camel-ai 0.2.13__py3-none-any.whl → 0.2.15a0__py3-none-any.whl

camel/schemas/outlines_converter.py

@@ -0,0 +1,249 @@
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+
+from typing import Any, Callable, List, Literal, Type, Union
+
+from pydantic import BaseModel
+
+from .base import BaseConverter
+
+
+class OutlinesConverter(BaseConverter):
+    r"""OutlinesConverter is a class that converts a string or a function
+    into a BaseModel schema.
+
+    Args:
+        model_type (str, optional): The model type to be used.
+        platform (str, optional): The platform to be used.
+            1. transformers
+            2. mamba
+            3. vllm
+            4. llamacpp
+            5. mlx
+            (default: "transformers")
+        **kwargs: The keyword arguments to be used. See the outlines
+            documentation for more details. See
+            https://dottxt-ai.github.io/outlines/latest/reference/models/models/
+    """
+
+    def __init__(
+        self,
+        model_type: str,
+        platform: Literal[
+            "vllm", "transformers", "mamba", "llamacpp", "mlx"
+        ] = "transformers",
+        **kwargs: Any,
+    ):
+        self.model_type = model_type
+        from outlines import models
+
+        match platform:
+            case "vllm":
+                self._outlines_model = models.vllm(model_type, **kwargs)
+            case "transformers":
+                self._outlines_model = models.transformers(
+                    model_type, **kwargs
+                )
+            case "mamba":
+                self._outlines_model = models.mamba(model_type, **kwargs)
+            case "llamacpp":
+                self._outlines_model = models.llamacpp(model_type, **kwargs)
+            case "mlx":
+                self._outlines_model = models.mlxlm(model_type, **kwargs)
+            case _:
+                raise ValueError(f"Unsupported platform: {platform}")
+
+    def convert_regex(self, content: str, regex_pattern: str) -> str:
+        r"""Convert the content to the specified regex pattern.
+
+        Args:
+            content (str): The content to be converted.
+            regex_pattern (str): The regex pattern to be used.
+
+        Returns:
+            str: The converted content.
+        """
+        import outlines
+
+        regex_generator = outlines.generate.regex(
+            self._outlines_model, regex_pattern
+        )
+        return regex_generator(content)
+
+    def convert_json(
+        self,
+        content: str,
+        output_schema: Union[str, Callable],
+    ) -> dict:
+        r"""Convert the content to the specified JSON schema given by
+        output_schema.
+
+        Args:
+            content (str): The content to be converted.
+            output_schema (Union[str, Callable]): The expected format of the
+                response.
+
+        Returns:
+            dict: The converted content in JSON format.
+        """
+        import outlines
+
+        json_generator = outlines.generate.json(
+            self._outlines_model, output_schema
+        )
+        return json_generator(content)
+
+    def convert_pydantic(
+        self,
+        content: str,
+        output_schema: Type[BaseModel],
+    ) -> BaseModel:
+        r"""Convert the content to the specified Pydantic schema.
+
+        Args:
+            content (str): The content to be converted.
+            output_schema (Type[BaseModel]): The expected format of the
+                response.
+
+        Returns:
+            BaseModel: The converted content in pydantic model format.
+        """
+        import outlines
+
+        json_generator = outlines.generate.json(
+            self._outlines_model, output_schema
+        )
+        return json_generator(content)
+
+    def convert_type(self, content: str, type_name: type) -> str:
+        r"""Convert the content to the specified type.
+
+        The following types are currently available:
+            1. int
+            2. float
+            3. bool
+            4. datetime.date
+            5. datetime.time
+            6. datetime.datetime
+            7. custom types (https://dottxt-ai.github.io/outlines/latest/reference/generation/types/)
+
+        Args:
+            content (str): The content to be converted.
+            type_name (type): The type to be used.
+
+        Returns:
+            str: The converted content.
+        """
+        import outlines
+
+        type_generator = outlines.generate.format(
+            self._outlines_model, type_name
+        )
+        return type_generator(content)
+
+    def convert_choice(self, content: str, choices: List[str]) -> str:
+        r"""Convert the content to the specified choice.
+
+        Args:
+            content (str): The content to be converted.
+            choices (List[str]): The choices to be used.
+
+        Returns:
+            str: The converted content.
+        """
+        import outlines
+
+        choices_generator = outlines.generate.choice(
+            self._outlines_model, choices
+        )
+        return choices_generator(content)
+
+    def convert_grammar(self, content: str, grammar: str) -> str:
+        r"""Convert the content to the specified grammar.
+
+        Args:
+            content (str): The content to be converted.
+            grammar (str): The grammar to be used.
+
+        Returns:
+            str: The converted content.
+        """
+        import outlines
+
+        grammar_generator = outlines.generate.cfg(
+            self._outlines_model, grammar
+        )
+        return grammar_generator(content)
+
+    def convert(  # type: ignore[override]
+        self,
+        content: str,
+        type: Literal["regex", "json", "type", "choice", "grammar"],
+        **kwargs,
+    ) -> Any:
+        r"""Formats the input content into the expected BaseModel.
+
+        Args:
+            type (Literal["regex", "json", "type", "choice", "grammar"]):
+                The type of conversion to perform. Options are:
+                    - "regex": Match the content against a regex pattern.
+                    - "pydantic": Convert the content into a pydantic model.
+                    - "json": Convert the content into a JSON based on a
+                      schema.
+                    - "type": Convert the content into a specified type.
+                    - "choice": Match the content against a list of valid
+                      choices.
+                    - "grammar": Convert the content using a specified grammar.
+            content (str): The content to be formatted.
+            **kwargs: Additional keyword arguments specific to the conversion
+                type.
+
+            - For "regex":
+                regex_pattern (str): The regex pattern to use for matching.
+
+            - For "pydantic":
+                output_schema (Type[BaseModel]): The schema to validate and
+                    format the pydantic model.
+
+            - For "json":
+                output_schema (Union[str, Callable]): The schema to validate
+                    and format the JSON object.
+
+            - For "type":
+                type_name (str): The target type name for the conversion.
+
+            - For "choice":
+                choices (List[str]): A list of valid choices to match against.
+
+            - For "grammar":
+                grammar (str): The grammar definition to use for content
+                    conversion.
+        """
+        match type:
+            case "regex":
+                return self.convert_regex(content, kwargs.get("regex_pattern"))  # type: ignore[arg-type]
+            case "pydantic":
+                return self.convert_pydantic(
+                    content, kwargs.get("output_schema")
+                )  # type: ignore[arg-type]
+            case "json":
+                return self.convert_json(content, kwargs.get("output_schema"))  # type: ignore[arg-type]
+            case "type":
+                return self.convert_type(content, kwargs.get("type_name"))  # type: ignore[arg-type]
+            case "choice":
+                return self.convert_choice(content, kwargs.get("choices"))  # type: ignore[arg-type]
+            case "grammar":
+                return self.convert_grammar(content, kwargs.get("grammar"))  # type: ignore[arg-type]
+            case _:
+                raise ValueError("Unsupported output schema type")

camel/societies/role_playing.py

@@ -509,8 +509,8 @@ class RolePlaying:
         # step and once in role play), and the model generates only one
         # response when multi-response support is enabled.
         if (
-            'n' in self.user_agent.model_config_dict.keys()
-            and self.user_agent.model_config_dict['n'] > 1
+            'n' in self.user_agent.model_backend.model_config_dict.keys()
+            and self.user_agent.model_backend.model_config_dict['n'] > 1
         ):
             self.user_agent.record_message(user_msg)
 
@@ -532,8 +532,8 @@ class RolePlaying:
         # step and once in role play), and the model generates only one
         # response when multi-response support is enabled.
         if (
-            'n' in self.assistant_agent.model_config_dict.keys()
-            and self.assistant_agent.model_config_dict['n'] > 1
+            'n' in self.assistant_agent.model_backend.model_config_dict.keys()
+            and self.assistant_agent.model_backend.model_config_dict['n'] > 1
         ):
             self.assistant_agent.record_message(assistant_msg)
 

camel/societies/workforce/workforce.py

@@ -251,7 +251,7 @@ class Workforce(BaseNode):
                 additional_info = "A Workforce node"
             elif isinstance(child, SingleAgentWorker):
                 additional_info = "tools: " + (
-                    ", ".join(child.worker.func_dict.keys())
+                    ", ".join(child.worker.tool_dict.keys())
                 )
             elif isinstance(child, RolePlayingWorker):
                 additional_info = "A Role playing node"
@@ -369,7 +369,7 @@ class Workforce(BaseNode):
             model_config_dict=model_config_dict,
         )
 
-        return ChatAgent(worker_sys_msg, model=model, tools=function_list)
+        return ChatAgent(worker_sys_msg, model=model, tools=function_list)  # type: ignore[arg-type]
 
     async def _get_returned_task(self) -> Task:
         r"""Get the task that's published by this node and just get returned

camel/storages/graph_storages/nebula_graph.py

@@ -12,9 +12,20 @@
 # limitations under the License.
 # ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
 
+import logging
+import re
 import time
 from typing import TYPE_CHECKING, Any, Dict, List, Tuple
 
+from camel.storages.graph_storages.base import BaseGraphStorage
+from camel.storages.graph_storages.graph_element import (
+    GraphElement,
+)
+from camel.utils.commons import dependencies_required
+
+logger = logging.getLogger(__name__)
+
+
 if TYPE_CHECKING:
     from nebula3.data.ResultSet import (  # type: ignore[import-untyped]
         ResultSet,
@@ -24,11 +35,6 @@ if TYPE_CHECKING:
         Session,
     )
 
-from camel.storages.graph_storages.base import BaseGraphStorage
-from camel.storages.graph_storages.graph_element import (
-    GraphElement,
-)
-from camel.utils.commons import dependencies_required
 
 MAX_RETRIES = 5
 RETRY_DELAY = 3
@@ -178,11 +184,21 @@ class NebulaGraph(BaseGraphStorage):
         """
         nodes = self._extract_nodes(graph_elements)
         for node in nodes:
-            self.add_node(node['id'], node['type'])
+            try:
+                self.add_node(node['id'], node['type'])
+            except Exception as e:
+                logger.warning(f"Failed to add node {node}. Error: {e}")
+                continue
 
         relationships = self._extract_relationships(graph_elements)
         for rel in relationships:
-            self.add_triplet(rel['subj']['id'], rel['obj']['id'], rel['type'])
+            try:
+                self.add_triplet(
+                    rel['subj']['id'], rel['obj']['id'], rel['type']
+                )
+            except Exception as e:
+                logger.warning(f"Failed to add relationship {rel}. Error: {e}")
+                continue
 
     def ensure_edge_type_exists(
         self,
@@ -253,6 +269,9 @@ class NebulaGraph(BaseGraphStorage):
             node_id (str): The ID of the node.
             tag_name (str): The tag name of the node.
         """
+        node_id = re.sub(r'[^a-zA-Z0-9\u4e00-\u9fa5]', '', node_id)
+        tag_name = re.sub(r'[^a-zA-Z0-9\u4e00-\u9fa5]', '', tag_name)
+
         self.ensure_tag_exists(tag_name)
 
         # Insert node without properties
@@ -409,6 +428,10 @@ class NebulaGraph(BaseGraphStorage):
             obj (str): The identifier for the object entity.
             rel (str): The relationship between the subject and object.
         """
+        subj = re.sub(r'[^a-zA-Z0-9\u4e00-\u9fa5]', '', subj)
+        obj = re.sub(r'[^a-zA-Z0-9\u4e00-\u9fa5]', '', obj)
+        rel = re.sub(r'[^a-zA-Z0-9\u4e00-\u9fa5]', '', rel)
+
         self.ensure_tag_exists(subj)
         self.ensure_tag_exists(obj)
         self.ensure_edge_type_exists(rel)

camel/types/enums.py

@@ -34,6 +34,7 @@ class ModelType(UnifiedModelType, Enum):
     GPT_4_TURBO = "gpt-4-turbo"
     GPT_4O = "gpt-4o"
     GPT_4O_MINI = "gpt-4o-mini"
+    O1 = "o1"
     O1_PREVIEW = "o1-preview"
     O1_MINI = "o1-mini"
 
@@ -153,9 +154,15 @@ class ModelType(UnifiedModelType, Enum):
             return self.value
         return "gpt-4o-mini"
 
+    @property
+    def support_native_structured_output(self) -> bool:
+        return self.is_openai
+
     @property
     def support_native_tool_calling(self) -> bool:
-        return any([self.is_openai, self.is_gemini, self.is_mistral])
+        return any(
+            [self.is_openai, self.is_gemini, self.is_mistral, self.is_qwen]
+        )
 
     @property
     def is_openai(self) -> bool:
@@ -166,6 +173,7 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.GPT_4_TURBO,
             ModelType.GPT_4O,
             ModelType.GPT_4O_MINI,
+            ModelType.O1,
             ModelType.O1_PREVIEW,
             ModelType.O1_MINI,
         }
@@ -452,6 +460,7 @@ class ModelType(UnifiedModelType, Enum):
         }:
             return 131_072
         elif self in {
+            ModelType.O1,
             ModelType.CLAUDE_2_1,
             ModelType.CLAUDE_3_OPUS,
             ModelType.CLAUDE_3_SONNET,

camel/types/unified_model_type.py

@@ -113,6 +113,11 @@ class UnifiedModelType(str):
         r"""Returns whether the model is a Qwen model."""
         return True
 
+    @property
+    def support_native_structured_output(self) -> bool:
+        r"""Returns whether the model supports native structured output."""
+        return False
+
     @property
     def support_native_tool_calling(self) -> bool:
         r"""Returns whether the model supports native tool calling."""

camel/utils/__init__.py

@@ -21,6 +21,7 @@ from .commons import (
     dependencies_required,
     download_tasks,
     func_string_to_callable,
+    generate_prompt_for_structured_output,
     get_first_int,
     get_prompt_template_key_words,
     get_pydantic_major_version,
@@ -78,4 +79,5 @@ __all__ = [
     "track_agent",
     "handle_http_error",
     "get_pydantic_model",
+    "generate_prompt_for_structured_output",
 ]

camel/utils/commons.py

@@ -12,7 +12,6 @@
 # limitations under the License.
 # ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
 import importlib
-import logging
 import os
 import platform
 import re
@@ -40,19 +39,14 @@ import pydantic
 import requests
 from pydantic import BaseModel
 
-from camel.logger import get_logger
 from camel.types import TaskType
 
 from .constants import Constants
 
 F = TypeVar('F', bound=Callable[..., Any])
 
-logger = get_logger(__name__)
 
-
-def print_text_animated(
-    text, delay: float = 0.02, end: str = "", log_level: int = logging.INFO
-):
+def print_text_animated(text, delay: float = 0.02, end: str = ""):
     r"""Prints the given text with an animated effect.
 
     Args:
@@ -61,22 +55,10 @@ def print_text_animated(
             (default: :obj:`0.02`)
         end (str, optional): The end character to print after each
             character of text. (default: :obj:`""`)
-        log_level (int, optional): The log level to use.
-            See https://docs.python.org/3/library/logging.html#levels
-            (default: :obj:`logging.INFO`)
     """
-    if logger.isEnabledFor(log_level):
-        # timestamp and other prefixes
-        logger.log(log_level, '')
-
-        for char in text:
-            print(char, end=end, flush=True)
-            time.sleep(delay)
-        # Close the log entry
-        logger.log(log_level, '')
-    else:
-        # This may be relevant for logging frameworks
-        logger.log(log_level, text)
+    for char in text:
+        print(char, end=end, flush=True)
+        time.sleep(delay)
 
 
 def get_prompt_template_key_words(template: str) -> Set[str]:
@@ -406,7 +388,8 @@ def json_to_function_code(json_obj: Dict) -> str:
     }
 
     for prop in required:
-        description = properties[prop]['description']
+        # if no description, return empty string
+        description = properties[prop].get('description', "")
         prop_type = properties[prop]['type']
         python_type = prop_to_python.get(prop_type, prop_type)
         args.append(f"{prop}: {python_type}")
@@ -624,3 +607,39 @@ def retry_request(
                 time.sleep(delay)
             else:
                 raise
+
+
+def generate_prompt_for_structured_output(
+    response_format: Optional[Type[BaseModel]],
+    user_message: str,
+) -> str:
+    """
+    This function generates a prompt based on the provided Pydantic model and
+    user message.
+
+    Args:
+        response_format (Type[BaseModel]): The Pydantic model class.
+        user_message (str): The user message to be used in the prompt.
+
+    Returns:
+        str: A prompt string for the LLM.
+    """
+    if response_format is None:
+        return user_message
+
+    json_schema = response_format.model_json_schema()
+    sys_prompt = (
+        "Given the user message, please generate a JSON response adhering "
+        "to the following JSON schema:\n"
+        f"{json_schema}\n"
+        "Make sure the JSON response is valid and matches the EXACT structure "
+        "defined in the schema. Your result should only be a valid json "
+        "object, without any other text or comments.\n"
+    )
+    user_prompt = f"User message: {user_message}\n"
+
+    final_prompt = f"""
+    {sys_prompt}
+    {user_prompt}
+    """
+    return final_prompt

camel/utils/token_counting.py

@@ -63,6 +63,7 @@ def get_model_encoding(value_for_tiktoken: str):
         encoding = tiktoken.encoding_for_model(value_for_tiktoken)
     except KeyError:
         if value_for_tiktoken in [
+            ModelType.O1.value,
             ModelType.O1_MINI.value,
             ModelType.O1_PREVIEW.value,
         ]:

{camel_ai-0.2.13.dist-info → camel_ai-0.2.15a0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: camel-ai
-Version: 0.2.13
+Version: 0.2.15a0
 Summary: Communicative Agents for AI Society Study
 Home-page: https://www.camel-ai.org/
 License: Apache-2.0
@@ -48,7 +48,7 @@ Requires-Dist: discord.py (>=2.3.2,<3.0.0) ; extra == "tools" or extra == "all"
 Requires-Dist: docker (>=7.1.0,<8.0.0) ; extra == "tools" or extra == "runtime" or extra == "all"
 Requires-Dist: docstring-parser (>=0.15,<0.16)
 Requires-Dist: docx2txt (>=0.8,<0.9) ; extra == "tools" or extra == "all"
-Requires-Dist: duckduckgo-search (>=6.2.12,<7.0.0) ; extra == "search-tools" or extra == "tools" or extra == "all"
+Requires-Dist: duckduckgo-search (>=6.3.5,<7.0.0) ; extra == "search-tools" or extra == "tools" or extra == "all"
 Requires-Dist: e2b-code-interpreter (>=1.0.3,<2.0.0) ; extra == "tools" or extra == "all"
 Requires-Dist: eval-type-backport (==0.2.0)
 Requires-Dist: ffmpeg-python (>=0.2.0,<0.3.0) ; extra == "tools" or extra == "all"
@@ -68,12 +68,12 @@ Requires-Dist: mock (>=5,<6) ; extra == "test"
 Requires-Dist: nebula3-python (==3.8.2) ; extra == "rag" or extra == "graph-storages" or extra == "all"
 Requires-Dist: neo4j (>=5.18.0,<6.0.0) ; extra == "rag" or extra == "graph-storages" or extra == "all"
 Requires-Dist: newspaper3k (>=0.2.8,<0.3.0) ; extra == "tools" or extra == "all"
-Requires-Dist: nltk (==3.9.1) ; extra == "tools" or extra == "all"
 Requires-Dist: notion-client (>=2.2.1,<3.0.0) ; extra == "tools" or extra == "all"
 Requires-Dist: numpy (>=1,<2)
-Requires-Dist: openai (>=1.45.0,<2.0.0)
+Requires-Dist: openai (>=1.58.1,<2.0.0)
 Requires-Dist: openapi-spec-validator (>=0.7.1,<0.8.0) ; extra == "tools" or extra == "all"
 Requires-Dist: opencv-python (>=4,<5) ; extra == "huggingface-agent" or extra == "all"
+Requires-Dist: outlines (>=0.1.7,<0.2.0) ; extra == "tools" or extra == "all"
 Requires-Dist: pandoc
 Requires-Dist: pathlib (>=1.0.1,<2.0.0)
 Requires-Dist: pdfplumber (>=0.11.0,<0.12.0) ; extra == "tools" or extra == "all"
@@ -109,7 +109,7 @@ Requires-Dist: tiktoken (>=0.7.0,<0.8.0)
 Requires-Dist: torch (==2.2.1) ; (platform_system == "Darwin" and platform_machine != "arm64") and (extra == "huggingface-agent" or extra == "all")
 Requires-Dist: torch (>=2,<3) ; (platform_system != "Darwin" or platform_machine == "arm64") and (extra == "huggingface-agent" or extra == "all")
 Requires-Dist: transformers (>=4,<5) ; extra == "huggingface-agent" or extra == "all"
-Requires-Dist: unstructured[all-docs] (==0.16.11) ; extra == "rag" or extra == "all"
+Requires-Dist: unstructured[all-docs] (==0.16.11) ; extra == "rag" or extra == "tools" or extra == "all" or extra == "all"
 Requires-Dist: wikipedia (>=1,<2) ; extra == "search-tools" or extra == "tools" or extra == "all"
 Requires-Dist: wolframalpha (>=5.0.0,<6.0.0) ; extra == "search-tools" or extra == "tools" or extra == "all"
 Requires-Dist: yt-dlp (>=2024.11.4,<2025.0.0) ; extra == "tools" or extra == "all"
@@ -263,7 +263,7 @@ conda create --name camel python=3.10
 conda activate camel
 
 # Clone github repo
-git clone -b v0.2.13 https://github.com/camel-ai/camel.git
+git clone -b v0.2.15a0 https://github.com/camel-ai/camel.git
 
 # Change directory into project directory
 cd camel