Refactor code and documentations

[nhungbi]

Issue #, if available:

Description of changes:

Documentations:

• Updated README with new features/import shortcuts.
• Updated pyproject to use strands-agents-evals instead of strands-evaluation
• Finalized examples for basic use cases in /examples (reorder steps so that task definition is first, updated imports, and got rid of redundant examples), also added examples for generating datasets

Structure:

• Moved files in utils to new folders and get rid of /utils. Helper functions to extract information in tasks are now in /extractors
• Changed the name of the main package from strands_evaluation to strands_evals
• Filled out init.py for nicer import shortcuts
• Got rid of mutable types as defaults
• Got rid of name == "main"
• Restructured /tests to have the same directory structure as /src and rename the tests to have the same format as strands

Code Logic:

• Added exception in _run_task in Dataset for not supporting async task function in synchronous _run_task
• Allowed user to return "input" in the task function to update the "input" for evaluation (will leave the original test case alone) for multi-shots or similar scenarios evaluation
• Got rid of @DataClass for Evaluator and BaseModel in Dataset
• In InteractionEvaluator, the prompt will now only include the input for the first interaction evaluation or iteratively select the correct input if the input is a list that matches the number of interactions
• Added TaskOutput typedDict
• Got rid of list[str] and use generic list in Interaction class to allow for more flexible definition of Interaction

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.