@eagleoutice/flowr 2.0.0 → 2.0.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (65) hide show
  1. package/benchmark/slicer.js +30 -5
  2. package/benchmark/stats/print.js +28 -9
  3. package/benchmark/stats/stats.d.ts +5 -0
  4. package/benchmark/summarizer/data.d.ts +10 -1
  5. package/benchmark/summarizer/first-phase/process.d.ts +1 -1
  6. package/benchmark/summarizer/first-phase/process.js +83 -30
  7. package/benchmark/summarizer/second-phase/process.js +20 -4
  8. package/cli/repl/commands/commands.js +19 -1
  9. package/cli/slicer-app.js +1 -1
  10. package/dataflow/environments/append.js +1 -2
  11. package/dataflow/environments/built-in.js +2 -1
  12. package/dataflow/environments/clone.js +1 -1
  13. package/dataflow/environments/diff.d.ts +1 -1
  14. package/dataflow/environments/diff.js +16 -18
  15. package/dataflow/environments/environment.d.ts +4 -7
  16. package/dataflow/environments/environment.js +5 -8
  17. package/dataflow/environments/identifier.d.ts +2 -1
  18. package/dataflow/environments/overwrite.js +1 -2
  19. package/dataflow/environments/scoping.js +1 -1
  20. package/dataflow/graph/diff.js +11 -6
  21. package/dataflow/graph/graph.d.ts +6 -2
  22. package/dataflow/graph/graph.js +13 -7
  23. package/dataflow/graph/vertex.d.ts +2 -1
  24. package/dataflow/info.d.ts +10 -1
  25. package/dataflow/info.js +54 -2
  26. package/dataflow/internal/linker.d.ts +1 -1
  27. package/dataflow/internal/linker.js +1 -2
  28. package/dataflow/internal/process/functions/call/built-in/built-in-assignment.js +5 -5
  29. package/dataflow/internal/process/functions/call/built-in/built-in-for-loop.js +1 -1
  30. package/dataflow/internal/process/functions/call/built-in/built-in-function-definition.js +21 -25
  31. package/dataflow/internal/process/functions/call/built-in/built-in-get.js +6 -1
  32. package/dataflow/internal/process/functions/call/built-in/built-in-if-then-else.js +10 -8
  33. package/dataflow/internal/process/functions/call/built-in/built-in-logical-bin-op.d.ts +1 -0
  34. package/dataflow/internal/process/functions/call/built-in/built-in-logical-bin-op.js +1 -2
  35. package/dataflow/internal/process/functions/call/built-in/built-in-while-loop.js +1 -1
  36. package/dataflow/internal/process/functions/call/default-call-handling.js +1 -1
  37. package/dataflow/internal/process/functions/call/unnamed-call-handling.js +1 -1
  38. package/dataflow/processor.d.ts +2 -3
  39. package/package.json +4 -2
  40. package/r-bridge/data/data.d.ts +1 -1
  41. package/r-bridge/data/data.js +1 -1
  42. package/r-bridge/lang-4.x/ast/model/nodes/r-function-call.d.ts +2 -2
  43. package/r-bridge/lang-4.x/ast/model/operators.js +1 -1
  44. package/r-bridge/lang-4.x/ast/model/processing/decorate.js +1 -1
  45. package/r-bridge/lang-4.x/ast/model/processing/stateful-fold.js +1 -1
  46. package/r-bridge/lang-4.x/ast/model/processing/visitor.js +2 -2
  47. package/r-bridge/lang-4.x/ast/parser/xml/internal/functions/normalize-call.js +2 -2
  48. package/r-bridge/lang-4.x/ast/parser/xml/internal/operators/normalize-binary.js +1 -1
  49. package/r-bridge/retriever.d.ts +1 -1
  50. package/r-bridge/retriever.js +3 -2
  51. package/r-bridge/shell.js +4 -3
  52. package/reconstruct/reconstruct.d.ts +3 -3
  53. package/reconstruct/reconstruct.js +40 -41
  54. package/slicing/criterion/filters/all-variables.js +1 -1
  55. package/slicing/static/static-slicer.js +2 -2
  56. package/statistics/features/common-syntax-probability.js +1 -1
  57. package/statistics/features/supported/control-flow/control-flow.js +1 -1
  58. package/statistics/features/supported/defined-functions/defined-functions.js +1 -1
  59. package/statistics/features/supported/loops/loops.js +1 -1
  60. package/statistics/features/supported/used-functions/used-functions.js +1 -1
  61. package/util/assert.d.ts +1 -1
  62. package/util/mermaid/ast.js +4 -0
  63. package/util/mermaid/dfg.js +15 -5
  64. package/util/mermaid/mermaid.js +21 -1
  65. package/util/version.js +1 -1
@@ -24,12 +24,17 @@ function processGet(name, args, rootId, data) {
24
24
  location: retrieve.location,
25
25
  namespace: undefined
26
26
  };
27
- const { information } = (0, known_call_handling_1.processKnownFunctionCall)({
27
+ const { information, processedArguments } = (0, known_call_handling_1.processKnownFunctionCall)({
28
28
  name,
29
29
  args: (0, make_argument_1.wrapArgumentsUnnamed)([treatTargetAsSymbol], data.completeAst.idMap),
30
30
  rootId,
31
31
  data
32
32
  });
33
+ const firstArg = processedArguments[0];
34
+ if (firstArg) {
35
+ // get 'reads' its first argument
36
+ information.graph.addEdge(rootId, firstArg.entryPoint, { type: 1 /* EdgeType.Reads */ });
37
+ }
33
38
  return information;
34
39
  }
35
40
  exports.processGet = processGet;
@@ -57,21 +57,23 @@ function processIfThenElse(name, args, rootId, data) {
57
57
  const thenEnvironment = then?.environment ?? cond.environment;
58
58
  // if there is no "else" case, we have to recover whatever we had before as it may be not executed
59
59
  const finalEnvironment = (0, append_1.appendEnvironment)(thenEnvironment, otherwise ? otherwise.environment : cond.environment);
60
+ const cdTrue = { id: rootId, when: true };
61
+ const cdFalse = { id: rootId, when: false };
60
62
  // again within an if-then-else we consider all actives to be read
61
63
  const ingoing = [
62
64
  ...cond.in,
63
- ...(makeThenMaybe ? (0, environment_1.makeAllMaybe)(then?.in, nextGraph, finalEnvironment, false, rootId) : then?.in ?? []),
64
- ...(makeOtherwiseMaybe ? (0, environment_1.makeAllMaybe)(otherwise?.in, nextGraph, finalEnvironment, false, rootId) : otherwise?.in ?? []),
65
+ ...(makeThenMaybe ? (0, environment_1.makeAllMaybe)(then?.in, nextGraph, finalEnvironment, false, cdTrue) : then?.in ?? []),
66
+ ...(makeOtherwiseMaybe ? (0, environment_1.makeAllMaybe)(otherwise?.in, nextGraph, finalEnvironment, false, cdFalse) : otherwise?.in ?? []),
65
67
  ...cond.unknownReferences,
66
- ...(makeThenMaybe ? (0, environment_1.makeAllMaybe)(then?.unknownReferences, nextGraph, finalEnvironment, false, rootId) : then?.unknownReferences ?? []),
67
- ...(makeOtherwiseMaybe ? (0, environment_1.makeAllMaybe)(otherwise?.unknownReferences, nextGraph, finalEnvironment, false, rootId) : otherwise?.unknownReferences ?? []),
68
+ ...(makeThenMaybe ? (0, environment_1.makeAllMaybe)(then?.unknownReferences, nextGraph, finalEnvironment, false, cdTrue) : then?.unknownReferences ?? []),
69
+ ...(makeOtherwiseMaybe ? (0, environment_1.makeAllMaybe)(otherwise?.unknownReferences, nextGraph, finalEnvironment, false, cdFalse) : otherwise?.unknownReferences ?? []),
68
70
  ];
69
71
  // we assign all with a maybe marker
70
72
  // we do not merge even if they appear in both branches because the maybe links will refer to different ids
71
73
  const outgoing = [
72
74
  ...cond.out,
73
- ...(makeThenMaybe ? (0, environment_1.makeAllMaybe)(then?.out, nextGraph, finalEnvironment, true, rootId) : then?.out ?? []),
74
- ...(makeOtherwiseMaybe ? (0, environment_1.makeAllMaybe)(otherwise?.out, nextGraph, finalEnvironment, true, rootId) : otherwise?.out ?? []),
75
+ ...(makeThenMaybe ? (0, environment_1.makeAllMaybe)(then?.out, nextGraph, finalEnvironment, true, cdTrue) : then?.out ?? []),
76
+ ...(makeOtherwiseMaybe ? (0, environment_1.makeAllMaybe)(otherwise?.out, nextGraph, finalEnvironment, true, cdFalse) : otherwise?.out ?? []),
75
77
  ];
76
78
  (0, common_1.patchFunctionCall)({
77
79
  nextGraph,
@@ -83,8 +85,8 @@ function processIfThenElse(name, args, rootId, data) {
83
85
  // as an if always evaluates its condition, we add a 'reads'-edge
84
86
  nextGraph.addEdge(name.info.id, cond.entryPoint, { type: 1 /* EdgeType.Reads */ });
85
87
  const exitPoints = [
86
- ...(then?.exitPoints ?? []).map(e => ({ ...e, controlDependencies: makeThenMaybe ? [...data.controlDependencies ?? []] : e.controlDependencies })),
87
- ...(otherwise?.exitPoints ?? []).map(e => ({ ...e, controlDependencies: makeOtherwiseMaybe ? [...data.controlDependencies ?? []] : e.controlDependencies }))
88
+ ...(then?.exitPoints ?? []).map(e => ({ ...e, controlDependencies: makeThenMaybe ? [...data.controlDependencies ?? [], { id: rootId, when: true }] : e.controlDependencies })),
89
+ ...(otherwise?.exitPoints ?? []).map(e => ({ ...e, controlDependencies: makeOtherwiseMaybe ? [...data.controlDependencies ?? [], { id: rootId, when: false }] : e.controlDependencies }))
88
90
  ];
89
91
  return {
90
92
  unknownReferences: [],
@@ -6,4 +6,5 @@ import type { RSymbol } from '../../../../../../r-bridge/lang-4.x/ast/model/node
6
6
  import type { NodeId } from '../../../../../../r-bridge/lang-4.x/ast/model/processing/node-id';
7
7
  export declare function processSpecialBinOp<OtherInfo>(name: RSymbol<OtherInfo & ParentInformation>, args: readonly RFunctionArgument<OtherInfo & ParentInformation>[], rootId: NodeId, data: DataflowProcessorInformation<OtherInfo & ParentInformation>, config: {
8
8
  lazy: boolean;
9
+ evalRhsWhen: boolean;
9
10
  }): DataflowInformation;
@@ -14,8 +14,7 @@ function processSpecialBinOp(name, args, rootId, data, config) {
14
14
  const { information, processedArguments } = (0, known_call_handling_1.processKnownFunctionCall)({ name, args, rootId, data,
15
15
  patchData: (d, i) => {
16
16
  if (i === 1) {
17
- // the rhs will be overshadowed by the lhs
18
- return { ...d, controlDependencies: [...d.controlDependencies ?? [], name.info.id] };
17
+ return { ...d, controlDependencies: [...d.controlDependencies ?? [], { id: name.info.id, when: config.evalRhsWhen }] };
19
18
  }
20
19
  return d;
21
20
  }
@@ -28,7 +28,7 @@ function processWhileLoop(name, args, rootId, data) {
28
28
  markAsNSE: [1],
29
29
  patchData: (d, i) => {
30
30
  if (i === 1) {
31
- return { ...d, controlDependencies: [...d.controlDependencies ?? [], name.info.id] };
31
+ return { ...d, controlDependencies: [...d.controlDependencies ?? [], { id: name.info.id, when: true }] };
32
32
  }
33
33
  return d;
34
34
  }
@@ -4,7 +4,7 @@ exports.processFunctionCall = void 0;
4
4
  const named_call_handling_1 = require("./named-call-handling");
5
5
  const unnamed_call_handling_1 = require("./unnamed-call-handling");
6
6
  function processFunctionCall(functionCall, data) {
7
- if (functionCall.flavor === 'named') {
7
+ if (functionCall.named) {
8
8
  return (0, named_call_handling_1.processNamedCall)(functionCall.functionName, functionCall.arguments, functionCall.info.id, data);
9
9
  }
10
10
  else {
@@ -50,7 +50,7 @@ function processUnnamedFunctionCall(functionCall, data) {
50
50
  out: calledFunction.out,
51
51
  graph: finalGraph,
52
52
  environment: finalEnv,
53
- entryPoint: calledRootId,
53
+ entryPoint: functionCall.info.id,
54
54
  exitPoints: calledFunction.exitPoints
55
55
  };
56
56
  }
@@ -1,11 +1,10 @@
1
1
  /**
2
2
  * Based on a two-way fold, this processor will automatically supply scope information
3
3
  */
4
- import type { DataflowInformation } from './info';
4
+ import type { ControlDependency, DataflowInformation } from './info';
5
5
  import type { NormalizedAst, ParentInformation, RNodeWithParent } from '../r-bridge/lang-4.x/ast/model/processing/decorate';
6
6
  import type { REnvironmentInformation } from './environments/environment';
7
7
  import type { RParseRequest } from '../r-bridge/retriever';
8
- import type { NodeId } from '../r-bridge/lang-4.x/ast/model/processing/node-id';
9
8
  import type { RNode } from '../r-bridge/lang-4.x/ast/model/model';
10
9
  export interface DataflowProcessorInformation<OtherInfo> {
11
10
  /**
@@ -33,7 +32,7 @@ export interface DataflowProcessorInformation<OtherInfo> {
33
32
  /**
34
33
  * The chain of control-flow {@link NodeId}s that lead to the current node (e.g. of known ifs).
35
34
  */
36
- readonly controlDependencies: NodeId[] | undefined;
35
+ readonly controlDependencies: ControlDependency[] | undefined;
37
36
  }
38
37
  export type DataflowProcessor<OtherInfo, NodeType extends RNodeWithParent<OtherInfo>> = (node: NodeType, data: DataflowProcessorInformation<OtherInfo>) => DataflowInformation;
39
38
  type NodeWithKey<OtherInfo, Key> = RNode<OtherInfo & ParentInformation> & {
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@eagleoutice/flowr",
3
- "version": "2.0.0",
3
+ "version": "2.0.2",
4
4
  "description": "Static Dataflow Analyzer and Program Slicer for the R Programming Language",
5
5
  "types": "dist/src/index.d.ts",
6
6
  "repository": {
@@ -12,7 +12,7 @@
12
12
  "url": "https://github.com/Code-Inspect/flowr/issues"
13
13
  },
14
14
  "scripts": {
15
- "main": "npm run build && node dist/src/cli/flowr.js",
15
+ "main": "npm run build:bundle-flowr && node dist/src/cli/flowr.min.js",
16
16
  "publish-library": "cp .npmignore package.json README.md LICENSE dist/src/ && cd dist/src && npm publish --access public",
17
17
  "release": "npx release-it --ci",
18
18
  "stats": "ts-node src/cli/statistics-app.ts",
@@ -24,6 +24,7 @@
24
24
  "export-quads": "ts-node src/cli/export-quads-app.ts",
25
25
  "capabilities-markdown": "ts-node src/r-bridge/data/print.ts",
26
26
  "build": "tsc --project .",
27
+ "build:bundle-flowr": "npm run build && esbuild --bundle dist/src/cli/flowr.js --platform=node --bundle --minify --target=node18 --outfile=dist/src/cli/flowr.min.js",
27
28
  "lint-local": "npx eslint --version && npx eslint src/ test/ --rule \"no-warning-comments: off\"",
28
29
  "lint": "npm run license-compat -- --summary && npx eslint --version && npx eslint src/ test/",
29
30
  "license-compat": "license-checker --onlyAllow 'MIT;MIT OR X11;GPLv2;LGPL;GNUGPL;ISC;Apache-2.0;FreeBSD;BSD-2-Clause;clearbsd;ModifiedBSD;BSD-3-Clause;Python-2.0;Unlicense;WTFPL;BlueOak-1.0.0;CC-BY-4.0;CC-BY-3.0;CC0-1.0;0BSD'",
@@ -200,6 +201,7 @@
200
201
  "@typescript-eslint/eslint-plugin": "^7.8.0",
201
202
  "chai": "^4.3.16",
202
203
  "chai-as-promised": "^7.1.1",
204
+ "esbuild": "0.21.3",
203
205
  "eslint": "^8.57.0",
204
206
  "eslint-plugin-check-file": "^2.8.0",
205
207
  "eslint-plugin-import": "^2.29.1",
@@ -373,7 +373,7 @@ export declare const flowrCapabilities: {
373
373
  readonly name: "Sequencing";
374
374
  readonly id: "built-in-sequencing";
375
375
  readonly supported: "not";
376
- readonly description: "_Handle `:`, `seq`, ... as they are used often._";
376
+ readonly description: "_Handle `:`, `seq`, ... by gathering value information using abstract interpretation._";
377
377
  }, {
378
378
  readonly name: "Internal and Primitive Functions";
379
379
  readonly id: "built-in-internal-and-primitive-functions";
@@ -464,7 +464,7 @@ exports.flowrCapabilities = {
464
464
  name: 'Sequencing',
465
465
  id: 'built-in-sequencing',
466
466
  supported: 'not',
467
- description: '_Handle `:`, `seq`, ... as they are used often._'
467
+ description: '_Handle `:`, `seq`, ... by gathering value information using abstract interpretation._'
468
468
  },
469
469
  {
470
470
  name: 'Internal and Primitive Functions',
@@ -11,7 +11,7 @@ export type RFunctionArgument<Info = NoInfo> = RArgument<Info> | typeof EmptyArg
11
11
  */
12
12
  export interface RNamedFunctionCall<Info = NoInfo> extends Base<Info>, Location {
13
13
  readonly type: RType.FunctionCall;
14
- readonly flavor: 'named';
14
+ readonly named: true;
15
15
  functionName: RSymbol<Info>;
16
16
  /** arguments can be empty, for example when calling as `a(1, ,3)` */
17
17
  readonly arguments: readonly RFunctionArgument<Info>[];
@@ -23,7 +23,7 @@ export interface RNamedFunctionCall<Info = NoInfo> extends Base<Info>, Location
23
23
  */
24
24
  export interface RUnnamedFunctionCall<Info = NoInfo> extends Base<Info>, Location {
25
25
  readonly type: RType.FunctionCall;
26
- readonly flavor: 'unnamed';
26
+ readonly named: false | undefined;
27
27
  calledFunction: RNode<Info>;
28
28
  /** marks function calls like `3 %xx% 4` which have been written in special infix notation; deprecated in v2 */
29
29
  infixSpecial?: boolean;
@@ -41,7 +41,7 @@ exports.OperatorDatabase = {
41
41
  '=': { name: 'equal assignment', stringUsedInRAst: "EQ_ASSIGN" /* RawRType.EqualAssign */, stringUsedInternally: '=', writtenAs: 'infix', arity: 2 /* OperatorArity.Binary */, usedAs: 'assignment', capabilities: ['binary-operator', 'infix-calls', 'assignment-functions', 'local-equal-assignment', 'function-calls'] },
42
42
  /* others */
43
43
  /* maybe introduce custom in-r-ast flavor for these? we consider it arithmetic, as it works on numbers => if we change this we have to create custom tests! (with arithmetic, there is the automatic test set) */
44
- ':': { name: 'sequence operator', stringUsedInRAst: ":" /* RawRType.Colon */, stringUsedInternally: ':', writtenAs: 'infix', arity: 2 /* OperatorArity.Binary */, usedAs: 'operation', capabilities: ['binary-operator', 'infix-calls', 'function-calls', 'built-in-sequencing'] },
44
+ ':': { name: 'sequence operator', stringUsedInRAst: ":" /* RawRType.Colon */, stringUsedInternally: ':', writtenAs: 'infix', arity: 2 /* OperatorArity.Binary */, usedAs: 'operation', capabilities: ['binary-operator', 'infix-calls', 'function-calls'] },
45
45
  '?': { name: 'question', stringUsedInRAst: "?" /* RawRType.Question */, stringUsedInternally: '?', writtenAs: 'prefix', arity: 1 /* OperatorArity.Unary */, usedAs: 'operation', capabilities: ['unary-operator', 'built-in-help'] }
46
46
  };
47
47
  /* eslint-enable */
@@ -269,7 +269,7 @@ function createFoldForFunctionCall(info) {
269
269
  return (data, functionName, args, depth) => {
270
270
  const id = info.getId(data);
271
271
  let decorated;
272
- if (data.flavor === 'named') {
272
+ if (data.named) {
273
273
  decorated = { ...data, info: { ...data.info, id, parent: undefined, depth }, functionName, arguments: args };
274
274
  }
275
275
  else {
@@ -37,7 +37,7 @@ function foldAstStateful(ast, down, folds) {
37
37
  case "RRepeatLoop" /* RType.RepeatLoop */:
38
38
  return folds.loop.foldRepeat(ast, foldAstStateful(ast.body, down, folds), down);
39
39
  case "RFunctionCall" /* RType.FunctionCall */:
40
- return folds.functions.foldFunctionCall(ast, foldAstStateful(ast.flavor === 'named' ? ast.functionName : ast.calledFunction, down, folds), ast.arguments.map(param => param === r_function_call_1.EmptyArgument ? param : foldAstStateful(param, down, folds)), down);
40
+ return folds.functions.foldFunctionCall(ast, foldAstStateful(ast.named ? ast.functionName : ast.calledFunction, down, folds), ast.arguments.map(param => param === r_function_call_1.EmptyArgument ? param : foldAstStateful(param, down, folds)), down);
41
41
  case "RFunctionDefinition" /* RType.FunctionDefinition */:
42
42
  return folds.functions.foldFunctionDefinition(ast, ast.parameters.map(param => foldAstStateful(param, down, folds)), foldAstStateful(ast.body, down, folds), down);
43
43
  case "RParameter" /* RType.Parameter */:
@@ -15,11 +15,11 @@ class NodeVisitor {
15
15
  if (this.onEnter?.(node)) {
16
16
  return;
17
17
  }
18
- /* let the type system know, that the type does not change */
18
+ /* let the type system know that the type does not change */
19
19
  const type = node.type;
20
20
  switch (type) {
21
21
  case "RFunctionCall" /* RType.FunctionCall */:
22
- this.visitSingle(node.flavor === 'named' ? node.functionName : node.calledFunction);
22
+ this.visitSingle(node.named ? node.functionName : node.calledFunction);
23
23
  this.visit(node.arguments);
24
24
  break;
25
25
  case "RFunctionDefinition" /* RType.FunctionDefinition */:
@@ -96,7 +96,7 @@ function tryParseUnnamedFunctionCall(data, mappedWithName, location, content) {
96
96
  }
97
97
  return {
98
98
  type: "RFunctionCall" /* RType.FunctionCall */,
99
- flavor: 'unnamed',
99
+ named: undefined,
100
100
  location,
101
101
  lexeme: content,
102
102
  calledFunction: calledFunction,
@@ -129,7 +129,7 @@ function parseNamedFunctionCall(data, symbolContent, mappedWithName, location, c
129
129
  const parsedArguments = parseArguments(mappedWithName, data);
130
130
  return {
131
131
  type: "RFunctionCall" /* RType.FunctionCall */,
132
- flavor: 'named',
132
+ named: true,
133
133
  location,
134
134
  lexeme: content,
135
135
  functionName,
@@ -37,7 +37,7 @@ function parseBinaryOp(data, lhs, operator, rhs) {
37
37
  // parse as infix function call!
38
38
  return {
39
39
  type: "RFunctionCall" /* RType.FunctionCall */,
40
- flavor: 'named',
40
+ named: true,
41
41
  infixSpecial: true,
42
42
  lexeme: data.currentLexeme ?? content,
43
43
  location,
@@ -49,4 +49,4 @@ export declare function removeRQuotes(str: string): string;
49
49
  /**
50
50
  * Needs to be called *after* {@link retrieveParseDataFromRCode} (or {@link retrieveNormalizedAstFromRCode})
51
51
  */
52
- export declare function retrieveNumberOfRTokensOfLastParse(shell: RShell): Promise<number>;
52
+ export declare function retrieveNumberOfRTokensOfLastParse(shell: RShell, ignoreComments?: boolean): Promise<number>;
@@ -95,8 +95,9 @@ exports.removeRQuotes = removeRQuotes;
95
95
  /**
96
96
  * Needs to be called *after* {@link retrieveParseDataFromRCode} (or {@link retrieveNormalizedAstFromRCode})
97
97
  */
98
- async function retrieveNumberOfRTokensOfLastParse(shell) {
99
- const result = await shell.sendCommandWithOutput(`cat(nrow(flowr_output),${(0, convert_values_1.ts2r)(shell.options.eol)})`);
98
+ async function retrieveNumberOfRTokensOfLastParse(shell, ignoreComments = false) {
99
+ const rows = ignoreComments ? `flowr_output[flowr_output$token != "${"COMMENT" /* RawRType.Comment */}", ]` : 'flowr_output';
100
+ const result = await shell.sendCommandWithOutput(`cat(nrow(${rows}),${(0, convert_values_1.ts2r)(shell.options.eol)})`);
100
101
  (0, assert_1.guard)(result.length === 1, () => `expected exactly one line to obtain the number of R tokens, but got: ${JSON.stringify(result)}`);
101
102
  return Number(result[0]);
102
103
  }
package/r-bridge/shell.js CHANGED
@@ -27,7 +27,7 @@ var __importDefault = (this && this.__importDefault) || function (mod) {
27
27
  };
28
28
  Object.defineProperty(exports, "__esModule", { value: true });
29
29
  exports.RShell = exports.DEFAULT_R_SHELL_OPTIONS = exports.DEFAULT_R_SHELL_EXEC_OPTIONS = exports.DEFAULT_R_PATH = exports.DEFAULT_OUTPUT_COLLECTOR_CONFIGURATION = void 0;
30
- const node_child_process_1 = require("node:child_process");
30
+ const child_process_1 = require("child_process");
31
31
  const objects_1 = require("../util/objects");
32
32
  const readline = __importStar(require("readline"));
33
33
  const log_1 = require("../util/log");
@@ -196,7 +196,8 @@ class RShell {
196
196
  */
197
197
  clearEnvironment() {
198
198
  this.log.debug('clearing environment');
199
- this._sendCommand('rm(list=ls())');
199
+ // run rm(list=ls()) but ignore 'flowr_get_ast', which is the compile command installed
200
+ this._sendCommand('rm(list=setdiff(ls(), "flowr_get_ast"))');
200
201
  }
201
202
  /**
202
203
  * Obtain the temporary directory used by R.
@@ -231,7 +232,7 @@ class RShellSession {
231
232
  options;
232
233
  collectionTimeout;
233
234
  constructor(options, log) {
234
- this.bareSession = (0, node_child_process_1.spawn)(options.pathToRExecutable, options.commandLineOptions, {
235
+ this.bareSession = (0, child_process_1.spawn)(options.pathToRExecutable, options.commandLineOptions, {
235
236
  env: options.env,
236
237
  cwd: options.cwd,
237
238
  windowsHide: true
@@ -14,8 +14,8 @@ export declare function doNotAutoSelect(_node: RNode<ParentInformation>): boolea
14
14
  export declare function autoSelectLibrary(node: RNode<ParentInformation>): boolean;
15
15
  export interface ReconstructionResult {
16
16
  code: string;
17
- /** number of nodes that triggered the `autoSelectIf` predicate {@link reconstructToCode} */
18
- autoSelected: number;
17
+ /** number of lines that contain nodes that triggered the `autoSelectIf` predicate {@link reconstructToCode} */
18
+ linesWithAutoSelected: number;
19
19
  }
20
20
  /**
21
21
  * Reconstructs parts of a normalized R ast into R code on an expression basis.
@@ -24,7 +24,7 @@ export interface ReconstructionResult {
24
24
  * @param selection - The selection of nodes to be reconstructed (probably the {@link NodeId|NodeIds} identified by the slicer)
25
25
  * @param autoSelectIf - A predicate that can be used to force the reconstruction of a node (for example to reconstruct library call statements, see {@link autoSelectLibrary}, {@link doNotAutoSelect})
26
26
  *
27
- * @returns The number of times `autoSelectIf` triggered, as well as the reconstructed code itself.
27
+ * @returns The number of lines for which `autoSelectIf` triggered, as well as the reconstructed code itself.
28
28
  */
29
29
  export declare function reconstructToCode<Info>(ast: NormalizedAst<Info>, selection: Selection, autoSelectIf?: AutoSelectPredicate): ReconstructionResult;
30
30
  export {};
@@ -129,31 +129,36 @@ function reconstructForLoop(loop, variable, vector, body, config) {
129
129
  ];
130
130
  }
131
131
  }
132
- function reconstructRepeatLoop(loop, body, configuration) {
133
- const sel = isSelected(configuration, loop);
134
- if (!sel) {
135
- return body;
132
+ function reconstructBodyWithHeader(header, body, onEmpty) {
133
+ if (body.length === 0) {
134
+ return [{ line: `${header.line}${onEmpty}`, indent: header.indent }];
136
135
  }
137
- else if (body.length <= 1) {
138
- // 'inline'
139
- return [{ line: `repeat ${body.length === 0 ? '{}' : body[0].line}`, indent: 0 }];
136
+ else if (body.length === 1) {
137
+ return [
138
+ { line: `${header.line} ${body[0].line}`, indent: header.indent }
139
+ ];
140
140
  }
141
141
  else if (body[0].line === '{' && body[body.length - 1].line === '}') {
142
- // 'block'
143
142
  return [
144
- { line: 'repeat {', indent: 0 },
143
+ { line: `${header.line} {`, indent: header.indent },
145
144
  ...body.slice(1, body.length - 1),
146
- { line: '}', indent: 0 }
145
+ { line: '}', indent: header.indent }
147
146
  ];
148
147
  }
149
148
  else {
150
- // unknown
151
149
  return [
152
- { line: 'repeat', indent: 0 },
150
+ header,
153
151
  ...indentBy(body, 1)
154
152
  ];
155
153
  }
156
154
  }
155
+ function reconstructRepeatLoop(loop, body, configuration) {
156
+ const sel = isSelected(configuration, loop);
157
+ if (!sel) {
158
+ return body;
159
+ }
160
+ return reconstructBodyWithHeader({ line: 'repeat', indent: 0 }, body, '{}');
161
+ }
157
162
  function reconstructIfThenElse(ifThenElse, condition, then, otherwise, config) {
158
163
  otherwise ??= [];
159
164
  if (then.length === 0 && otherwise.length === 0) {
@@ -169,10 +174,7 @@ function reconstructIfThenElse(ifThenElse, condition, then, otherwise, config) {
169
174
  }
170
175
  else if (otherwise.length === 0) {
171
176
  if (isSelected(config, ifThenElse)) {
172
- return [
173
- { line: `if(${getLexeme(ifThenElse.condition)}) ${then[0].line}`, indent: 0 },
174
- ...indentBy(then.splice(1), 1)
175
- ];
177
+ return reconstructBodyWithHeader({ line: `if(${getLexeme(ifThenElse.condition)})`, indent: 0 }, then, '{}');
176
178
  }
177
179
  else {
178
180
  return then;
@@ -180,10 +182,7 @@ function reconstructIfThenElse(ifThenElse, condition, then, otherwise, config) {
180
182
  }
181
183
  else if (then.length === 0) {
182
184
  if (isSelected(config, ifThenElse)) {
183
- return [
184
- { line: `if(${getLexeme(ifThenElse.condition)}) { } else ${otherwise[0].line}`, indent: 0 },
185
- ...indentBy(otherwise.splice(1), 1)
186
- ];
185
+ return reconstructBodyWithHeader({ line: `if(${getLexeme(ifThenElse.condition)}) { } else`, indent: 0 }, then, '{}');
187
186
  }
188
187
  else {
189
188
  return otherwise;
@@ -283,7 +282,7 @@ function reconstructFunctionDefinition(definition, functionParameters, body, con
283
282
  const empty = body === undefined || body.length === 0;
284
283
  const selected = isSelected(config, definition);
285
284
  if (empty && selected) { // give function stub
286
- return plain(`function(${reconstructParameters(definition.parameters).join(', ')}) { }`);
285
+ return plain(`${definition.lexeme}(${reconstructParameters(definition.parameters).join(', ')}) { }`);
287
286
  }
288
287
  else if (!selected) { // do not require function
289
288
  return body;
@@ -294,25 +293,25 @@ function reconstructFunctionDefinition(definition, functionParameters, body, con
294
293
  // 'inline'
295
294
  const bodyStr = body.length === 0 ? '{ }' : `${body[0].line}`;
296
295
  // we keep the braces in every case because I do not like no-brace functions
297
- return [{ line: `function(${parameters}) ${bodyStr}`, indent: 0 }];
296
+ return [{ line: `${definition.lexeme}(${parameters}) ${bodyStr}`, indent: 0 }];
298
297
  }
299
298
  else {
300
299
  // 'block'
301
300
  return [
302
- { line: `function(${parameters}) ${body[0].line}`, indent: 0 },
301
+ { line: `${definition.lexeme}(${parameters}) ${body[0].line}`, indent: 0 },
303
302
  ...body.slice(1),
304
303
  ];
305
304
  }
306
305
  }
307
306
  function reconstructSpecialInfixFunctionCall(args, call) {
308
307
  (0, assert_1.guard)(args.length === 2, () => `infix special call must have exactly two arguments, got: ${args.length} (${JSON.stringify(args)})`);
309
- (0, assert_1.guard)(call.flavor === 'named', `infix special call must be named, got: ${call.flavor}`);
308
+ (0, assert_1.guard)(call.named, `infix special call must be named, got: ${call.named}`);
310
309
  const [lhs, rhs] = args;
311
310
  if ((lhs === undefined || lhs.length === 0) && (rhs === undefined || rhs.length === 0)) {
312
311
  return [];
313
312
  }
314
313
  // else if (rhs === undefined || rhs.length === 0) {
315
- // if rhs is undefined we still have to keep both now, but reconstruct manually :/
314
+ // if rhs is undefined we still have to keep both now, but reconstruct manually :/
316
315
  if (lhs !== r_function_call_1.EmptyArgument && lhs.length > 0) {
317
316
  const lhsText = lhs.map(l => `${getIndentString(l.indent)}${l.line}`).join('\n');
318
317
  if (rhs !== r_function_call_1.EmptyArgument && rhs.length > 0) {
@@ -339,7 +338,7 @@ function reconstructFunctionCall(call, functionName, args, configuration) {
339
338
  if (call.infixSpecial === true) {
340
339
  return reconstructSpecialInfixFunctionCall(args, call);
341
340
  }
342
- if (call.flavor === 'named' && selected) {
341
+ if (call.named && selected) {
343
342
  return plain(getLexeme(call));
344
343
  }
345
344
  const filteredArgs = args.filter(a => a !== undefined && a.length > 0);
@@ -349,14 +348,12 @@ function reconstructFunctionCall(call, functionName, args, configuration) {
349
348
  if (args.length === 0) {
350
349
  (0, assert_1.guard)(functionName.length > 0, `without args, we need the function name to be present! got: ${JSON.stringify(functionName)}`);
351
350
  const last = functionName[functionName.length - 1];
352
- if (call.flavor === 'unnamed' && !last.line.endsWith(')')) {
351
+ if (!call.named && !last.line.endsWith(')')) {
353
352
  functionName[0].line = `(${functionName[0].line}`;
354
353
  last.line += ')';
355
354
  }
356
- if (!last.line.endsWith('()')) {
357
- // add empty call braces if not present
358
- last.line += '()';
359
- }
355
+ // add empty call braces if not present
356
+ last.line += '()';
360
357
  return functionName;
361
358
  }
362
359
  else {
@@ -369,7 +366,7 @@ function doNotAutoSelect(_node) {
369
366
  exports.doNotAutoSelect = doNotAutoSelect;
370
367
  const libraryFunctionCall = /^(library|require|((require|load|attach)Namespace))$/;
371
368
  function autoSelectLibrary(node) {
372
- if (node.type !== "RFunctionCall" /* RType.FunctionCall */ || node.flavor !== 'named') {
369
+ if (node.type !== "RFunctionCall" /* RType.FunctionCall */ || !node.named) {
373
370
  return false;
374
371
  }
375
372
  return libraryFunctionCall.test(node.functionName.content);
@@ -416,13 +413,13 @@ function getIndentString(indent) {
416
413
  function prettyPrintCodeToString(code, lf = '\n') {
417
414
  return code.map(({ line, indent }) => `${getIndentString(indent)}${line}`).join(lf);
418
415
  }
419
- function removeOuterExpressionListIfApplicable(result, autoSelected) {
416
+ function removeOuterExpressionListIfApplicable(result, linesWithAutoSelected) {
420
417
  if (result.length > 1 && result[0].line === '{' && result[result.length - 1].line === '}') {
421
418
  // remove outer block
422
- return { code: prettyPrintCodeToString(indentBy(result.slice(1, result.length - 1), -1)), autoSelected };
419
+ return { code: prettyPrintCodeToString(indentBy(result.slice(1, result.length - 1), -1)), linesWithAutoSelected };
423
420
  }
424
421
  else {
425
- return { code: prettyPrintCodeToString(result), autoSelected };
422
+ return { code: prettyPrintCodeToString(result), linesWithAutoSelected };
426
423
  }
427
424
  }
428
425
  /**
@@ -432,25 +429,27 @@ function removeOuterExpressionListIfApplicable(result, autoSelected) {
432
429
  * @param selection - The selection of nodes to be reconstructed (probably the {@link NodeId|NodeIds} identified by the slicer)
433
430
  * @param autoSelectIf - A predicate that can be used to force the reconstruction of a node (for example to reconstruct library call statements, see {@link autoSelectLibrary}, {@link doNotAutoSelect})
434
431
  *
435
- * @returns The number of times `autoSelectIf` triggered, as well as the reconstructed code itself.
432
+ * @returns The number of lines for which `autoSelectIf` triggered, as well as the reconstructed code itself.
436
433
  */
437
434
  function reconstructToCode(ast, selection, autoSelectIf = autoSelectLibrary) {
438
435
  if (exports.reconstructLogger.settings.minLevel <= 1 /* LogLevel.Trace */) {
439
436
  exports.reconstructLogger.trace(`reconstruct ast with ids: ${JSON.stringify([...selection])}`);
440
437
  }
441
- // we use a wrapper to count the number of times the autoSelectIf predicate triggered
442
- let autoSelected = 0;
438
+ // we use a wrapper to count the number of lines for which the autoSelectIf predicate triggered
439
+ const linesWithAutoSelected = new Set();
443
440
  const autoSelectIfWrapper = (node) => {
444
441
  const result = autoSelectIf(node);
445
- if (result) {
446
- autoSelected++;
442
+ if (result && node.location) {
443
+ for (let i = node.location[0]; i <= node.location[2]; i++) {
444
+ linesWithAutoSelected.add(i);
445
+ }
447
446
  }
448
447
  return result;
449
448
  };
450
449
  // fold of the normalized ast
451
450
  const result = (0, stateful_fold_1.foldAstStateful)(ast.ast, { selection, autoSelectIf: autoSelectIfWrapper }, reconstructAstFolds);
452
451
  (0, log_1.expensiveTrace)(exports.reconstructLogger, () => `reconstructed ast before string conversion: ${JSON.stringify(result)}`);
453
- return removeOuterExpressionListIfApplicable(result, autoSelected);
452
+ return removeOuterExpressionListIfApplicable(result, linesWithAutoSelected.size);
454
453
  }
455
454
  exports.reconstructToCode = reconstructToCode;
456
455
  //# sourceMappingURL=reconstruct.js.map
@@ -38,7 +38,7 @@ const defaultAllVariablesCollectorFolds = {
38
38
  foldFunctionDefinition: (_, a, b) => [...a.flat(), ...b],
39
39
  foldFunctionCall: (c, a, b) => {
40
40
  const args = b.flatMap(b => b !== r_function_call_1.EmptyArgument ? b.flat() : []);
41
- if (c.flavor === 'named') {
41
+ if (c.named) {
42
42
  return c.functionName.content === 'library' ? args.slice(1) : args;
43
43
  }
44
44
  else {
@@ -48,8 +48,8 @@ function staticSlicing(graph, ast, criteria, threshold = 75) {
48
48
  if (currentVertex.controlDependencies) {
49
49
  const topLevel = graph.isRoot(id) || sliceSeedIds.has(id);
50
50
  for (const cd of currentVertex.controlDependencies) {
51
- if (!topLevel || (ast.idMap.get(cd)?.info.depth ?? 0) <= minDepth) {
52
- queue.add(cd, baseEnvironment, baseEnvFingerprint, false);
51
+ if (!topLevel || (ast.idMap.get(cd.id)?.info.depth ?? 0) <= minDepth) {
52
+ queue.add(cd.id, baseEnvironment, baseEnvFingerprint, false);
53
53
  }
54
54
  }
55
55
  }
@@ -79,7 +79,7 @@ function updateCommonSyntaxTypeCounts(current, ...nodes) {
79
79
  }
80
80
  break;
81
81
  case "RFunctionCall" /* RType.FunctionCall */:
82
- if (node.flavor === 'unnamed') {
82
+ if (!node.named) {
83
83
  current.unnamedCall++;
84
84
  }
85
85
  else {
@@ -21,7 +21,7 @@ function visitIfThenElse(info, input) {
21
21
  const ifThenElseStack = [];
22
22
  (0, visitor_1.visitAst)(input.normalizedRAst.ast, node => {
23
23
  if (node.type !== "RIfThenElse" /* RType.IfThenElse */) {
24
- if (node.type === "RFunctionCall" /* RType.FunctionCall */ && node.flavor === 'named' && node.functionName.content === 'switch') {
24
+ if (node.type === "RFunctionCall" /* RType.FunctionCall */ && node.named && node.functionName.content === 'switch') {
25
25
  const initialArg = (0, unpack_argument_1.unpackArgument)(node.arguments[0]);
26
26
  if (initialArg) {
27
27
  info.switchCase = (0, common_syntax_probability_1.updateCommonSyntaxTypeCounts)(info.switchCase, initialArg);
@@ -100,7 +100,7 @@ function visitDefinitions(info, input) {
100
100
  // track all calls with the same name that do not already have a bound calls edge, superfluous if recursive tracking is explicit
101
101
  const recursiveCalls = [];
102
102
  (0, visitor_1.visitAst)(node.body, n => {
103
- if (n.type === "RFunctionCall" /* RType.FunctionCall */ && n.flavor === 'named' && assigned.has(n.functionName.lexeme)) {
103
+ if (n.type === "RFunctionCall" /* RType.FunctionCall */ && n.named && assigned.has(n.functionName.lexeme)) {
104
104
  recursiveCalls.push(n);
105
105
  }
106
106
  });
@@ -33,7 +33,7 @@ function visitLoops(info, input) {
33
33
  info.breakStatements++;
34
34
  return;
35
35
  case "RFunctionCall" /* RType.FunctionCall */:
36
- if (node.flavor === 'named' && isImplicitLoop.test(node.functionName.lexeme)) {
36
+ if (node.named && isImplicitLoop.test(node.functionName.lexeme)) {
37
37
  info.implicitLoops++;
38
38
  (0, statistics_file_1.appendStatisticsFile)(exports.loops.name, 'implicit-loop', [node.functionName.info.fullLexeme ?? node.functionName.lexeme], input.filepath);
39
39
  }