@eagleoutice/flowr 2.6.2 → 2.7.0

package/abstract-interpretation/data-frame/mappers/function-mapper.js

@@ -519,16 +519,16 @@ const DataFrameFunctionParamsMapper = {
 };
 /**
  * Maps a concrete data frame function call to abstract data frame operations.
- * @param node   - The R node of the function call
- * @param dfg    - The data flow graph for resolving the arguments
- * @param config - The flowR configuration to use
+ * @param node - The R node of the function call
+ * @param dfg  - The data flow graph for resolving the arguments
+ * @param ctx  - The current flowR analyzer context
  * @returns Data frame expression info containing the mapped abstract data frame operations, or `undefined` if the node does not represent a data frame function call
  */
-function mapDataFrameFunctionCall(node, dfg, config) {
+function mapDataFrameFunctionCall(node, dfg, ctx) {
     if (node.type !== type_1.RType.FunctionCall || !node.named) {
         return;
     }
-    const resolveInfo = { graph: dfg, idMap: dfg.idMap, full: true, resolve: config_1.VariableResolve.Alias };
+    const resolveInfo = { graph: dfg, idMap: dfg.idMap, full: true, resolve: config_1.VariableResolve.Alias, ctx };
     let operations;
     if (isDataFrameFunction(node.functionName.content)) {
         const functionName = node.functionName.content;
@@ -539,7 +539,7 @@ function mapDataFrameFunctionCall(node, dfg, config) {
             operations = [{ operation: 'unknown', operand: undefined }];
         }
         else {
-            operations = mapper(args, params, resolveInfo, config);
+            operations = mapper(args, params, resolveInfo, ctx);
         }
     }
     else {
@@ -605,10 +605,10 @@ function mapDataFrameConvert(args, params, info) {
             operand: dataFrame.value.info.id
         }];
 }
-function mapDataFrameRead(args, params, info, config) {
+function mapDataFrameRead(args, params, info, ctx) {
     const fileNameArg = (0, arguments_1.getFunctionArgument)(args, params.fileName, info);
     const textArg = params.text ? (0, arguments_1.getFunctionArgument)(args, params.text, info) : undefined;
-    const { source, request } = getRequestFromRead(fileNameArg, textArg, params, info, config);
+    const { source, request } = getRequestFromRead(fileNameArg, textArg, params, info, ctx);
     const header = (0, arguments_1.getArgumentValue)(args, params.header, info);
     const separator = (0, arguments_1.getArgumentValue)(args, params.separator, info);
     const quote = (0, arguments_1.getArgumentValue)(args, params.quote, info);
@@ -618,7 +618,7 @@ function mapDataFrameRead(args, params, info, config) {
     const noDupNames = (0, arguments_1.getArgumentValue)(args, params.noDupNames, info);
     const validArguments = typeof header === 'boolean' && typeof separator === 'string' && typeof quote === 'string' && typeof comment === 'string' &&
         typeof skipLines === 'number' && typeof checkNames === 'boolean' && typeof noDupNames === 'boolean';
-    if (request === undefined || !config.abstractInterpretation.dataFrame.readLoadedData.readExternalFiles || !validArguments) {
+    if (request === undefined || !ctx.config.abstractInterpretation.dataFrame.readLoadedData.readExternalFiles || !validArguments) {
         return [{
                 operation: 'read',
                 operand: undefined,
@@ -643,7 +643,7 @@ function mapDataFrameRead(args, params, info, config) {
             }
         }
     };
-    const allLines = parseRequestContent(request, parseLine, config.abstractInterpretation.dataFrame.readLoadedData.maxReadLines);
+    const allLines = parseRequestContent(request, parseLine, ctx.config.abstractInterpretation.dataFrame.readLoadedData.maxReadLines);
     let colnames;
     if (header) {
         colnames = (0, arguments_1.filterValidNames)(firstLine, checkNames, noDupNames, params.noEmptyNames);
@@ -946,20 +946,20 @@ function mapDataFrameMutate(args, params, info) {
             columns: accessedNames
         });
     }
-    if (deletedCols === undefined || deletedCols.length > 0) {
+    if (mutatedCols === undefined || mutatedCols.length > 0 || deletedCols?.length === 0) {
         result.push({
-            operation: 'removeCols',
+            operation: 'mutateCols',
             operand: operand?.info.id,
-            colnames: deletedCols,
-            options: { maybe: true }
+            colnames: mutatedCols
         });
         operand = undefined;
     }
-    if (mutatedCols === undefined || mutatedCols.length > 0 || deletedCols?.length === 0) {
+    if (deletedCols === undefined || deletedCols.length > 0) {
         result.push({
-            operation: 'mutateCols',
+            operation: 'removeCols',
             operand: operand?.info.id,
-            colnames: mutatedCols
+            colnames: deletedCols,
+            options: { maybe: true }
         });
         operand = undefined;
     }
@@ -1002,8 +1002,11 @@ function mapDataFrameSummarize(args, params, info) {
     }
     const result = [];
     const summarizeArgs = args.filter(arg => arg !== dataFrame);
-    const accessedNames = summarizeArgs.flatMap(arg => (0, arguments_1.getUnresolvedSymbolsInExpression)(arg, info.graph));
     const summarizedCols = summarizeArgs.map(arg => (0, resolve_args_1.resolveIdToArgName)(arg, info));
+    // only column names that are not created by summarize are preconditions on the operand
+    const accessedNames = summarizeArgs
+        .flatMap(arg => (0, arguments_1.getUnresolvedSymbolsInExpression)(arg, info.graph))
+        .filter(arg => !summarizedCols.includes(arg));
     if (accessedNames.length > 0) {
         result.push({
             operation: 'accessCols',
@@ -1018,7 +1021,7 @@ function mapDataFrameSummarize(args, params, info) {
     });
     return result;
 }
-function mapDataFrameJoin(args, params, info, config) {
+function mapDataFrameJoin(args, params, info, ctx) {
     const dataFrame = (0, arguments_1.getFunctionArgument)(args, params.dataFrame, info);
     const joinAll = (0, arguments_1.getArgumentValue)(args, params.joinAll, info);
     const joinLeft = (0, arguments_1.getArgumentValue)(args, params.joinLeft, info);
@@ -1035,7 +1038,7 @@ function mapDataFrameJoin(args, params, info, config) {
     const result = [];
     const otherArg = (0, arguments_1.getFunctionArgument)(args, params.otherDataFrame, info);
     const byArg = (0, arguments_1.getFunctionArgument)(args, params.by, info);
-    const otherDataFrame = (0, shape_inference_1.resolveIdToDataFrameShape)(otherArg, info.graph) ?? dataframe_domain_1.DataFrameDomain.top(config.abstractInterpretation.dataFrame.maxColNames);
+    const otherDataFrame = (0, shape_inference_1.resolveIdToDataFrameShape)(otherArg, info.graph) ?? dataframe_domain_1.DataFrameDomain.top(ctx.config.abstractInterpretation.dataFrame.maxColNames);
     let byCols;
     const joinType = getJoinType(joinAll, joinLeft, joinRight);
     if (byArg !== undefined) {
@@ -1101,19 +1104,19 @@ function mapDataFrameUnknown(args, params, info) {
             ...(params.constraintType !== undefined ? { type: params.constraintType } : {})
         }];
 }
-function getRequestFromRead(fileNameArg, textArg, params, info, config) {
+function getRequestFromRead(fileNameArg, textArg, params, info, ctx) {
     let source;
     let request;
     if (fileNameArg !== undefined && fileNameArg !== r_function_call_1.EmptyArgument) {
         const fileName = (0, resolve_args_1.resolveIdToArgValue)(fileNameArg, info);
         if (typeof fileName === 'string') {
             source = fileName;
-            const referenceChain = fileNameArg.info.file ? [(0, retriever_1.requestFromInput)(`file://${fileNameArg.info.file}`)] : [];
-            const sources = (0, built_in_source_1.findSource)(config.solver.resolveSource, fileName, { referenceChain });
+            const referenceChain = fileNameArg.info.file ? [fileNameArg.info.file] : [];
+            const sources = (0, built_in_source_1.findSource)(ctx.config.solver.resolveSource, fileName, { referenceChain, ctx });
             if (sources?.length === 1) {
                 source = sources[0];
                 // create request from resolved source file path
-                request = (0, built_in_source_1.getSourceProvider)().createRequest(source);
+                request = { request: 'file', content: sources[0] };
             }
             else if (params.text === undefined && (0, resolve_args_1.unescapeSpecialChars)(fileName).includes('\n')) {
                 // create request from string if file name argument contains newline
@@ -1128,6 +1131,7 @@ function getRequestFromRead(fileNameArg, textArg, params, info, config) {
             request = (0, retriever_1.requestFromInput)((0, resolve_args_1.unescapeSpecialChars)(text));
         }
     }
+    request = request ? ctx.files.resolveRequest(request).r : undefined;
     return { source, request };
 }
 function parseRequestContent(request, parser, maxLines) {

package/abstract-interpretation/data-frame/mappers/replacement-mapper.d.ts

@@ -2,10 +2,12 @@ import type { DataflowGraph } from '../../../dataflow/graph/graph';
 import type { RNode } from '../../../r-bridge/lang-4.x/ast/model/model';
 import type { ParentInformation } from '../../../r-bridge/lang-4.x/ast/model/processing/decorate';
 import type { DataFrameExpressionInfo } from '../absint-info';
+import type { ReadOnlyFlowrAnalyzerContext } from '../../../project/context/flowr-analyzer-context';
 /**
  * Maps a concrete data frame replacement function to abstract data frame operations.
  * @param node - The R node of the replacement function
  * @param dfg  - The data flow graph for resolving the arguments
+ * @param ctx - The read-only Flowr analysis context
  * @returns Data frame expression info containing the mapped abstract data frame operations, or `undefined` if the node does not represent a data frame replacement function
  */
-export declare function mapDataFrameReplacementFunction(node: RNode<ParentInformation>, expression: RNode<ParentInformation>, dfg: DataflowGraph): DataFrameExpressionInfo | undefined;
+export declare function mapDataFrameReplacementFunction(node: RNode<ParentInformation>, expression: RNode<ParentInformation>, dfg: DataflowGraph, ctx: ReadOnlyFlowrAnalyzerContext): DataFrameExpressionInfo | undefined;

package/abstract-interpretation/data-frame/mappers/replacement-mapper.js

@@ -21,11 +21,12 @@ const DataFrameReplacementFunctionMapper = {
  * Maps a concrete data frame replacement function to abstract data frame operations.
  * @param node - The R node of the replacement function
  * @param dfg  - The data flow graph for resolving the arguments
+ * @param ctx - The read-only Flowr analysis context
  * @returns Data frame expression info containing the mapped abstract data frame operations, or `undefined` if the node does not represent a data frame replacement function
  */
-function mapDataFrameReplacementFunction(node, expression, dfg) {
+function mapDataFrameReplacementFunction(node, expression, dfg, ctx) {
     const parent = hasParentReplacement(node, dfg) ? dfg.idMap?.get(node.info.parent) : undefined;
-    const resolveInfo = { graph: dfg, idMap: dfg.idMap, full: true, resolve: config_1.VariableResolve.Alias };
+    const resolveInfo = { graph: dfg, idMap: dfg.idMap, full: true, resolve: config_1.VariableResolve.Alias, ctx };
     let operations;
     if (node.type === type_1.RType.Access) {
         if (node.access.every(arg => arg === r_function_call_1.EmptyArgument)) {

package/abstract-interpretation/data-frame/semantics.js

@@ -70,7 +70,7 @@ function getConstraintType(operation) {
     return DataFrameSemanticsMapper[operation].type;
 }
 function applyCreateSemantics(value, { colnames, rows }) {
-    const colnamesValue = colnames?.every(assert_1.isNotUndefined) ? colnames : lattice_1.Top;
+    const colnamesValue = setRange(colnames);
     const colsValue = colnames !== undefined ? [colnames.length, colnames.length] : positive_interval_domain_1.PosIntervalTop;
     const rowsValue = Array.isArray(rows) ? rows : typeof rows === 'number' ? [rows, rows] : positive_interval_domain_1.PosIntervalTop;
     return new dataframe_domain_1.DataFrameDomain({
@@ -85,7 +85,7 @@ function applyReadSemantics(value, { colnames, rows }) {
 function applyAccessColsSemantics(value, { columns }) {
     if (columns?.every(col => typeof col === 'string')) {
         return new dataframe_domain_1.DataFrameDomain({
-            colnames: value.colnames.join(columns),
+            colnames: value.colnames.union(setRange(columns)),
             cols: value.cols,
             rows: value.rows
         });
@@ -93,7 +93,7 @@ function applyAccessColsSemantics(value, { columns }) {
     else if (columns?.every(col => typeof col === 'number')) {
         return new dataframe_domain_1.DataFrameDomain({
             colnames: value.colnames,
-            cols: columns?.reduce((current, col) => current.max([col, col]), value.cols),
+            cols: columns.reduce((current, col) => current.max([col, col]), value.cols),
             rows: value.rows
         });
     }
@@ -113,21 +113,21 @@ function applyAssignColsSemantics(value, { columns }) {
     if (columns?.every(col => typeof col === 'string')) {
         const cols = columns.length;
         return new dataframe_domain_1.DataFrameDomain({
-            colnames: value.colnames.join(columns),
+            colnames: value.colnames.union(setRange(columns)),
             cols: value.cols.add([0, cols]).max([cols, cols]),
             rows: value.rows
         });
     }
     else if (columns?.every(col => typeof col === 'number')) {
         return new dataframe_domain_1.DataFrameDomain({
-            colnames: value.colnames.top(),
+            colnames: value.colnames.widenUp(),
             cols: columns.reduce((current, col) => current.max([col, col]), value.cols),
             rows: value.rows
         });
     }
     return new dataframe_domain_1.DataFrameDomain({
-        colnames: value.colnames.top(),
-        cols: value.cols.extendUp(),
+        colnames: value.colnames.widenUp(),
+        cols: value.cols.widenUp(),
         rows: value.rows
     });
 }
@@ -142,49 +142,57 @@ function applyAssignRowsSemantics(value, { rows }) {
     return new dataframe_domain_1.DataFrameDomain({
         colnames: value.colnames,
         cols: value.cols,
-        rows: value.rows.extendUp()
+        rows: value.rows.widenUp()
     });
 }
 function applySetColNamesSemantics(value, { colnames }, options) {
     if (options?.partial) {
         return new dataframe_domain_1.DataFrameDomain({
-            colnames: colnames?.every(assert_1.isNotUndefined) ? value.colnames.join(colnames) : value.colnames.top(),
+            colnames: value.colnames.widenDown().union(setRange(colnames)),
             cols: value.cols,
             rows: value.rows
         });
     }
     const allColNames = colnames?.every(assert_1.isNotUndefined) && value.cols.value !== lattice_1.Bottom && colnames.length >= value.cols.value[1];
     return new dataframe_domain_1.DataFrameDomain({
-        colnames: allColNames ? value.colnames.create(colnames) : value.colnames.top(),
+        colnames: allColNames ? value.colnames.create(setRange(colnames)) : value.colnames.widenDown().union(setRange(colnames)).widenUp(),
         cols: value.cols,
         rows: value.rows
     });
 }
 function applyAddColsSemantics(value, { colnames }) {
     return new dataframe_domain_1.DataFrameDomain({
-        colnames: colnames?.every(assert_1.isNotUndefined) ? value.colnames.join(colnames) : value.colnames.top(),
-        cols: colnames !== undefined ? value.cols.add([colnames.length, colnames.length]) : value.cols.extendUp(),
+        colnames: colnames !== undefined ? value.colnames.union(setRange(colnames)) : value.colnames.widenUp(),
+        cols: colnames !== undefined ? value.cols.add([colnames.length, colnames.length]) : value.cols.widenUp(),
         rows: value.rows
     });
 }
 function applyAddRowsSemantics(value, { rows }) {
+    if (value.cols.value !== lattice_1.Bottom && value.cols.value[0] === 0) {
+        return new dataframe_domain_1.DataFrameDomain({
+            ...value,
+            colnames: value.colnames.top(),
+            cols: rows !== undefined ? value.cols.add([1, 1]) : value.cols.top(),
+            rows: rows !== undefined ? value.rows.add([rows, rows]) : value.rows.widenUp()
+        });
+    }
     return new dataframe_domain_1.DataFrameDomain({
         colnames: value.colnames,
         cols: value.cols,
-        rows: rows !== undefined ? value.rows.add([rows, rows]) : value.rows.extendUp()
+        rows: rows !== undefined ? value.rows.add([rows, rows]) : value.rows.widenUp()
     });
 }
 function applyRemoveColsSemantics(value, { colnames }, options) {
     if (options?.maybe) {
         return new dataframe_domain_1.DataFrameDomain({
-            colnames: colnames !== undefined ? value.colnames.subtract(colnames.filter(assert_1.isNotUndefined)) : value.colnames,
-            cols: colnames !== undefined ? value.cols.subtract([colnames.length, 0]) : value.cols.extendDown(),
+            colnames: colnames !== undefined ? value.colnames.subtract(setRange(colnames)) : value.colnames.widenDown(),
+            cols: colnames !== undefined ? value.cols.subtract([colnames.length, 0]) : value.cols.widenDown(),
             rows: value.rows
         });
     }
     return new dataframe_domain_1.DataFrameDomain({
-        colnames: colnames !== undefined ? value.colnames.subtract(colnames.filter(assert_1.isNotUndefined)) : value.colnames,
-        cols: colnames !== undefined ? value.cols.subtract([colnames.length, colnames.length]) : value.cols.extendDown(),
+        colnames: colnames !== undefined ? value.colnames.subtract(setRange(colnames)) : value.colnames.widenDown(),
+        cols: colnames !== undefined ? value.cols.subtract([colnames.length, colnames.length]) : value.cols.widenDown(),
         rows: value.rows
     });
 }
@@ -193,18 +201,18 @@ function applyRemoveRowsSemantics(value, { rows }, options) {
         return new dataframe_domain_1.DataFrameDomain({
             colnames: value.colnames,
             cols: value.cols,
-            rows: rows !== undefined ? value.rows.subtract([rows, 0]) : value.rows.extendDown()
+            rows: rows !== undefined ? value.rows.subtract([rows, 0]) : value.rows.widenDown()
         });
     }
     return new dataframe_domain_1.DataFrameDomain({
         colnames: value.colnames,
         cols: value.cols,
-        rows: rows !== undefined ? value.rows.subtract([rows, rows]) : value.rows.extendDown()
+        rows: rows !== undefined ? value.rows.subtract([rows, rows]) : value.rows.widenDown()
     });
 }
 function applyConcatColsSemantics(value, { other }) {
     return new dataframe_domain_1.DataFrameDomain({
-        colnames: value.colnames.join(other.colnames),
+        colnames: value.colnames.union(other.colnames),
         cols: value.cols.add(other.cols),
         rows: value.rows
     });
@@ -235,13 +243,13 @@ function applySubsetColsSemantics(value, { colnames }, options) {
     else if (options?.renamedCols) {
         return new dataframe_domain_1.DataFrameDomain({
             colnames: value.colnames.top(),
-            cols: colnames !== undefined ? value.cols.min([colnames.length, colnames.length]) : value.cols.extendDown(),
+            cols: colnames !== undefined ? value.cols.min([colnames.length, colnames.length]) : value.cols.widenDown(),
             rows: value.rows
         });
     }
     return new dataframe_domain_1.DataFrameDomain({
-        colnames: colnames?.every(assert_1.isNotUndefined) ? value.colnames.meet(colnames) : value.colnames,
-        cols: colnames !== undefined ? value.cols.min([colnames.length, colnames.length]) : value.cols.extendDown(),
+        colnames: colnames !== undefined ? value.colnames.intersect(setRange(colnames)) : value.colnames.widenDown(),
+        cols: colnames !== undefined ? value.cols.min([colnames.length, colnames.length]) : value.cols.widenDown(),
         rows: value.rows
     });
 }
@@ -256,27 +264,27 @@ function applySubsetRowsSemantics(value, { rows }, options) {
     return new dataframe_domain_1.DataFrameDomain({
         colnames: value.colnames,
         cols: value.cols,
-        rows: rows !== undefined ? value.rows.min([rows, rows]) : value.rows.extendDown()
+        rows: rows !== undefined ? value.rows.min([rows, rows]) : value.rows.widenDown()
     });
 }
 function applyFilterRowsSemantics(value, { condition }) {
     return new dataframe_domain_1.DataFrameDomain({
         colnames: value.colnames,
         cols: value.cols,
-        rows: condition ? value.rows : condition === false ? value.rows.create([0, 0]) : value.rows.extendDown()
+        rows: condition ? value.rows : condition === false ? value.rows.create([0, 0]) : value.rows.widenDown()
     });
 }
 function applyMutateColsSemantics(value, { colnames }) {
     return new dataframe_domain_1.DataFrameDomain({
-        colnames: colnames?.every(assert_1.isNotUndefined) ? value.colnames.join(colnames) : value.colnames.top(),
-        cols: colnames !== undefined ? value.cols.add([0, colnames.length]).max([colnames.length, colnames.length]) : value.cols.extendUp(),
+        colnames: colnames !== undefined ? value.colnames.union(setRange(colnames)) : value.colnames.widenUp(),
+        cols: colnames !== undefined ? value.cols.add([0, colnames.length]).max([colnames.length, colnames.length]) : value.cols.widenUp(),
         rows: value.rows
     });
 }
 function applyGroupBySemantics(value, { by }, options) {
     if (options?.mutatedCols) {
         return new dataframe_domain_1.DataFrameDomain({
-            colnames: by.every(assert_1.isNotUndefined) ? value.colnames.join(by) : value.colnames.top(),
+            colnames: value.colnames.union(setRange(by)),
             cols: value.cols.add([0, by.length]),
             rows: value.rows
         });
@@ -286,8 +294,8 @@ function applyGroupBySemantics(value, { by }, options) {
 }
 function applySummarizeSemantics(value, { colnames }) {
     return new dataframe_domain_1.DataFrameDomain({
-        colnames: colnames?.every(assert_1.isNotUndefined) ? value.colnames.join(colnames) : value.colnames.top(),
-        cols: colnames !== undefined ? value.cols.add([0, colnames.length]).min([colnames.length, +Infinity]) : value.cols.extendUp(),
+        colnames: colnames !== undefined ? value.colnames.join(setRange([])).union(setRange(colnames)) : value.colnames.widenUp(),
+        cols: colnames !== undefined ? value.cols.add([0, colnames.length]).min([colnames.length, +Infinity]) : value.cols.widenUp(),
         rows: value.rows.min([1, +Infinity]).max([0, 1])
     });
 }
@@ -310,35 +318,36 @@ function applyJoinSemantics(value, { other, by }, options) {
             return new positive_interval_domain_1.PosIntervalDomain([lower.value[0], interval1.value[1] * interval2.value[1]]);
         }
     };
-    const commonCols = value.colnames.meet(other.colnames);
-    let duplicateCols; // whether columns may be renamed due to occurrence in both data frames
+    let duplicateCols; // columns that may be renamed due to occurring in both data frames
     let productRows; // whether the resulting rows may be a Cartesian product of the rows of the data frames
     if (options?.natural) {
-        duplicateCols = false;
-        productRows = commonCols.isValue() && commonCols.value.size === 0;
+        const commonCols = value.colnames.intersect(other.colnames).upper();
+        duplicateCols = [];
+        productRows = commonCols !== lattice_1.Bottom && commonCols !== lattice_1.Top && commonCols.size === 0;
     }
     else if (by === undefined) {
-        duplicateCols = true;
+        duplicateCols = undefined;
         productRows = true;
     }
     else if (by.length === 0) {
-        duplicateCols = commonCols.isTop() || (commonCols.isValue() && commonCols.value.size > 0);
+        const commonCols = value.colnames.intersect(other.colnames).upper();
+        duplicateCols = commonCols !== lattice_1.Bottom ? commonCols !== lattice_1.Top ? [...commonCols] : undefined : [];
         productRows = true;
     }
     else if (by.every(assert_1.isNotUndefined)) {
-        const remainingCols = commonCols.subtract(by);
-        duplicateCols = remainingCols.isTop() || (remainingCols.isValue() && remainingCols.value.size > 0);
+        const remainingCols = value.colnames.intersect(other.colnames).subtract(setRange(by)).upper();
+        duplicateCols = remainingCols !== lattice_1.Bottom ? remainingCols !== lattice_1.Top ? [...remainingCols] : undefined : [];
         productRows = false;
     }
     else {
-        duplicateCols = true;
+        duplicateCols = undefined;
         productRows = false;
     }
     const joinType = options?.join ?? 'inner';
     let rows;
     switch (joinType) {
         case 'inner':
-            rows = value.rows.min(other.rows).extendDown();
+            rows = value.rows.min(other.rows).widenDown();
             break;
         case 'left':
             rows = value.rows;
@@ -354,7 +363,7 @@ function applyJoinSemantics(value, { other, by }, options) {
     }
     return new dataframe_domain_1.DataFrameDomain({
         ...value,
-        colnames: duplicateCols ? value.colnames.top() : value.colnames.join(other.colnames),
+        colnames: duplicateCols === undefined ? value.colnames.top() : duplicateCols.length > 0 ? value.colnames.union(other.colnames).subtract(setRange(duplicateCols)).widenUp() : value.colnames.union(other.colnames),
         cols: by !== undefined ? value.cols.add(other.cols).subtract([by.length, by.length]) : mergeInterval(value.cols, other.cols),
         rows: productRows ? productInterval(rows, value.rows, other.rows) : rows
     });
@@ -369,4 +378,8 @@ function applyUnknownSemantics(value,
 _args) {
     return value.top();
 }
+function setRange(colnames) {
+    const names = colnames?.filter(assert_1.isNotUndefined) ?? [];
+    return { min: names, range: names.length === colnames?.length ? [] : lattice_1.Top };
+}
 //# sourceMappingURL=semantics.js.map

package/abstract-interpretation/data-frame/shape-inference.d.ts

@@ -1,6 +1,6 @@
-import type { FlowrConfigOptions } from '../../config';
 import { type ControlFlowInformation } from '../../control-flow/control-flow-graph';
 import type { DataflowGraph } from '../../dataflow/graph/graph';
+import type { ReadOnlyFlowrAnalyzerContext } from '../../project/context/flowr-analyzer-context';
 import type { RNode } from '../../r-bridge/lang-4.x/ast/model/model';
 import type { NormalizedAst, ParentInformation } from '../../r-bridge/lang-4.x/ast/model/processing/decorate';
 import type { NodeId } from '../../r-bridge/lang-4.x/ast/model/processing/node-id';
@@ -12,11 +12,11 @@ import { type DataFrameDomain, DataFrameStateDomain } from './dataframe-domain';
  * @param cfinfo - The control flow information containing the control flow graph
  * @param dfg    - The data flow graph to resolve variable origins and function arguments
  * @param ast    - The abstract syntax tree to resolve node IDs to AST nodes
- * @param config - The flowR configuration to use for the shape inference
+ * @param ctx    - The current flowr analyzer context
  * @returns The abstract data frame state at the exit node of the control flow graph (see {@link DataFrameStateDomain}).
  * The abstract data frame states for all other nodes are attached to the AST.
  */
-export declare function inferDataFrameShapes(cfinfo: ControlFlowInformation, dfg: DataflowGraph, ast: NormalizedAst<ParentInformation & AbstractInterpretationInfo>, config: FlowrConfigOptions): DataFrameStateDomain;
+export declare function inferDataFrameShapes(cfinfo: ControlFlowInformation, dfg: DataflowGraph, ast: NormalizedAst<ParentInformation & AbstractInterpretationInfo>, ctx: ReadOnlyFlowrAnalyzerContext): DataFrameStateDomain;
 /**
  * Resolves the abstract data frame shape of a node in the AST.
  * This requires that the data frame shape inference has been executed before using {@link inferDataFrameShapes}.

package/abstract-interpretation/data-frame/shape-inference.js

@@ -19,12 +19,12 @@ const dataframe_domain_1 = require("./dataframe-domain");
  * @param cfinfo - The control flow information containing the control flow graph
  * @param dfg    - The data flow graph to resolve variable origins and function arguments
  * @param ast    - The abstract syntax tree to resolve node IDs to AST nodes
- * @param config - The flowR configuration to use for the shape inference
+ * @param ctx    - The current flowr analyzer context
  * @returns The abstract data frame state at the exit node of the control flow graph (see {@link DataFrameStateDomain}).
  * The abstract data frame states for all other nodes are attached to the AST.
  */
-function inferDataFrameShapes(cfinfo, dfg, ast, config) {
-    const visitor = new absint_visitor_1.DataFrameShapeInferenceVisitor({ controlFlow: cfinfo, dfg: dfg, normalizedAst: ast, flowrConfig: config });
+function inferDataFrameShapes(cfinfo, dfg, ast, ctx) {
+    const visitor = new absint_visitor_1.DataFrameShapeInferenceVisitor({ controlFlow: cfinfo, dfg: dfg, normalizedAst: ast, ctx });
     visitor.start();
     const exitPoints = cfinfo.exitPoints.map(id => cfinfo.graph.getVertex(id)).filter(assert_1.isNotUndefined);
     const exitNodes = exitPoints.map(vertex => ast.idMap.get((0, control_flow_graph_1.getVertexRootId)(vertex))).filter(assert_1.isNotUndefined);

package/abstract-interpretation/domains/abstract-domain.d.ts

@@ -5,6 +5,7 @@ import { type Lattice, Top } from './lattice';
 export declare const DEFAULT_INFERENCE_LIMIT = 50;
 /**
  * An abstract domain as complete lattice with a widening operator, narrowing operator, concretization function, and abstraction function.
+ * All operations of value abstract domains should not modify the domain in-place but return new values using {@link create}.
  * @template Concrete - Type of an concrete element of the concrete domain for the abstract domain
  * @template Abstract - Type of an abstract element of the abstract domain representing possible elements (excludes `Top` and `Bot`)
  * @template Top      - Type of the Top element of the abstract domain representing all possible elements

package/abstract-interpretation/domains/abstract-domain.js

@@ -11,6 +11,7 @@ const lattice_1 = require("./lattice");
 exports.DEFAULT_INFERENCE_LIMIT = 50;
 /**
  * An abstract domain as complete lattice with a widening operator, narrowing operator, concretization function, and abstraction function.
+ * All operations of value abstract domains should not modify the domain in-place but return new values using {@link create}.
  * @template Concrete - Type of an concrete element of the concrete domain for the abstract domain
  * @template Abstract - Type of an abstract element of the abstract domain representing possible elements (excludes `Top` and `Bot`)
  * @template Top      - Type of the Top element of the abstract domain representing all possible elements
@@ -73,10 +74,10 @@ function domainElementToString(value) {
         return value.toString();
     }
     else if (value === lattice_1.Top) {
-        return '⊤';
+        return lattice_1.TopSymbol;
     }
     else if (value === lattice_1.Bottom) {
-        return '⊥';
+        return lattice_1.BottomSymbol;
     }
     return JSON.stringify(value);
 }

package/abstract-interpretation/domains/bounded-set-domain.js

@@ -125,7 +125,7 @@ class BoundedSetDomain extends abstract_domain_1.AbstractDomain {
     }
     toString() {
         if (this.value === lattice_1.Top) {
-            return '⊤';
+            return lattice_1.TopSymbol;
         }
         const string = this.value.values().map(abstract_domain_1.domainElementToString).toArray().join(', ');
         return `{${string}}`;

package/abstract-interpretation/domains/interval-domain.d.ts

@@ -55,11 +55,11 @@ export declare class IntervalDomain<Value extends IntervalLift = IntervalLift> e
     /**
      * Extends the lower bound of the current abstract value down to -∞.
      */
-    extendDown(): this;
+    widenDown(): this;
     /**
      * Extends the upper bound of the current abstract value up to +∞.
      */
-    extendUp(): this;
+    widenUp(): this;
     toJson(): unknown;
     toString(): string;
     isTop(): this is IntervalDomain<IntervalTop>;

package/abstract-interpretation/domains/interval-domain.js

@@ -75,9 +75,6 @@ class IntervalDomain extends abstract_domain_1.AbstractDomain {
         if (this.value === lattice_1.Bottom || otherValue === lattice_1.Bottom) {
             return this.bottom();
         }
-        else if (Math.max(this.value[0], otherValue[0]) > Math.min(this.value[1], otherValue[1])) {
-            return this.bottom();
-        }
         else {
             return this.create([Math.max(this.value[0], otherValue[0]), Math.min(this.value[1], otherValue[1])]);
         }
@@ -222,7 +219,7 @@ class IntervalDomain extends abstract_domain_1.AbstractDomain {
     /**
      * Extends the lower bound of the current abstract value down to -∞.
      */
-    extendDown() {
+    widenDown() {
         if (this.value === lattice_1.Bottom) {
             return this.bottom();
         }
@@ -233,7 +230,7 @@ class IntervalDomain extends abstract_domain_1.AbstractDomain {
     /**
      * Extends the upper bound of the current abstract value up to +∞.
      */
-    extendUp() {
+    widenUp() {
         if (this.value === lattice_1.Bottom) {
             return this.bottom();
         }
@@ -249,7 +246,7 @@ class IntervalDomain extends abstract_domain_1.AbstractDomain {
     }
     toString() {
         if (this.value === lattice_1.Bottom) {
-            return '⊥';
+            return lattice_1.BottomSymbol;
         }
         return `[${isFinite(this.value[0]) ? this.value[0] : '-∞'}, ${isFinite(this.value[1]) ? this.value[1] : '+∞'}]`;
     }

package/abstract-interpretation/domains/lattice.d.ts

@@ -2,10 +2,12 @@
  * The Top symbol to represent the Top element of complete lattices (e.g. of abstract domains).
  */
 export declare const Top: unique symbol;
+export declare const TopSymbol = "\u22A4";
 /**
  * The Bottom symbol to represent the Bottom element of complete lattices (e.g. of abstract domains).
  */
 export declare const Bottom: unique symbol;
+export declare const BottomSymbol = "\u22A5";
 /**
  * A complete lattice with a partially ordered set, join operator (LUB), meet operator (GLB), top element, and bottom element (e.g. for abstract domains).
  * @template Value - Type of a lattice element representing a value (may exclude `Top` and `Bot`)

package/abstract-interpretation/domains/lattice.js

@@ -1,12 +1,14 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.Bottom = exports.Top = void 0;
+exports.BottomSymbol = exports.Bottom = exports.TopSymbol = exports.Top = void 0;
 /**
  * The Top symbol to represent the Top element of complete lattices (e.g. of abstract domains).
  */
 exports.Top = Symbol('top');
+exports.TopSymbol = '⊤';
 /**
  * The Bottom symbol to represent the Bottom element of complete lattices (e.g. of abstract domains).
  */
 exports.Bottom = Symbol('bottom');
+exports.BottomSymbol = '⊥';
 //# sourceMappingURL=lattice.js.map

package/abstract-interpretation/domains/positive-interval-domain.d.ts

@@ -30,7 +30,7 @@ export declare class PosIntervalDomain<Value extends PosIntervalLift = PosInterv
     /**
      * Extends the lower bound of the current abstract value down to 0.
      */
-    extendDown(): this;
+    widenDown(): this;
     isTop(): this is PosIntervalDomain<PosIntervalTop>;
 }
 export {};

package/abstract-interpretation/domains/positive-interval-domain.js

@@ -84,7 +84,7 @@ class PosIntervalDomain extends interval_domain_1.IntervalDomain {
     /**
      * Extends the lower bound of the current abstract value down to 0.
      */
-    extendDown() {
+    widenDown() {
         if (this.value === lattice_1.Bottom) {
             return this.bottom();
         }