@optave/codegraph 3.7.0 → 3.8.0

package/src/extractors/r.ts

@@ -0,0 +1,253 @@
+import type { ExtractorOutput, SubDeclaration, TreeSitterNode, TreeSitterTree } from '../types.js';
+import { findChild, nodeEndLine } from './helpers.js';
+
+/**
+ * Extract symbols from R files.
+ *
+ * tree-sitter-r grammar (r-lib/tree-sitter-r) notes:
+ * - Assignments: binary_operator with `<-` or `=` operator
+ * - Functions: function_definition as RHS of assignment
+ * - Calls: call node with function/arguments fields
+ * - Imports: library() and require() calls
+ * - S4 classes: setClass(), setRefClass()
+ */
+export function extractRSymbols(tree: TreeSitterTree, _filePath: string): ExtractorOutput {
+  const ctx: ExtractorOutput = {
+    definitions: [],
+    calls: [],
+    imports: [],
+    classes: [],
+    exports: [],
+    typeMap: new Map(),
+  };
+
+  walkRNode(tree.rootNode, ctx);
+  return ctx;
+}
+
+function walkRNode(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  switch (node.type) {
+    case 'binary_operator':
+      handleBinaryOp(node, ctx);
+      break;
+    case 'call':
+      handleCall(node, ctx);
+      break;
+  }
+
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (child) walkRNode(child, ctx);
+  }
+}
+
+function handleBinaryOp(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  // binary_operator: child[0]=LHS, child[1]=operator (<- or =), child[2]=RHS
+  if (node.childCount < 3) return;
+
+  const lhs = node.child(0);
+  const op = node.child(1);
+  const rhs = node.child(2);
+
+  if (!lhs || !op || !rhs) return;
+  if (op.text !== '<-' && op.text !== '=' && op.text !== '<<-') return;
+  if (lhs.type !== 'identifier') return;
+
+  if (rhs.type === 'function_definition') {
+    const params = extractRParams(rhs);
+    ctx.definitions.push({
+      name: lhs.text,
+      kind: 'function',
+      line: node.startPosition.row + 1,
+      endLine: nodeEndLine(node),
+      children: params.length > 0 ? params : undefined,
+    });
+  } else {
+    // Variable assignment — only record top-level
+    if (node.parent?.type === 'program') {
+      ctx.definitions.push({
+        name: lhs.text,
+        kind: 'variable',
+        line: node.startPosition.row + 1,
+        endLine: nodeEndLine(node),
+      });
+    }
+  }
+}
+
+function extractRParams(funcDef: TreeSitterNode): SubDeclaration[] {
+  const params: SubDeclaration[] = [];
+  const paramsNode = findChild(funcDef, 'parameters');
+  if (!paramsNode) return params;
+
+  for (let i = 0; i < paramsNode.childCount; i++) {
+    const child = paramsNode.child(i);
+    if (!child) continue;
+    if (child.type === 'parameter') {
+      // parameter node has name and possibly default value
+      const nameNode = child.childForFieldName('name') || findChild(child, 'identifier');
+      if (nameNode) {
+        params.push({ name: nameNode.text, kind: 'parameter', line: child.startPosition.row + 1 });
+      } else if (child.text && child.text !== ',' && child.text !== '(' && child.text !== ')') {
+        // Some grammars have the param as plain text
+        params.push({ name: child.text, kind: 'parameter', line: child.startPosition.row + 1 });
+      }
+    }
+    if (child.type === 'identifier') {
+      params.push({ name: child.text, kind: 'parameter', line: child.startPosition.row + 1 });
+    }
+  }
+  return params;
+}
+
+function handleCall(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  // call: child[0]=function, then arguments
+  const funcNode = node.child(0);
+  if (!funcNode) return;
+
+  const funcName = funcNode.text;
+
+  // library() and require() are imports
+  if (funcName === 'library' || funcName === 'require') {
+    handleLibraryCall(node, ctx);
+    return;
+  }
+
+  // source() is a file import
+  if (funcName === 'source') {
+    handleSourceCall(node, ctx);
+    return;
+  }
+
+  // setClass / setRefClass for S4
+  if (funcName === 'setClass' || funcName === 'setRefClass') {
+    handleSetClass(node, ctx);
+    return;
+  }
+
+  if (funcName === 'setGeneric' || funcName === 'setMethod') {
+    handleSetGeneric(node, ctx);
+    return;
+  }
+
+  // Regular call
+  if (funcNode.type === 'identifier') {
+    ctx.calls.push({ name: funcName, line: node.startPosition.row + 1 });
+  } else if (funcNode.type === 'namespace_operator') {
+    // pkg::func
+    const parts = funcName.split('::');
+    if (parts.length >= 2) {
+      ctx.calls.push({
+        name: parts[parts.length - 1]!,
+        receiver: parts.slice(0, -1).join('::'),
+        line: node.startPosition.row + 1,
+      });
+    }
+  }
+}
+
+function handleLibraryCall(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  // Find the package name in arguments
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (!child) continue;
+    if (child.type === 'arguments') {
+      for (let j = 0; j < child.childCount; j++) {
+        const arg = child.child(j);
+        if (!arg) continue;
+        if (arg.type === 'identifier') {
+          ctx.imports.push({
+            source: arg.text,
+            names: [arg.text],
+            line: node.startPosition.row + 1,
+          });
+          return;
+        }
+        if (arg.type === 'string' || arg.type === 'string_content') {
+          const text = arg.text.replace(/^["']|["']$/g, '');
+          ctx.imports.push({
+            source: text,
+            names: [text],
+            line: node.startPosition.row + 1,
+          });
+          return;
+        }
+        // Argument might be wrapped
+        if (arg.type === 'argument') {
+          const id = findChild(arg, 'identifier') || findChild(arg, 'string');
+          if (id) {
+            const text = id.text.replace(/^["']|["']$/g, '');
+            ctx.imports.push({
+              source: text,
+              names: [text],
+              line: node.startPosition.row + 1,
+            });
+            return;
+          }
+        }
+      }
+    }
+  }
+}
+
+function handleSourceCall(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (!child || child.type !== 'arguments') continue;
+    for (let j = 0; j < child.childCount; j++) {
+      const arg = child.child(j);
+      if (!arg) continue;
+      if (arg.type === 'string') {
+        const text = arg.text.replace(/^["']|["']$/g, '');
+        ctx.imports.push({
+          source: text,
+          names: ['source'],
+          line: node.startPosition.row + 1,
+        });
+        return;
+      }
+    }
+  }
+}
+
+function handleSetClass(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (!child || child.type !== 'arguments') continue;
+    for (let j = 0; j < child.childCount; j++) {
+      const arg = child.child(j);
+      if (!arg) continue;
+      if (arg.type === 'string') {
+        const name = arg.text.replace(/^["']|["']$/g, '');
+        ctx.definitions.push({
+          name,
+          kind: 'class',
+          line: node.startPosition.row + 1,
+          endLine: nodeEndLine(node),
+        });
+        return;
+      }
+    }
+  }
+}
+
+function handleSetGeneric(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (!child || child.type !== 'arguments') continue;
+    for (let j = 0; j < child.childCount; j++) {
+      const arg = child.child(j);
+      if (!arg) continue;
+      if (arg.type === 'string') {
+        const name = arg.text.replace(/^["']|["']$/g, '');
+        ctx.definitions.push({
+          name,
+          kind: 'function',
+          line: node.startPosition.row + 1,
+          endLine: nodeEndLine(node),
+        });
+        return;
+      }
+    }
+  }
+}

package/src/extractors/solidity.ts

@@ -0,0 +1,398 @@
+import type {
+  Call,
+  ExtractorOutput,
+  SubDeclaration,
+  TreeSitterNode,
+  TreeSitterTree,
+} from '../types.js';
+import {
+  extractModifierVisibility,
+  findChild,
+  findParentNode,
+  nodeEndLine,
+  stripQuotes,
+} from './helpers.js';
+
+/**
+ * Extract symbols from Solidity files.
+ *
+ * Solidity's tree-sitter grammar covers contracts, interfaces, libraries,
+ * structs, enums, events, errors, functions, modifiers, and import paths.
+ */
+export function extractSoliditySymbols(tree: TreeSitterTree, _filePath: string): ExtractorOutput {
+  const ctx: ExtractorOutput = {
+    definitions: [],
+    calls: [],
+    imports: [],
+    classes: [],
+    exports: [],
+    typeMap: new Map(),
+  };
+
+  walkSolidityNode(tree.rootNode, ctx);
+  return ctx;
+}
+
+function walkSolidityNode(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  switch (node.type) {
+    case 'contract_declaration':
+      handleContractDecl(node, ctx, 'class');
+      break;
+    case 'interface_declaration':
+      handleContractDecl(node, ctx, 'interface');
+      break;
+    case 'library_declaration':
+      handleContractDecl(node, ctx, 'module');
+      break;
+    case 'struct_declaration':
+      handleStructDecl(node, ctx);
+      break;
+    case 'enum_declaration':
+      handleEnumDecl(node, ctx);
+      break;
+    case 'function_definition':
+      handleFunctionDef(node, ctx);
+      break;
+    case 'modifier_definition':
+      handleModifierDef(node, ctx);
+      break;
+    case 'event_definition':
+      handleEventDef(node, ctx);
+      break;
+    case 'error_declaration':
+      handleErrorDecl(node, ctx);
+      break;
+    case 'state_variable_declaration':
+      handleStateVarDecl(node, ctx);
+      break;
+    case 'import_directive':
+      handleImportDirective(node, ctx);
+      break;
+    case 'call_expression':
+    case 'function_call':
+      handleCallExpression(node, ctx);
+      break;
+  }
+
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (child) walkSolidityNode(child, ctx);
+  }
+}
+
+// ── Handlers ───────────────────────────────────────────────────────────────
+
+const SOL_PARENT_TYPES = [
+  'contract_declaration',
+  'interface_declaration',
+  'library_declaration',
+] as const;
+
+function handleContractDecl(
+  node: TreeSitterNode,
+  ctx: ExtractorOutput,
+  kind: 'class' | 'interface' | 'module',
+): void {
+  const nameNode = node.childForFieldName('name');
+  if (!nameNode) return;
+  const name = nameNode.text;
+
+  const members: SubDeclaration[] = [];
+  const body = node.childForFieldName('body') || findChild(node, 'contract_body');
+  if (body) {
+    for (let i = 0; i < body.childCount; i++) {
+      const child = body.child(i);
+      if (!child) continue;
+      if (child.type === 'function_definition') {
+        const fnName = child.childForFieldName('name');
+        if (fnName) {
+          members.push({ name: fnName.text, kind: 'method', line: child.startPosition.row + 1 });
+        }
+      } else if (child.type === 'state_variable_declaration') {
+        const varName = child.childForFieldName('name');
+        if (varName) {
+          members.push({
+            name: varName.text,
+            kind: 'property',
+            line: child.startPosition.row + 1,
+            visibility: extractSolVisibility(child),
+          });
+        }
+      } else if (child.type === 'event_definition') {
+        const evName = child.childForFieldName('name');
+        if (evName) {
+          members.push({
+            name: evName.text,
+            kind: 'property',
+            decorators: ['event'],
+            line: child.startPosition.row + 1,
+          });
+        }
+      } else if (child.type === 'error_declaration') {
+        const errName = child.childForFieldName('name');
+        if (errName) {
+          members.push({
+            name: errName.text,
+            kind: 'property',
+            decorators: ['error'],
+            line: child.startPosition.row + 1,
+          });
+        }
+      } else if (child.type === 'modifier_definition') {
+        const modName = child.childForFieldName('name');
+        if (modName) {
+          members.push({
+            name: modName.text,
+            kind: 'method',
+            decorators: ['modifier'],
+            line: child.startPosition.row + 1,
+          });
+        }
+      }
+    }
+  }
+
+  ctx.definitions.push({
+    name,
+    kind,
+    line: node.startPosition.row + 1,
+    endLine: nodeEndLine(node),
+    children: members.length > 0 ? members : undefined,
+  });
+
+  // Inheritance
+  const inheritance = findChild(node, 'inheritance_specifier');
+  if (inheritance) {
+    for (let i = 0; i < inheritance.childCount; i++) {
+      const child = inheritance.child(i);
+      if (!child) continue;
+      if (child.type === 'user_defined_type' || child.type === 'identifier') {
+        ctx.classes.push({ name, extends: child.text, line: node.startPosition.row + 1 });
+      }
+    }
+  }
+}
+
+function handleStructDecl(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  const nameNode = node.childForFieldName('name');
+  if (!nameNode) return;
+
+  const members: SubDeclaration[] = [];
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (child && child.type === 'struct_member') {
+      const memberName = child.childForFieldName('name');
+      if (memberName) {
+        members.push({
+          name: memberName.text,
+          kind: 'property',
+          line: child.startPosition.row + 1,
+        });
+      }
+    }
+  }
+
+  const parent = findParentNode(node, SOL_PARENT_TYPES);
+  const fullName = parent ? `${parent}.${nameNode.text}` : nameNode.text;
+
+  ctx.definitions.push({
+    name: fullName,
+    kind: 'struct',
+    line: node.startPosition.row + 1,
+    endLine: nodeEndLine(node),
+    children: members.length > 0 ? members : undefined,
+  });
+}
+
+function handleEnumDecl(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  const nameNode = node.childForFieldName('name');
+  if (!nameNode) return;
+
+  const members: SubDeclaration[] = [];
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (child && child.type === 'enum_value') {
+      members.push({ name: child.text, kind: 'constant', line: child.startPosition.row + 1 });
+    }
+  }
+
+  const parent = findParentNode(node, SOL_PARENT_TYPES);
+  const fullName = parent ? `${parent}.${nameNode.text}` : nameNode.text;
+
+  ctx.definitions.push({
+    name: fullName,
+    kind: 'enum',
+    line: node.startPosition.row + 1,
+    endLine: nodeEndLine(node),
+    children: members.length > 0 ? members : undefined,
+  });
+}
+
+function handleFunctionDef(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  const nameNode = node.childForFieldName('name');
+  if (!nameNode) return;
+  const parent = findParentNode(node, SOL_PARENT_TYPES);
+  const fullName = parent ? `${parent}.${nameNode.text}` : nameNode.text;
+  const kind = parent ? 'method' : 'function';
+
+  const params = extractSolParams(node);
+  ctx.definitions.push({
+    name: fullName,
+    kind,
+    line: node.startPosition.row + 1,
+    endLine: nodeEndLine(node),
+    children: params.length > 0 ? params : undefined,
+    visibility: extractSolVisibility(node),
+  });
+}
+
+function handleModifierDef(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  const nameNode = node.childForFieldName('name');
+  if (!nameNode) return;
+  const parent = findParentNode(node, SOL_PARENT_TYPES);
+  const fullName = parent ? `${parent}.${nameNode.text}` : nameNode.text;
+
+  ctx.definitions.push({
+    name: fullName,
+    kind: 'function',
+    line: node.startPosition.row + 1,
+    endLine: nodeEndLine(node),
+    decorators: ['modifier'],
+  });
+}
+
+function handleEventDef(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  const nameNode = node.childForFieldName('name');
+  if (!nameNode) return;
+  const parent = findParentNode(node, SOL_PARENT_TYPES);
+  const fullName = parent ? `${parent}.${nameNode.text}` : nameNode.text;
+
+  ctx.definitions.push({
+    name: fullName,
+    kind: 'type',
+    line: node.startPosition.row + 1,
+    endLine: nodeEndLine(node),
+    decorators: ['event'],
+  });
+}
+
+function handleErrorDecl(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  const nameNode = node.childForFieldName('name');
+  if (!nameNode) return;
+  const parent = findParentNode(node, SOL_PARENT_TYPES);
+  const fullName = parent ? `${parent}.${nameNode.text}` : nameNode.text;
+
+  ctx.definitions.push({
+    name: fullName,
+    kind: 'type',
+    line: node.startPosition.row + 1,
+    endLine: nodeEndLine(node),
+    decorators: ['error'],
+  });
+}
+
+function handleStateVarDecl(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  const nameNode = node.childForFieldName('name');
+  if (!nameNode) return;
+  const parent = findParentNode(node, SOL_PARENT_TYPES);
+  const fullName = parent ? `${parent}.${nameNode.text}` : nameNode.text;
+
+  ctx.definitions.push({
+    name: fullName,
+    kind: 'variable',
+    line: node.startPosition.row + 1,
+    endLine: nodeEndLine(node),
+    visibility: extractSolVisibility(node),
+  });
+}
+
+function handleImportDirective(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  // import "path"; or import { X } from "path"; or import "path" as Alias;
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (!child) continue;
+    if (child.type === 'string' || child.type === 'string_literal') {
+      const source = stripQuotes(child.text);
+      const names: string[] = [];
+      // Look for imported symbols
+      for (let j = 0; j < node.childCount; j++) {
+        const sibling = node.child(j);
+        if (sibling && sibling.type === 'identifier') names.push(sibling.text);
+        if (sibling && sibling.type === 'import_declaration') {
+          const id = findChild(sibling, 'identifier');
+          if (id) names.push(id.text);
+        }
+      }
+      ctx.imports.push({
+        source,
+        names: names.length > 0 ? names : ['*'],
+        line: node.startPosition.row + 1,
+      });
+      return;
+    }
+    // source_import: handles `import * as X from "path"`
+    if (child.type === 'source_import' || child.type === 'import_clause') {
+      const strNode = findChild(child, 'string') || findChild(child, 'string_literal');
+      if (strNode) {
+        ctx.imports.push({
+          source: stripQuotes(strNode.text),
+          names: ['*'],
+          line: node.startPosition.row + 1,
+        });
+        return;
+      }
+    }
+  }
+}
+
+function handleCallExpression(node: TreeSitterNode, ctx: ExtractorOutput): void {
+  const funcNode = node.childForFieldName('function') || node.childForFieldName('callee');
+  if (!funcNode) return;
+
+  const call: Call = { name: '', line: node.startPosition.row + 1 };
+  if (funcNode.type === 'member_expression' || funcNode.type === 'member_access') {
+    const prop = funcNode.childForFieldName('property') || funcNode.childForFieldName('member');
+    const obj = funcNode.childForFieldName('object') || funcNode.childForFieldName('expression');
+    if (prop) call.name = prop.text;
+    if (obj) call.receiver = obj.text;
+  } else {
+    call.name = funcNode.text;
+  }
+  if (call.name) ctx.calls.push(call);
+}
+
+// ── Helpers ────────────────────────────────────────────────────────────────
+
+function extractSolParams(funcNode: TreeSitterNode): SubDeclaration[] {
+  const params: SubDeclaration[] = [];
+  const paramList =
+    funcNode.childForFieldName('parameters') || findChild(funcNode, 'parameter_list');
+  if (!paramList) return params;
+
+  for (let i = 0; i < paramList.childCount; i++) {
+    const param = paramList.child(i);
+    if (!param || param.type !== 'parameter') continue;
+    const nameNode = param.childForFieldName('name');
+    if (nameNode) {
+      params.push({ name: nameNode.text, kind: 'parameter', line: param.startPosition.row + 1 });
+    }
+  }
+  return params;
+}
+
+function extractSolVisibility(
+  node: TreeSitterNode,
+): 'public' | 'private' | 'protected' | undefined {
+  // Solidity visibility is embedded as child keywords or visibility nodes
+  const vis = extractModifierVisibility(node);
+  if (vis) return vis;
+  for (let i = 0; i < node.childCount; i++) {
+    const child = node.child(i);
+    if (!child) continue;
+    const t = child.text;
+    if (t === 'public' || t === 'external') return 'public';
+    if (t === 'private') return 'private';
+    if (t === 'internal') return 'protected';
+  }
+  return undefined;
+}