Integrate web Tree sitter

Author: AndreasArvidsson

packages/common/package.json

@@ -26,7 +26,8 @@
     "cross-spawn": "7.0.5",
     "fast-check": "3.22.0",
     "js-yaml": "^4.1.0",
-    "mocha": "^10.7.3"
+    "mocha": "^10.7.3",
+    "web-tree-sitter": "^0.25.3"
   },
   "types": "./out/index.d.ts",
   "exports": {

packages/common/src/types/TreeSitter.ts

@@ -1,11 +1,11 @@
 import type { Range, TextDocument } from "@cursorless/common";
-import type { Language, SyntaxNode, Tree } from "web-tree-sitter";
+import type { Language, Node, Tree } from "web-tree-sitter";
 
 export interface TreeSitter {
   /**
    * Function to access nodes in the tree sitter.
    */
-  getNodeAtLocation(document: TextDocument, range: Range): SyntaxNode;
+  getNodeAtLocation(document: TextDocument, range: Range): Node;
 
   /**
    * Function to access the tree sitter tree.

packages/cursorless-engine/package.json

@@ -32,6 +32,7 @@
     "nearley": "2.20.1",
     "talon-snippets": "1.3.0",
     "uuid": "^10.0.0",
+    "web-tree-sitter": "^0.25.3",
     "zod": "3.23.8"
   },
   "devDependencies": {

packages/cursorless-engine/src/disabledComponents/DisabledLanguageDefinitions.ts

@@ -1,5 +1,5 @@
 import type { TextDocument, Range, Listener } from "@cursorless/common";
-import type { SyntaxNode } from "web-tree-sitter";
+import type { Node } from "web-tree-sitter";
 import type { LanguageDefinition } from "../languages/LanguageDefinition";
 import type { LanguageDefinitions } from "../languages/LanguageDefinitions";
 
@@ -16,10 +16,7 @@ export class DisabledLanguageDefinitions implements LanguageDefinitions {
     return undefined;
   }
 
-  getNodeAtLocation(
-    _document: TextDocument,
-    _range: Range,
-  ): SyntaxNode | undefined {
+  getNodeAtLocation(_document: TextDocument, _range: Range): Node | undefined {
     return undefined;
   }
 

packages/cursorless-engine/src/disabledComponents/DisabledTreeSitter.ts

@@ -1,5 +1,5 @@
 import type { Range, TextDocument, TreeSitter } from "@cursorless/common";
-import type { Language, SyntaxNode, Tree } from "web-tree-sitter";
+import type { Language, Node, Tree } from "web-tree-sitter";
 
 export class DisabledTreeSitter implements TreeSitter {
   getTree(_document: TextDocument): Tree {
@@ -14,7 +14,7 @@ export class DisabledTreeSitter implements TreeSitter {
     throw new Error("Tree sitter not provided");
   }
 
-  getNodeAtLocation(_document: TextDocument, _range: Range): SyntaxNode {
+  getNodeAtLocation(_document: TextDocument, _range: Range): Node {
     throw new Error("Tree sitter not provided");
   }
 }

packages/cursorless-engine/src/languages/LanguageDefinition.ts

@@ -11,6 +11,7 @@ import {
   type IDE,
   type TextDocument,
 } from "@cursorless/common";
+import { Query } from "web-tree-sitter";
 import { TreeSitterScopeHandler } from "../processTargets/modifiers/scopeHandlers";
 import { TreeSitterQuery } from "./TreeSitterQuery";
 import type { QueryCapture } from "./TreeSitterQuery/QueryCapture";
@@ -59,9 +60,14 @@ export class LanguageDefinition {
       return undefined;
     }
 
-    const rawQuery = treeSitter
-      .getLanguage(languageId)!
-      .query(rawLanguageQueryString);
+    const language = treeSitter.getLanguage(languageId);
+
+    if (language == null) {
+      throw Error(`Could not get Tree sitter language for ${languageId}`);
+    }
+
+    const rawQuery = new Query(language, rawLanguageQueryString);
+
     const query = TreeSitterQuery.create(languageId, treeSitter, rawQuery);
 
     return new LanguageDefinition(query);

packages/cursorless-engine/src/languages/LanguageDefinitions.ts

@@ -8,7 +8,7 @@ import {
   type TreeSitter,
 } from "@cursorless/common";
 import { toString } from "lodash-es";
-import type { SyntaxNode } from "web-tree-sitter";
+import type { Node } from "web-tree-sitter";
 import { LanguageDefinition } from "./LanguageDefinition";
 import { treeSitterQueryCache } from "./TreeSitterQuery/treeSitterQueryCache";
 
@@ -36,10 +36,7 @@ export interface LanguageDefinitions {
   /**
    * @deprecated Only for use in legacy containing scope stage
    */
-  getNodeAtLocation(
-    document: TextDocument,
-    range: Range,
-  ): SyntaxNode | undefined;
+  getNodeAtLocation(document: TextDocument, range: Range): Node | undefined;
 }
 
 /**
@@ -163,7 +160,7 @@ export class LanguageDefinitionsImpl
     return definition === LANGUAGE_UNDEFINED ? undefined : definition;
   }
 
-  public getNodeAtLocation(document: TextDocument, range: Range): SyntaxNode {
+  public getNodeAtLocation(document: TextDocument, range: Range): Node {
     return this.treeSitter.getNodeAtLocation(document, range);
   }
 

packages/cursorless-engine/src/languages/TreeSitterQuery/PredicateOperatorSchemaTypes.ts

@@ -32,7 +32,7 @@ export type InferSchemaType<T extends HasSchema> = T["schema"];
  * Maps from an operand schema output type to the type of the argument that
  * will be passed to the `accept` function of the predicate operator.  For example:
  *
- * - `{type: "capture", name: string}` -> `SyntaxNode`
+ * - `{type: "capture", name: string}` -> `Node`
  * - `{type: "integer", value: number}` -> `number`
  */
 type PredicateParameterType<T extends SchemaOutputType> = T extends {

packages/cursorless-engine/src/languages/TreeSitterQuery/QueryPredicateOperator.ts

@@ -1,4 +1,4 @@
-import type { PredicateOperand } from "web-tree-sitter";
+import type { PredicateStep } from "web-tree-sitter";
 import type { z } from "zod";
 import type {
   AcceptFunctionArgs,
@@ -41,7 +41,7 @@ export abstract class QueryPredicateOperator<T extends HasSchema> {
    *
    * @param args The arguments to the operator, converted to the types specified
    * in the schema.  For example, if the schema is `z.tuple([q.node, q.string])`,
-   * then `args` will be `SyntaxNode, string`.
+   * then `args` will be `Node, string`.
    */
   protected abstract run(
     ...args: AcceptFunctionArgs<z.infer<InferSchemaType<T>>>
@@ -71,7 +71,7 @@ export abstract class QueryPredicateOperator<T extends HasSchema> {
    * @returns Either a predicate function, or a list of error messages if the operands
    * were invalid.
    */
-  createPredicate(inputOperands: PredicateOperand[]): PredicateResult {
+  createPredicate(inputOperands: PredicateStep[]): PredicateResult {
     const result = this.schema.safeParse(inputOperands);
 
     return result.success

packages/cursorless-engine/src/languages/TreeSitterQuery/constructZodErrorMessages.ts

@@ -1,10 +1,10 @@
-import type { PredicateOperand } from "web-tree-sitter";
+import type { PredicateStep } from "web-tree-sitter";
 import type { z } from "zod";
 import { operandToString } from "./predicateToString";
 
 export function constructZodErrorMessages(
-  inputOperands: PredicateOperand[],
-  error: z.ZodError<PredicateOperand[]>,
+  inputOperands: PredicateStep[],
+  error: z.ZodError<PredicateStep[]>,
 ): string[] {
   return error.errors
     .filter(
@@ -21,7 +21,7 @@ export function constructZodErrorMessages(
     .map((error) => getErrorMessage(inputOperands, error));
 }
 
-function getErrorMessage(inputOperands: PredicateOperand[], error: z.ZodIssue) {
+function getErrorMessage(inputOperands: PredicateStep[], error: z.ZodIssue) {
   if (error.path.length === 0) {
     if (error.code === "too_small") {
       return "Too few arguments";