Pass tests

Makefile

@@ -8,19 +8,19 @@ extra_quality_checks:
 	python utils/check_copies.py
 	python utils/check_dummies.py
 	python utils/check_repo.py
-	doc-builder style agents docs/source --max_len 119
+	doc-builder style smolagents docs/source --max_len 119
 
 # this target runs checks on all files
 quality:
 	ruff check $(check_dirs)
 	ruff format --check $(check_dirs)
-	doc-builder style agents docs/source --max_len 119 --check_only
+	doc-builder style smolagents docs/source --max_len 119 --check_only
 
 # Format source code automatically and check is there are any problems left that need manual fixing
 style:
 	ruff check $(check_dirs) --fix
 	ruff format $(check_dirs)
-	doc-builder style agents docs/source --max_len 119
+	doc-builder style smolagents docs/source --max_len 119
 	
 # Run tests for the library
 test_big_modeling:

src/smolagents/agents.py

@@ -370,8 +370,7 @@ class MultiStepAgent:
         try:
             return self.model(self.input_messages)
         except Exception as e:
-            error_msg = f"Error in generating final LLM output:\n{e}"
-            raise AgentGenerationError(error_msg)
+            return f"Error in generating final LLM output:\n{e}"
 
     def execute_tool_call(self, tool_name: str, arguments: Dict[str, str]) -> Any:
         """

src/smolagents/default_tools.py

@@ -153,8 +153,8 @@ class DuckDuckGoSearchTool(Tool):
     }
     output_type = "any"
 
-    def __init__(self):
-        super().__init__(self)
+    def __init__(self, **kwargs):
+        super().__init__(self, **kwargs)
         try:
             from duckduckgo_search import DDGS
         except ImportError:

src/smolagents/models.py

@@ -410,7 +410,12 @@ class TransformersModel(Model):
 
 
 class LiteLLMModel(Model):
-    def __init__(self, model_id="anthropic/claude-3-5-sonnet-20240620", api_base=None, api_key=None):
+    def __init__(
+        self,
+        model_id="anthropic/claude-3-5-sonnet-20240620",
+        api_base=None,
+        api_key=None,
+    ):
         super().__init__()
         self.model_id = model_id
         # IMPORTANT - Set this to TRUE to add the function to the prompt for Non OpenAI LLMs

src/smolagents/prompts.py

@@ -517,5 +517,6 @@ __all__ = [
     "PLAN_UPDATE_FINAL_PLAN_REDACTION",
     "SINGLE_STEP_CODE_SYSTEM_PROMPT",
     "CODE_SYSTEM_PROMPT",
+    "TOOL_CALLING_SYSTEM_PROMPT",
     "MANAGED_AGENT_PROMPT",
 ]

src/smolagents/tool_validation.py

@@ -7,8 +7,6 @@ from .utils import BASE_BUILTIN_MODULES
 
 _BUILTIN_NAMES = set(vars(builtins))
 
-IMPORTED_PACKAGES = BASE_BUILTIN_MODULES
-
 
 class MethodChecker(ast.NodeVisitor):
     """
@@ -91,7 +89,7 @@ class MethodChecker(ast.NodeVisitor):
         if isinstance(node.ctx, ast.Load):
             if not (
                 node.id in _BUILTIN_NAMES
-                or node.id in IMPORTED_PACKAGES
+                or node.id in BASE_BUILTIN_MODULES
                 or node.id in self.arg_names
                 or node.id == "self"
                 or node.id in self.class_attributes
@@ -105,7 +103,7 @@ class MethodChecker(ast.NodeVisitor):
         if isinstance(node.func, ast.Name):
             if not (
                 node.func.id in _BUILTIN_NAMES
-                or node.func.id in IMPORTED_PACKAGES
+                or node.func.id in BASE_BUILTIN_MODULES
                 or node.func.id in self.arg_names
                 or node.func.id == "self"
                 or node.func.id in self.class_attributes

src/smolagents/tools.py

@@ -854,7 +854,7 @@ def load_tool(
         main_module = importlib.import_module("smolagents")
         tools_module = main_module
         tool_class = getattr(tools_module, tool_class_name)
-        return tool_class(model_repo_id, token=token, **kwargs)
+        return tool_class(token=token, **kwargs)
     else:
         return Tool.from_hub(
             task_or_repo_id,

src/smolagents/types.py

@@ -104,7 +104,9 @@ class AgentImage(AgentType, ImageType):
         self._raw = None
         self._tensor = None
 
-        if isinstance(value, ImageType):
+        if isinstance(value, AgentImage):
+            self._raw, self._path, self._tensor = value._raw, value._path, value._tensor
+        elif isinstance(value, ImageType):
             self._raw = value
         elif isinstance(value, bytes):
             self._raw = Image.open(BytesIO(value))

tests/test_all_docs.py

@@ -123,6 +123,7 @@ class TestDocs:
                 "ToolCollection",
                 "image_generation_tool",
                 "from_langchain",
+                "while llm_should_continue(memory):",
             ]
             code_blocks = [
                 block.replace("<YOUR_HUGGINGFACEHUB_API_TOKEN>", self.hf_token).replace(

tests/test_final_answer.py

@@ -59,6 +59,6 @@ class FinalAnswerToolTester(unittest.TestCase, ToolTesterMixin):
     def test_agent_type_output(self):
         inputs = self.create_inputs()
         for input_type, input in inputs.items():
-            output = self.tool(**input)
+            output = self.tool(**input, sanitize_inputs_outputs=True)
             agent_type = AGENT_TYPE_MAPPING[input_type]
             self.assertTrue(isinstance(output, agent_type))

tests/test_monitoring.py

@@ -55,8 +55,8 @@ final_answer('This is the final answer.')
                 self.last_input_token_count = 10
                 self.last_output_token_count = 20
 
-            def __call__(self, prompt, **kwargs):
-                return 'Action:{"action": "final_answer", "action_input": {"answer": "image"}}'
+            def get_tool_call(self, prompt, **kwargs):
+                return "final_answer", {"answer": "image"}, "fake_id"
 
         agent = ToolCallingAgent(
             tools=[],
@@ -96,18 +96,21 @@ final_answer('This is the final answer.')
                 self.last_output_token_count = 20
 
             def __call__(self, prompt, **kwargs):
-                raise AgentError
+                self.last_input_token_count = 10
+                self.last_output_token_count = 0
+                raise Exception("Cannot generate")
 
         agent = CodeAgent(
             tools=[],
             model=FakeLLMModel(),
             max_iterations=1,
         )
-
         agent.run("Fake task")
 
-        self.assertEqual(agent.monitor.total_input_token_count, 20)
-        self.assertEqual(agent.monitor.total_output_token_count, 40)
+        self.assertEqual(
+            agent.monitor.total_input_token_count, 20
+        )  # Should have done two monitoring callbacks
+        self.assertEqual(agent.monitor.total_output_token_count, 0)
 
     def test_streaming_agent_text_output(self):
         def dummy_model(prompt, **kwargs):
@@ -132,14 +135,16 @@ final_answer('This is the final answer.')
         self.assertIn("This is the final answer.", final_message.content)
 
     def test_streaming_agent_image_output(self):
-        def dummy_model(prompt, **kwargs):
-            return (
-                'Action:{"action": "final_answer", "action_input": {"answer": "image"}}'
-            )
+        class FakeLLM:
+            def __init__(self):
+                pass
+
+            def get_tool_call(self, messages, **kwargs):
+                return "final_answer", {"answer": "image"}, "fake_id"
 
         agent = ToolCallingAgent(
             tools=[],
-            model=dummy_model,
+            model=FakeLLM(),
             max_iterations=1,
         )
 
@@ -148,7 +153,7 @@ final_answer('This is the final answer.')
             stream_to_gradio(
                 agent,
                 task="Test task",
-                image=AgentImage(value="path.png"),
+                additional_args=dict(image=AgentImage(value="path.png")),
                 test_mode=True,
             )
         )

tests/test_python_interpreter.py

@@ -41,17 +41,16 @@ class PythonInterpreterToolTester(unittest.TestCase, ToolTesterMixin):
 
     def test_exact_match_arg(self):
         result = self.tool("(2 / 2) * 4")
-        self.assertEqual(result, "4.0")
+        self.assertEqual(result, "Stdout:\n\nOutput: 4.0")
 
     def test_exact_match_kwarg(self):
         result = self.tool(code="(2 / 2) * 4")
-        self.assertEqual(result, "4.0")
+        self.assertEqual(result, "Stdout:\n\nOutput: 4.0")
 
     def test_agent_type_output(self):
         inputs = ["2 * 2"]
-        output = self.tool(*inputs)
+        output = self.tool(*inputs, sanitize_inputs_outputs=True)
         output_type = AGENT_TYPE_MAPPING[self.tool.output_type]
-        print("OKK", type(output), output_type, AGENT_TYPE_MAPPING)
         self.assertTrue(isinstance(output, output_type))
 
     def test_agent_types_inputs(self):
@@ -71,7 +70,7 @@ class PythonInterpreterToolTester(unittest.TestCase, ToolTesterMixin):
                 _inputs.append(AGENT_TYPE_MAPPING[input_type](_input))
 
         # Should not raise an error
-        output = self.tool(*inputs)
+        output = self.tool(*inputs, sanitize_inputs_outputs=True)
         output_type = AGENT_TYPE_MAPPING[self.tool.output_type]
         self.assertTrue(isinstance(output, output_type))
 

tests/test_search.py

@@ -27,4 +27,4 @@ class DuckDuckGoSearchToolTester(unittest.TestCase, ToolTesterMixin):
 
     def test_exact_match_arg(self):
         result = self.tool("Agents")
-        assert isinstance(result, list) and isinstance(result[0], dict)
+        assert isinstance(result, str)

tests/test_tools.py

@@ -93,7 +93,7 @@ class ToolTesterMixin:
 
     def test_agent_type_output(self):
         inputs = create_inputs(self.tool.inputs)
-        output = self.tool(**inputs)
+        output = self.tool(**inputs, sanitize_inputs_outputs=True)
         if self.tool.output_type != "any":
             agent_type = AGENT_TYPE_MAPPING[self.tool.output_type]
             self.assertTrue(isinstance(output, agent_type))
@@ -164,20 +164,20 @@ class ToolTests(unittest.TestCase):
             assert coolfunc.output_type == "number"
         assert "docstring has no description for the argument" in str(e)
 
-    def test_tool_definition_raises_error_imports_outside_function(self):
+    def test_saving_tool_raises_error_imports_outside_function(self):
         with pytest.raises(Exception) as e:
-            from datetime import datetime
+            import numpy as np
 
             @tool
             def get_current_time() -> str:
                 """
                 Gets the current time.
                 """
-                return str(datetime.now())
+                return str(np.random.random())
 
             get_current_time.save("output")
 
-        assert "datetime" in str(e)
+        assert "np" in str(e)
 
         # Also test with classic definition
         with pytest.raises(Exception) as e:
@@ -189,12 +189,12 @@ class ToolTests(unittest.TestCase):
                 output_type = "string"
 
                 def forward(self):
-                    return str(datetime.now())
+                    return str(np.random.random())
 
             get_current_time = GetCurrentTimeTool()
             get_current_time.save("output")
 
-        assert "datetime" in str(e)
+        assert "np" in str(e)
 
     def test_tool_definition_raises_no_error_imports_in_function(self):
         @tool

tests/test_types.py

@@ -20,7 +20,6 @@ from pathlib import Path
 
 from smolagents.types import AgentAudio, AgentImage, AgentText
 from transformers.testing_utils import (
-    get_tests_dir,
     require_soundfile,
     require_torch,
     require_vision,
@@ -91,7 +90,7 @@ class AgentImageTests(unittest.TestCase):
         self.assertTrue(os.path.exists(path))
 
     def test_from_string(self):
-        path = Path(get_tests_dir("fixtures/")) / "000000039769.png"
+        path = Path("tests/fixtures/000000039769.png")
         image = Image.open(path)
         agent_type = AgentImage(path)
 
@@ -103,7 +102,7 @@ class AgentImageTests(unittest.TestCase):
         self.assertTrue(os.path.exists(path))
 
     def test_from_image(self):
-        path = Path(get_tests_dir("fixtures/")) / "000000039769.png"
+        path = Path("tests/fixtures/000000039769.png")
         image = Image.open(path)
         agent_type = AgentImage(image)