynamicModule to symbols missing from the variable list. | | no-shadowed-pattern | error | no | Reports pattern variables that shadow local variables from Module, With, or Block. | | no-unused-module-var | warn | no | Reports unused variables declared in Module, With, Block, or DynamicModule. |

Parse Runtime

The plugin uses a bundled tree-sitter WebAssembly grammar (src/parser/tree-sitter-wolfram.wasm) to parse Wolfram Language. No Wolfram Engine, Mathematica installation, or kernel process is required at runtime.

VS Code Extension

Build the standalone .vsix that bundles Prettier and this plugin:

npm run package:vscode:standalone

This writes vscode-extension/prettier-vscode-wolfram-<version>.vsix.

For Marketplace pre-release publishing, use the workspace-safe publish wrapper:

npm run publish:vscode:pre-release

The extension README with editor setup, settings, diagnostics, and file association behavior lives at vscode-extension/README.md.

Publishing To npm

Log in:

npm login

Preview package contents:

npm pack --dry-run

Publish:

npm publish

Verify:

npm view prettier-plugin-wolfram