Improve typed task outputs (#1651)

* V1 working

* clean up imports and prints

* more clean up and add tests

* fixing tests

* fix test

* fix linting

* Fix tests

* Fix linting

* add doc string as requested by eduardo

src/crewai/agent.py

@@ -11,10 +11,12 @@ from crewai.agents.crew_agent_executor import CrewAgentExecutor
 from crewai.cli.constants import ENV_VARS
 from crewai.llm import LLM
 from crewai.memory.contextual.contextual_memory import ContextualMemory
+from crewai.task import Task
 from crewai.tools import BaseTool
 from crewai.tools.agent_tools.agent_tools import AgentTools
 from crewai.utilities import Converter, Prompts
 from crewai.utilities.constants import TRAINED_AGENTS_DATA_FILE, TRAINING_DATA_FILE
+from crewai.utilities.converter import generate_model_description
 from crewai.utilities.token_counter_callback import TokenCalcHandler
 from crewai.utilities.training_handler import CrewTrainingHandler
 
@@ -237,7 +239,7 @@ class Agent(BaseAgent):
 
     def execute_task(
         self,
-        task: Any,
+        task: Task,
         context: Optional[str] = None,
         tools: Optional[List[BaseTool]] = None,
     ) -> str:
@@ -256,6 +258,22 @@ class Agent(BaseAgent):
 
         task_prompt = task.prompt()
 
+        # If the task requires output in JSON or Pydantic format,
+        # append specific instructions to the task prompt to ensure
+        # that the final answer does not include any code block markers
+        if task.output_json or task.output_pydantic:
+            # Generate the schema based on the output format
+            if task.output_json:
+                # schema = json.dumps(task.output_json, indent=2)
+                schema = generate_model_description(task.output_json)
+
+            elif task.output_pydantic:
+                schema = generate_model_description(task.output_pydantic)
+
+            task_prompt += "\n" + self.i18n.slice("formatted_task_instructions").format(
+                output_format=schema
+            )
+
         if context:
             task_prompt = self.i18n.slice("task_with_context").format(
                 task=task_prompt, context=context

src/crewai/task.py

@@ -279,9 +279,7 @@ class Task(BaseModel):
             content = (
                 json_output
                 if json_output
-                else pydantic_output.model_dump_json()
-                if pydantic_output
-                else result
+                else pydantic_output.model_dump_json() if pydantic_output else result
             )
             self._save_file(content)
 

src/crewai/translations/en.json

@@ -21,7 +21,8 @@
     "summarizer_system_message": "You are a helpful assistant that summarizes text.",
     "sumamrize_instruction": "Summarize the following text, make sure to include all the important information: {group}",
     "summary": "This is a summary of our conversation so far:\n{merged_summary}",
-    "manager_request": "Your best answer to your coworker asking you this, accounting for the context shared."
+    "manager_request": "Your best answer to your coworker asking you this, accounting for the context shared.",
+    "formatted_task_instructions": "Ensure your final answer contains only the content in the following format: {output_format}\n\nEnsure the final output does not include any code block markers like ```json or ```python."
   },
   "errors": {
     "force_final_answer_error": "You can't keep going, this was the best you could do.\n {formatted_answer.text}",

src/crewai/utilities/converter.py

@@ -1,6 +1,6 @@
 import json
 import re
-from typing import Any, Optional, Type, Union
+from typing import Any, Optional, Type, Union, get_args, get_origin
 
 from pydantic import BaseModel, ValidationError
 
@@ -214,3 +214,38 @@ def create_converter(
         raise Exception("No output converter found or set.")
 
     return converter
+
+
+def generate_model_description(model: Type[BaseModel]) -> str:
+    """
+    Generate a string description of a Pydantic model's fields and their types.
+
+    This function takes a Pydantic model class and returns a string that describes
+    the model's fields and their respective types. The description includes handling
+    of complex types such as `Optional`, `List`, and `Dict`, as well as nested Pydantic
+    models.
+    """
+
+    def describe_field(field_type):
+        origin = get_origin(field_type)
+        args = get_args(field_type)
+
+        if origin is Union and type(None) in args:
+            non_none_args = [arg for arg in args if arg is not type(None)]
+            return f"Optional[{describe_field(non_none_args[0])}]"
+        elif origin is list:
+            return f"List[{describe_field(args[0])}]"
+        elif origin is dict:
+            key_type = describe_field(args[0])
+            value_type = describe_field(args[1])
+            return f"Dict[{key_type}, {value_type}]"
+        elif isinstance(field_type, type) and issubclass(field_type, BaseModel):
+            return generate_model_description(field_type)
+        else:
+            return field_type.__name__
+
+    fields = model.__annotations__
+    field_descriptions = [
+        f'"{name}": {describe_field(type_)}' for name, type_ in fields.items()
+    ]
+    return "{\n  " + ",\n  ".join(field_descriptions) + "\n}"

tests/utilities/test_converter.py

@@ -1,7 +1,10 @@
 import json
+from typing import Dict, List, Optional
 from unittest.mock import MagicMock, Mock, patch
 
 import pytest
+from pydantic import BaseModel
+
 from crewai.llm import LLM
 from crewai.utilities.converter import (
     Converter,
@@ -9,12 +12,11 @@ from crewai.utilities.converter import (
     convert_to_model,
     convert_with_instructions,
     create_converter,
+    generate_model_description,
     get_conversion_instructions,
     handle_partial_json,
     validate_model,
 )
-from pydantic import BaseModel
-
 from crewai.utilities.pydantic_schema_parser import PydanticSchemaParser
 
 
@@ -269,3 +271,45 @@ def test_create_converter_fails_without_agent_or_converter_cls():
         create_converter(
             llm=Mock(), text="Sample", model=SimpleModel, instructions="Convert"
         )
+
+
+def test_generate_model_description_simple_model():
+    description = generate_model_description(SimpleModel)
+    expected_description = '{\n  "name": str,\n  "age": int\n}'
+    assert description == expected_description
+
+
+def test_generate_model_description_nested_model():
+    description = generate_model_description(NestedModel)
+    expected_description = (
+        '{\n  "id": int,\n  "data": {\n  "name": str,\n  "age": int\n}\n}'
+    )
+    assert description == expected_description
+
+
+def test_generate_model_description_optional_field():
+    class ModelWithOptionalField(BaseModel):
+        name: Optional[str]
+        age: int
+
+    description = generate_model_description(ModelWithOptionalField)
+    expected_description = '{\n  "name": Optional[str],\n  "age": int\n}'
+    assert description == expected_description
+
+
+def test_generate_model_description_list_field():
+    class ModelWithListField(BaseModel):
+        items: List[int]
+
+    description = generate_model_description(ModelWithListField)
+    expected_description = '{\n  "items": List[int]\n}'
+    assert description == expected_description
+
+
+def test_generate_model_description_dict_field():
+    class ModelWithDictField(BaseModel):
+        attributes: Dict[str, int]
+
+    description = generate_model_description(ModelWithDictField)
+    expected_description = '{\n  "attributes": Dict[str, int]\n}'
+    assert description == expected_description