@eagleoutice/flowr-dev/dataflow/processor.d.ts

Static Dataflow Analyzer and Program Slicer for the R Programming Language

/**
 * Based on a two-way fold, this processor will automatically supply scope information
 */
import type { ControlDependency, DataflowInformation } from './info';
import type { NormalizedAst, ParentInformation, RNodeWithParent } from '../r-bridge/lang-4.x/ast/model/processing/decorate';
import type { REnvironmentInformation } from './environments/environment';
import type { RNode } from '../r-bridge/lang-4.x/ast/model/model';
import type { KnownParserType, Parser } from '../r-bridge/parser';
import type { FlowrAnalyzerContext } from '../project/context/flowr-analyzer-context';
import type { NodeId } from '../r-bridge/lang-4.x/ast/model/processing/node-id';
export interface DataflowProcessorInformation<OtherInfo> {
    readonly parser: Parser<KnownParserType>;
    /**
     * Initial and frozen ast-information
     */
    readonly completeAst: NormalizedAst<OtherInfo>;
    /**
     * Correctly contains pushed local scopes introduced by `function` scopes.
     * Will by default *not* contain any symbol-bindings introduced along the way; they have to be decorated when moving up the tree.
     */
    readonly environment: REnvironmentInformation;
    /**
     * Other processors to be called by the given functions
     */
    readonly processors: DataflowProcessors<OtherInfo>;
    /**
     * The chain of file paths that lead to this inclusion.
     * The most recent (last) entry is expected to always be the current one.
     */
    readonly referenceChain: (string | undefined)[];
    /**
     * The chain of control-flow {@link NodeId}s that lead to the current node (e.g., of known ifs).
     */
    readonly cds: ControlDependency[] | undefined;
    /**
     * The flowr context used for environment seeding, files, and precision control, ...
     */
    readonly ctx: FlowrAnalyzerContext;
    /**
     * If set, the function call with this id is known to resolve to built-ins only,
     * so its vertex needs no environment snapshot (it would be discarded by markAsOnlyBuiltIn anyway).
     */
    readonly builtInNoEnv?: NodeId;
    /**
     * Escape hatch for hot recursive paths.
     * When set and its `rootId` matches the function call currently being processed by {@link processAllArguments},
     * its `info` is used as the already-processed first argument instead of processing that argument again.
     */
    readonly precomputedFirstArg?: {
        readonly rootId: NodeId;
        readonly info: DataflowInformation;
    };
    /**
     * Companion to {@link precomputedFirstArg}: when set and its `nodeId` matches the node
     * {@link processFunctionArgument} is about to process as an argument's value, its `info` is used instead of
     * processing (and recursing into) that node again. This allows to separate arg wrappers from their content!
     */
    readonly precomputedValue?: {
        readonly nodeId: NodeId;
        readonly info: DataflowInformation;
    };
}
export type DataflowProcessor<OtherInfo, NodeType extends RNodeWithParent<OtherInfo>> = (node: NodeType, data: DataflowProcessorInformation<OtherInfo>) => DataflowInformation;
type NodeWithKey<OtherInfo, Key> = RNode<OtherInfo & ParentInformation> & {
    type: Key;
};
/**
 * This way, a processor mapped to a {@link RType#Symbol} require a {@link RSymbol} as first parameter and so on.
 */
export type DataflowProcessors<OtherInfo> = {
    [key in RNode['type']]: DataflowProcessor<OtherInfo, NodeWithKey<OtherInfo, key>>;
};
/**
 * Originally, dataflow processor was written as a two-way fold, but this produced problems when trying to resolve function calls
 * which require information regarding the calling *and* definition context. While this only is a problem for late bindings as they happen
 * with functions (and probably quote'd R-expressions), it is still a problem that must be dealt with.
 * Therefore, the dataflow processor has no complete control over the traversal and merge strategy of the graph, with each processor being in
 * the position to call the other processors as needed for its children.
 * <p>
 * Now this method can be called recursively within the other processors to parse the dataflow for nodes that you cannot narrow down
 * in type or context.
 * @param current - The current node to start processing from
 * @param data    - The initial (/current) information to be passed down
 */
export declare function processDataflowFor<OtherInfo>(current: RNode<OtherInfo & ParentInformation>, data: DataflowProcessorInformation<OtherInfo & ParentInformation>): DataflowInformation;
export {};