cwyodmodules 0.3.32__py3-none-any.whl → 0.3.33__py3-none-any.whl

cwyodmodules/batch/utilities/tools/question_answer_tool.py

@@ -10,14 +10,10 @@ from ..search.search import Search
 from .answering_tool_base import AnsweringToolBase
 from openai.types.chat import ChatCompletion
 
-from logging import getLogger
-from opentelemetry import trace, baggage
-from opentelemetry.propagate import extract
-
-# logger = getLogger("__main__" + ".base_package")
-logger = getLogger("__main__")
-# tracer = trace.get_tracer("__main__" + ".base_package")
-tracer = trace.get_tracer("__main__")
+from logging_config import logger
+env_helper: EnvHelper = EnvHelper()
+log_args = env_helper.LOG_ARGS
+log_result = env_helper.LOG_RESULT
 
 
 class QuestionAnswerTool(AnsweringToolBase):
@@ -31,21 +27,20 @@ class QuestionAnswerTool(AnsweringToolBase):
         """
         Initialize the QuestionAnswerTool with necessary helpers and configurations.
         """
-        with tracer.start_as_current_span("QuestionAnswerTool.__init__"):
-            logger.info("Initializing QuestionAnswerTool...")
-            self.name = "QuestionAnswer"
-            self.env_helper = EnvHelper()
-            self.llm_helper = LLMHelper()
-            self.search_handler = Search.get_search_handler(env_helper=self.env_helper)
-            self.verbose = True
+        logger.info("Initializing QuestionAnswerTool...")
+        self.name = "QuestionAnswer"
+        self.llm_helper = LLMHelper()
+        self.search_handler = Search.get_search_handler(env_helper=self.env_helper)
+        self.verbose = True
 
-            self.config = ConfigHelper.get_active_config_or_default()
-            logger.info("QuestionAnswerTool initialized with configuration.")
+        self.config = ConfigHelper.get_active_config_or_default()
+        logger.info("QuestionAnswerTool initialized with configuration.")
 
     def __str__(self):
         return self.answer
 
     @staticmethod
+    @logger.trace_function(log_args=False, log_result=False)
     def json_remove_whitespace(obj: str) -> str:
         """
         Remove whitespace from a JSON string.
@@ -56,14 +51,14 @@ class QuestionAnswerTool(AnsweringToolBase):
         Returns:
             str: The JSON string without whitespace.
         """
-        with tracer.start_as_current_span("QuestionAnswerTool.json_remove_whitespace"):
-            try:
-                return json.dumps(json.loads(obj), separators=(",", ":"))
-            except json.JSONDecodeError:
-                logger.exception("Failed to parse JSON in json_remove_whitespace.")
-                return obj
+        try:
+            return json.dumps(json.loads(obj), separators=(",", ":"))
+        except json.JSONDecodeError:
+            logger.exception("Failed to parse JSON in json_remove_whitespace.")
+            return obj
 
     @staticmethod
+    @logger.trace_function(log_args=False, log_result=False)
     def clean_chat_history(chat_history: list[dict]) -> list[dict]:
         """
         Clean the chat history by retaining only the content and role of each message.
@@ -74,20 +69,20 @@ class QuestionAnswerTool(AnsweringToolBase):
         Returns:
             list[dict]: The cleaned chat history.
         """
-        with tracer.start_as_current_span("QuestionAnswerTool.clean_chat_history"):
-            logger.info("Cleaning chat history...")
-            cleaned_history = [
-                {
-                    "content": message["content"],
-                    "role": message["role"],
-                }
-                for message in chat_history
-            ]
-            logger.info(
-                f"Chat history cleaned. Returning {len(cleaned_history)} messages."
-            )
-            return cleaned_history
-
+        logger.info("Cleaning chat history...")
+        cleaned_history = [
+            {
+                "content": message["content"],
+                "role": message["role"],
+            }
+            for message in chat_history
+        ]
+        logger.info(
+            f"Chat history cleaned. Returning {len(cleaned_history)} messages."
+        )
+        return cleaned_history
+
+    @logger.trace_function(log_args=False, log_result=False)
     def generate_messages(self, question: str, sources: list[SourceDocument]):
         """
         Generate messages for the language model based on the question and source documents.
@@ -99,25 +94,25 @@ class QuestionAnswerTool(AnsweringToolBase):
         Returns:
             list[dict]: The generated messages.
         """
-        with tracer.start_as_current_span("QuestionAnswerTool.generate_messages"):
-            sources_text = "\n\n".join(
-                [f"[doc{i+1}]: {source.content}" for i, source in enumerate(sources)]
-            )
-
-            logger.info(
-                f"Generating messages for question: {question} with {len(sources)} sources."
-            )
-            messages = [
-                {
-                    "content": self.config.prompts.answering_user_prompt.format(
-                        question=question, sources=sources_text
-                    ),
-                    "role": "user",
-                },
-            ]
-            logger.debug(f"Generated messages: {messages}")
-            return messages
-
+        sources_text = "\n\n".join(
+            [f"[doc{i+1}]: {source.content}" for i, source in enumerate(sources)]
+        )
+
+        logger.info(
+            f"Generating messages for question: {question} with {len(sources)} sources."
+        )
+        messages = [
+            {
+                "content": self.config.prompts.answering_user_prompt.format(
+                    question=question, sources=sources_text
+                ),
+                "role": "user",
+            },
+        ]
+        logger.debug(f"Generated messages: {messages}")
+        return messages
+
+    @logger.trace_function(log_args=False, log_result=False)
     def generate_on_your_data_messages(
         self,
         question: str,
@@ -137,93 +132,91 @@ class QuestionAnswerTool(AnsweringToolBase):
         Returns:
             list[dict]: The generated messages.
         """
-        with tracer.start_as_current_span(
-            "QuestionAnswerTool.generate_on_your_data_messages"
-        ):
-            logger.info(f"Generating On Your Data messages for question: {question}")
-            examples = []
-
-            few_shot_example = {
-                "sources": self.config.example.documents.strip(),
-                "question": self.config.example.user_question.strip(),
-                "answer": self.config.example.answer.strip(),
-            }
+        logger.info(f"Generating On Your Data messages for question: {question}")
+        examples = []
+
+        few_shot_example = {
+            "sources": self.config.example.documents.strip(),
+            "question": self.config.example.user_question.strip(),
+            "answer": self.config.example.answer.strip(),
+        }
+
+        if few_shot_example["sources"]:
+            few_shot_example["sources"] = QuestionAnswerTool.json_remove_whitespace(
+                few_shot_example["sources"]
+            )
 
-            if few_shot_example["sources"]:
-                few_shot_example["sources"] = QuestionAnswerTool.json_remove_whitespace(
-                    few_shot_example["sources"]
+        if any(few_shot_example.values()):
+            if all((few_shot_example.values())):
+                examples.append(
+                    {
+                        "content": self.config.prompts.answering_user_prompt.format(
+                            sources=few_shot_example["sources"],
+                            question=few_shot_example["question"],
+                        ),
+                        "name": "example_user",
+                        "role": "system",
+                    }
+                )
+                examples.append(
+                    {
+                        "content": few_shot_example["answer"],
+                        "name": "example_assistant",
+                        "role": "system",
+                    }
+                )
+            else:
+                warnings.warn(
+                    "Not all example fields are set in the config. Skipping few-shot example."
                 )
 
-            if any(few_shot_example.values()):
-                if all((few_shot_example.values())):
-                    examples.append(
-                        {
-                            "content": self.config.prompts.answering_user_prompt.format(
-                                sources=few_shot_example["sources"],
-                                question=few_shot_example["question"],
-                            ),
-                            "name": "example_user",
-                            "role": "system",
-                        }
-                    )
-                    examples.append(
-                        {
-                            "content": few_shot_example["answer"],
-                            "name": "example_assistant",
-                            "role": "system",
-                        }
-                    )
-                else:
-                    warnings.warn(
-                        "Not all example fields are set in the config. Skipping few-shot example."
-                    )
-
-            documents = json.dumps(
-                {
-                    "retrieved_documents": [
-                        {f"[doc{i+1}]": {"content": source.content}}
-                        for i, source in enumerate(sources)
-                    ],
-                },
-                separators=(",", ":"),
-            )
-
-            messages = [
-                {
-                    "role": "system",
-                    "content": self.config.prompts.answering_system_prompt,
-                },
-                *examples,
-                {
-                    "role": "system",
-                    "content": self.env_helper.AZURE_OPENAI_SYSTEM_MESSAGE,
-                },
-                *QuestionAnswerTool.clean_chat_history(chat_history),
-                {
-                    "role": "user",
-                    "content": [
-                        {
-                            "type": "text",
-                            "text": self.config.prompts.answering_user_prompt.format(
-                                sources=documents,
-                                question=question,
-                            ),
-                        },
-                        *(
-                            [
-                                {
-                                    "type": "image_url",
-                                    "image_url": {"url": image_url},
-                                }
-                                for image_url in image_urls
-                            ]
+        documents = json.dumps(
+            {
+                "retrieved_documents": [
+                    {f"[doc{i+1}]": {"content": source.content}}
+                    for i, source in enumerate(sources)
+                ],
+            },
+            separators=(",", ":"),
+        )
+
+        messages = [
+            {
+                "role": "system",
+                "content": self.config.prompts.answering_system_prompt,
+            },
+            *examples,
+            {
+                "role": "system",
+                "content": self.env_helper.AZURE_OPENAI_SYSTEM_MESSAGE,
+            },
+            *QuestionAnswerTool.clean_chat_history(chat_history),
+            {
+                "role": "user",
+                "content": [
+                    {
+                        "type": "text",
+                        "text": self.config.prompts.answering_user_prompt.format(
+                            sources=documents,
+                            question=question,
                         ),
-                    ],
-                },
-            ]
-            logger.debug(f"Generated On Your Data messages: {messages}")
-            return messages
+                    },
+                    *(
+                        [
+                            {
+                                "type": "image_url",
+                                "image_url": {"url": image_url},
+                            }
+                            for image_url in image_urls
+                        ]
+                    ),
+                ],
+            },
+        ]
+        logger.debug(f"Generated On Your Data messages: {messages}")
+        return messages
 
+    @logger.trace_function(log_args=False, log_result=False)
     def answer_question(self, question: str, chat_history: list[dict], **kwargs):
         """
         Answer the given question using the chat history and additional parameters.
@@ -236,44 +229,44 @@ class QuestionAnswerTool(AnsweringToolBase):
         Returns:
             Answer: The formatted answer.
         """
-        with tracer.start_as_current_span("QuestionAnswerTool.answer_question"):
-            logger.info("Answering question")
-            source_documents = Search.get_source_documents(
-                self.search_handler, question
-            )
-
-            if self.env_helper.USE_ADVANCED_IMAGE_PROCESSING:
-                image_urls = self.create_image_url_list(source_documents)
-                logger.info(
-                    f"Generated {len(image_urls)} image URLs for advanced image processing."
-                )
-            else:
-                image_urls = []
-
-            model = self.env_helper.AZURE_OPENAI_VISION_MODEL if image_urls else None
+        logger.info("Answering question")
+        source_documents = Search.get_source_documents(
+            self.search_handler, question
+        )
 
-            if self.config.prompts.use_on_your_data_format:
-                messages = self.generate_on_your_data_messages(
-                    question, chat_history, source_documents, image_urls
-                )
-            else:
-                warnings.warn(
-                    "Azure OpenAI On Your Data prompt format is recommended and should be enabled in the Admin app.",
-                )
-                messages = self.generate_messages(question, source_documents)
+        if self.env_helper.USE_ADVANCED_IMAGE_PROCESSING:
+            image_urls = self.create_image_url_list(source_documents)
+            logger.info(
+                f"Generated {len(image_urls)} image URLs for advanced image processing."
+            )
+        else:
+            image_urls = []
 
-            llm_helper = LLMHelper()
+        model = self.env_helper.AZURE_OPENAI_VISION_MODEL if image_urls else None
 
-            response = llm_helper.get_chat_completion(
-                messages, model=model, temperature=0
+        if self.config.prompts.use_on_your_data_format:
+            messages = self.generate_on_your_data_messages(
+                question, chat_history, source_documents, image_urls
             )
-            clean_answer = self.format_answer_from_response(
-                response, question, source_documents
+        else:
+            warnings.warn(
+                "Azure OpenAI On Your Data prompt format is recommended and should be enabled in the Admin app.",
             )
-            logger.info("Cleaned answer generated successfully.")
-            logger.debug(f"Answer: {clean_answer.answer}")
-            return clean_answer
+            messages = self.generate_messages(question, source_documents)
+
+        llm_helper = LLMHelper()
+
+        response = llm_helper.get_chat_completion(
+            messages, model=model, temperature=0
+        )
+        clean_answer = self.format_answer_from_response(
+            response, question, source_documents
+        )
+        logger.info("Cleaned answer generated successfully.")
+        logger.debug(f"Answer: {clean_answer.answer}")
+        return clean_answer
 
+    @logger.trace_function(log_args=False, log_result=log_result)
     def create_image_url_list(self, source_documents):
         """
         Create a list of image URLs from the source documents.
@@ -284,23 +277,23 @@ class QuestionAnswerTool(AnsweringToolBase):
         Returns:
             list[str]: The list of image URLs.
         """
-        with tracer.start_as_current_span("QuestionAnswerTool.create_image_url_list"):
-            image_types = self.config.get_advanced_image_processing_image_types()
+        image_types = self.config.get_advanced_image_processing_image_types()
 
-            blob_client = AzureBlobStorageClient()
-            container_sas = blob_client.get_container_sas()
+        blob_client = AzureBlobStorageClient()
+        container_sas = blob_client.get_container_sas()
 
-            image_urls = [
-                doc.source.replace("_SAS_TOKEN_PLACEHOLDER_", container_sas)
-                for doc in source_documents
-                if doc.title is not None and doc.title.split(".")[-1] in image_types
-            ][: self.env_helper.ADVANCED_IMAGE_PROCESSING_MAX_IMAGES]
+        image_urls = [
+            doc.source.replace("_SAS_TOKEN_PLACEHOLDER_", container_sas)
+            for doc in source_documents
+            if doc.title is not None and doc.title.split(".")[-1] in image_types
+        ][: self.env_helper.ADVANCED_IMAGE_PROCESSING_MAX_IMAGES]
 
-            logger.info(
-                f"Generated {len(image_urls)} image URLs for {len(source_documents)} source documents."
-            )
-            return image_urls
+        logger.info(
+            f"Generated {len(image_urls)} image URLs for {len(source_documents)} source documents."
+        )
+        return image_urls
 
+    @logger.trace_function(log_args=False, log_result=False)
     def format_answer_from_response(
         self,
         response: ChatCompletion,
@@ -318,26 +311,23 @@ class QuestionAnswerTool(AnsweringToolBase):
         Returns:
             Answer: The formatted answer.
         """
-        with tracer.start_as_current_span(
-            "QuestionAnswerTool.format_answer_from_response"
-        ):
-            answer = response.choices[0].message.content
-            logger.debug(f"Answer format_answer_from_response: {answer}")
-
-            # Append document citations to the answer
-            citations = "".join([
-                f"[doc{i+1}]"
-                for i in range(len(source_documents))
-                if f"[doc{i+1}]" not in answer
-            ])
-            answer_with_citations = f"{answer} {citations}"
-            # Generate Answer Object
-            clean_answer = Answer(
-                question=question,
-                answer=answer_with_citations,  # Use the answer with citations
-                source_documents=source_documents,
-                prompt_tokens=response.usage.prompt_tokens,
-                completion_tokens=response.usage.completion_tokens,
-            )
-
-            return clean_answer
+        answer = response.choices[0].message.content
+        logger.debug(f"Answer format_answer_from_response: {answer}")
+
+        # Append document citations to the answer
+        citations = "".join([
+            f"[doc{i+1}]"
+            for i in range(len(source_documents))
+            if f"[doc{i+1}]" not in answer
+        ])
+        answer_with_citations = f"{answer} {citations}"
+        # Generate Answer Object
+        clean_answer = Answer(
+            question=question,
+            answer=answer_with_citations,  # Use the answer with citations
+            source_documents=source_documents,
+            prompt_tokens=response.usage.prompt_tokens,
+            completion_tokens=response.usage.completion_tokens,
+        )
+
+        return clean_answer

cwyodmodules/batch/utilities/tools/text_processing_tool.py

@@ -3,51 +3,48 @@ from ..helpers.llm_helper import LLMHelper
 from .answering_tool_base import AnsweringToolBase
 from ..common.answer import Answer
 
-from logging import getLogger
-from opentelemetry import trace, baggage
-from opentelemetry.propagate import extract
-
-# logger = getLogger("__main__" + ".base_package")
-logger = getLogger("__main__")
-# tracer = trace.get_tracer("__main__" + ".base_package")
-tracer = trace.get_tracer("__main__")
+from ...utilities.helpers.env_helper import EnvHelper
+from logging_config import logger
+env_helper: EnvHelper = EnvHelper()
+log_args = env_helper.LOG_ARGS
+log_result = env_helper.LOG_RESULT
 
 
 class TextProcessingTool(AnsweringToolBase):
     def __init__(self) -> None:
         self.name = "TextProcessing"
 
+    @logger.trace_function(log_args=False, log_result=False)
     def answer_question(self, question: str, chat_history: List[dict] = [], **kwargs):
-        with tracer.start_as_current_span("TextProcessingTool.answer_question"):
-            logger.info(f"Answering question: {question}")
-            llm_helper = LLMHelper()
-            text = kwargs.get("text")
-            operation = kwargs.get("operation")
-            user_content = (
-                f"{operation} the following TEXT: {text}"
-                if (text and operation)
-                else question
+        logger.info(f"Answering question: {question}")
+        llm_helper = LLMHelper()
+        text = kwargs.get("text")
+        operation = kwargs.get("operation")
+        user_content = (
+            f"{operation} the following TEXT: {text}"
+            if (text and operation)
+            else question
+        )
+
+        system_message = """You are an AI assistant for the user."""
+
+        try:
+            result = llm_helper.get_chat_completion(
+                [
+                    {"role": "system", "content": system_message},
+                    {"role": "user", "content": user_content},
+                ]
             )
 
-            system_message = """You are an AI assistant for the user."""
-
-            try:
-                result = llm_helper.get_chat_completion(
-                    [
-                        {"role": "system", "content": system_message},
-                        {"role": "user", "content": user_content},
-                    ]
-                )
-
-                answer = Answer(
-                    question=question,
-                    answer=result.choices[0].message.content,
-                    source_documents=[],
-                    prompt_tokens=result.usage.prompt_tokens,
-                    completion_tokens=result.usage.completion_tokens,
-                )
-                logger.info(f"Answer generated successfully.")
-                return answer
-            except Exception as e:
-                logger.error(f"Error during get_chat_completion: {e}", exc_info=True)
-                raise
+            answer = Answer(
+                question=question,
+                answer=result.choices[0].message.content,
+                source_documents=[],
+                prompt_tokens=result.usage.prompt_tokens,
+                completion_tokens=result.usage.completion_tokens,
+            )
+            logger.info(f"Answer generated successfully.")
+            return answer
+        except Exception as e:
+            logger.error(f"Error during get_chat_completion: {e}", exc_info=True)
+            raise

cwyodmodules/logging_config.py

@@ -0,0 +1,15 @@
+import os
+from azpaddypy.mgmt.logging import create_app_logger
+
+application_insights_connection_string=os.getenv("APPLICATIONINSIGHTS_CONNECTION_STRING")
+
+# Create a single instance of the logger
+logger = create_app_logger(
+        connection_string=application_insights_connection_string,
+        service_name=__name__,
+        service_version="1.0.0",
+        enable_console_logging=True,
+    )
+
+# Export the logger instance
+__all__ = ['logger'] 

{cwyodmodules-0.3.32.dist-info → cwyodmodules-0.3.33.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cwyodmodules
-Version: 0.3.32
+Version: 0.3.33
 Summary: Add your description here
 Author-email: Patrik <patrikhartl@gmail.com>
 Classifier: Operating System :: OS Independent
@@ -40,6 +40,7 @@ Requires-Dist: azure-search-documents==11.6.0b4
 Requires-Dist: semantic-kernel==1.3.0
 Requires-Dist: pydantic==2.7.4
 Requires-Dist: pandas>=2.2.3
+Requires-Dist: azpaddypy>=0.2.4
 Dynamic: license-file
 
 # paddypy