Task Definition Format

Available tasks are provided to the constructor as task-definition strings. Each task definition uses the format:

Example:

Required Class

• Each entry in availableTaskDefinitions defines one task id and its ordered functionalities.
• All constructor inputs are guaranteed valid.
• The following is a valid example:

Supported Functionalities

removeEmptyStrings

• Removes strings that are exactly empty or contain only spaces.
• List.of("cat", "", "dog", " ") -> List.of("cat", "dog")

trimEachAndRemoveEmpty

• Trims every string, then removes strings that become empty after trimming.
• List.of(" cat ", " ", "dog") -> List.of("cat", "dog")

removeDuplicateStringsIgnoreCaseKeepFirst

• Keeps only the first occurrence of each string ignoring case.
• List.of("cat", "dog", "Cat", "DOG") -> List.of("cat", "dog")

sortStringsLexicographically

• Sorts the list in lexicographical order.
• List.of("banana", "apple", "cat") -> List.of("apple", "banana", "cat")

sortByLength

• Sorts strings by length in ascending order. If lengths are equal, keep their existing relative order.
• List.of("elephant", "cat", "dog") -> List.of("cat", "dog", "elephant")

reverseList

• Reverses the order of the list.
• List.of("a", "b", "c") -> List.of("c", "b", "a")

lowercase

• Converts every string in the list to lowercase.
• List.of("Cat", "DOG") -> List.of("cat", "dog")

uppercase

• Converts every string in the list to uppercase.
• List.of("Cat", "dog") -> List.of("CAT", "DOG")

Execution Rules

• Inside one task, functionalities are applied in the exact order given in its task definition.
• In a sequential workflow, the output of one task becomes the input of the next task.
• In a parallel workflow, every task receives the same original input list of that workflow.
• For a parallel workflow, task outputs are merged by concatenating the output lists in the same task order used in the workflow definition.
• If workflow A is connected to workflow B, then the final output of A becomes the input of B.
• Each workflow may have at most one direct downstream and at most one direct upstream workflow.
• If a workflow chain is executed, executeWorkflow returns the final output of the last workflow in that chain.

Top Methods

boolean submitWorkflow(String workflowId, List<String> taskIds, boolean runInParallel)

• Creates a new workflow with the given task ids and execution mode.
• workflowId must be unique.
• Each task id in taskIds must exist in availableTaskDefinitions.
• 1 ≤ taskIds.size()
• Return true if the workflow is created successfully, otherwise return false.

boolean connectWorkflows(String sourceWorkflowId, String targetWorkflowId)

• Connects one workflow to another.
• The final output of sourceWorkflowId becomes the input of targetWorkflowId.
• Both workflows must already exist.
• sourceWorkflowId and targetWorkflowId must be different.
• The connection must not create a cycle.
• sourceWorkflowId must not already have a direct downstream workflow.
• targetWorkflowId must not already have a direct upstream workflow.
• Therefore, workflows can form only a linear chain. Branching and merging are not supported.
• Return true if the connection is added successfully, otherwise return false.

List<String> executeWorkflow(String workflowId, List<String> inputData)

• Executes the given workflow using the provided input list.
• workflowId must exist.
• Every task in the workflow must apply its internal functionalities in order.
• If the workflow is sequential, task outputs flow one by one through the workflow.
• If the workflow is parallel, all tasks use the same input list and their output lists are concatenated in task order.
• If a downstream workflow is connected, its execution starts automatically with the current workflow's final output.
• Return the final List<String> produced by the last workflow in the chain.

String getWorkflowDetails(String workflowId)

• Returns a string representation of the workflow configuration.
• The response must include workflow id, execution mode, task order, and directly connected downstream workflow if present.
• If the workflow does not exist, return "WORKFLOW_NOT_FOUND".
• The returned string must exactly follow this format:
  workflowId=<workflowId>, mode=<SEQUENTIAL|PARALLEL>, tasks=<task1->task2->...>, downstream=<downstreamWorkflowId or empty string>

Output Format Example

• workflowId=WF-MAIN, mode=SEQUENTIAL, tasks=validate->clean->refine, downstream=WF-NOTIFY
• workflowId=WF-MAIN, mode=SEQUENTIAL, tasks=validate->clean->refine, downstream

Constraints

• 1 ≤ availableTaskDefinitions.size() ≤ 10^4
• 1 ≤ taskIds.size() ≤ 50 for one workflow
• 0 ≤ inputData.size() ≤ 10^4
• Each task id and workflow id is a non-empty string.
• Task ids are unique across availableTaskDefinitions.
• Each task definition contains at least one functionality.
• A workflow cannot be submitted more than once with the same id.
• Self-connections are invalid.
• Cyclic workflow connections are invalid.
• All operations should be handled in memory.

Example 1

    WorkflowManager(
    availableTaskDefinitions=List.of(
    "validate-removeEmptyStrings,trimEachAndRemoveEmpty",
    "clean-lowercase,sortByLength",
    "refine-removeDuplicateStringsIgnoreCaseKeepFirst,sortStringsLexicographically",
    "email-uppercase",
    "sms-reverseList"
    )
    )

Example 2

    WorkflowManager(
    availableTaskDefinitions=List.of(
    "validate-removeEmptyStrings,trimEachAndRemoveEmpty",
    "clean-lowercase,sortByLength",
    "refine-removeDuplicateStringsIgnoreCaseKeepFirst,sortStringsLexicographically",
    "email-uppercase",
    "sms-reverseList"
    )
    )