ai-parrot 0.8.3__cp39-cp39-manylinux2014_x86_64.manylinux_2_17_x86_64.whl

parrot/bots/data.py

@@ -0,0 +1,730 @@
+from pathlib import Path
+from typing import Any, List, Dict, Union, Optional
+import re
+from datetime import datetime, timezone, timedelta
+from string import Template
+import redis.asyncio as aioredis
+import pandas as pd
+from aiohttp import web
+from datamodel.typedefs import SafeDict
+from datamodel.parsers.json import json_encoder, json_decoder  # pylint: disable=E0611 # noqa
+from langchain_core.exceptions import OutputParserException
+from langchain_core.messages import HumanMessage
+from langchain_experimental.tools.python.tool import PythonAstREPLTool
+from langchain_experimental.agents.agent_toolkits import create_pandas_dataframe_agent
+from navconfig import BASE_DIR
+from navconfig.logging import logging
+from querysource.queries.qs import QS
+from querysource.queries.multi import MultiQS
+from ..tools import AbstractTool
+from ..tools.docx import DocxGeneratorTool
+from .agent import BasicAgent
+from ..models import AgentResponse
+from ..conf import BASE_STATIC_URL, REDIS_HISTORY_URL
+from .prompts import AGENT_PROMPT_SUFFIX, FORMAT_INSTRUCTIONS
+from .prompts.data import (
+    TOOL_CALLING_PROMPT_PREFIX,
+    TOOL_CALLING_PROMPT_SUFFIX,
+    REACT_PROMPT_PREFIX
+)
+
+## Enable Debug:
+from langchain.globals import set_debug, set_verbose
+
+# Enable verbosity for debugging
+# set_debug(True)
+# set_verbose(True)
+
+
+def brace_escape(text: str) -> str:
+    return text.replace('{', '{{').replace('}', '}}')
+
+class PandasAgent(BasicAgent):
+    """
+    A simple agent that uses the pandas library to perform data analysis tasks.
+    TODO
+    - add notify tool (email, telegram, teams)
+    - specific teams tool to send private messages from agents
+    """
+
+    def __init__(
+        self,
+        name: str = 'Agent',
+        agent_type: str = None,
+        llm: Optional[str] = None,
+        tools: List[AbstractTool] = None,
+        system_prompt: str = None,
+        human_prompt: str = None,
+        prompt_template: str = None,
+        df: Union[list[pd.DataFrame], Dict[str, pd.DataFrame]] = None,
+        query: Union[List[str], dict] = None,
+        **kwargs
+    ):
+        self.chatbot_id: str = None
+        self._queries = query
+        self.df = self._define_dataframe(df)
+        self.df_locals: dict = {}
+        # Agent ID:
+        self._prompt_prefix = None
+        # Must be one of 'tool-calling', 'openai-tools', 'openai-functions', or 'zero-shot-react-description'.
+        self.agent_type = agent_type or "tool-calling"
+        if self.agent_type == "tool-calling":
+            self._prompt_template = prompt_template or TOOL_CALLING_PROMPT_PREFIX
+            self._format_instructions = ''
+        else:
+            self._prompt_template = prompt_template or REACT_PROMPT_PREFIX
+            self._format_instructions: str = kwargs.get('format_instructions', FORMAT_INSTRUCTIONS)
+        self._capabilities: str = kwargs.get('capabilities', None)
+        self.name = name or "Pandas Agent"
+        self.description = "A simple agent that uses the pandas library to perform data analysis tasks."
+        self._static_path = BASE_DIR.joinpath('static')
+        self.agent_report_dir = self._static_path.joinpath('reports', 'agents')
+        super().__init__(
+            name=name,
+            llm=llm,
+            system_prompt=system_prompt,
+            human_prompt=human_prompt,
+            tools=tools,
+            agent_type=self.agent_type,
+            **kwargs
+        )
+
+    def _define_dataframe(
+        self,
+        df: Union[list[pd.DataFrame], Dict[str, pd.DataFrame]]
+    ) -> Dict[str, pd.DataFrame]:
+        """Define the dataframe."""
+        _df = {}
+        if isinstance(df, pd.DataFrame):
+            # if data is one single dataframe
+            _df['df1'] = df
+        elif isinstance(df, list):
+            # if data is a list of dataframes
+            for i, dataframe in enumerate(df):
+                df_name = f"df{i + 1}"
+                _df[df_name] = dataframe.copy()
+        elif isinstance(df, pd.Series):
+            # if data is a pandas series
+            # convert it to a dataframe
+            df = pd.DataFrame(df)
+            _df['df1'] = df
+        elif isinstance(df, dict):
+            _df = df
+        else:
+            raise ValueError(
+                f"Expected pandas DataFrame, got {type(df)}"
+            )
+        return _df
+
+    def get_query(self) -> Union[List[str], dict]:
+        """Get the query."""
+        return self._queries
+
+    def get_capabilities(self) -> str:
+        """Get the capabilities of the agent."""
+        return self._capabilities
+
+    def pandas_agent(self, **kwargs):
+        """
+        Creates a Pandas Agent.
+
+        This agent uses reasoning and tool execution iteratively to generate responses.
+
+        Returns:
+            RunnableMultiActionAgent: A Pandas-based agent.
+
+        ✅ Use Case: Best for decision-making and reasoning tasks where the agent must break problems down into multiple steps.
+
+        """
+        # Create the pandas agent
+        dfs = list(self.df.values()) if isinstance(self.df, dict) else self.df
+        # bind the tools to the LLM:
+        llm = self._llm.bind_tools(self.tools)
+        return create_pandas_dataframe_agent(
+            llm,
+            dfs,
+            verbose=True,
+            agent_type=self.agent_type,
+            allow_dangerous_code=True,
+            # prefix=self._prompt_prefix,
+            max_iterations=10,
+            extra_tools=self.tools,
+            agent_executor_kwargs={"memory": self.memory, "handle_parsing_errors": True},
+            return_intermediate_steps=True,
+            **kwargs
+        )
+
+    async def configure(self, df: pd.DataFrame = None, app=None) -> None:
+        """Basic Configuration of Pandas Agent.
+        """
+        if app:
+            if isinstance(app, web.Application):
+                self.app = app  # register the app into the Extension
+            else:
+                self.app = app.get_app()  # Nav Application
+        # adding this configured chatbot to app:
+        if self.app:
+            self.app[f"{self.name.lower()}_bot"] = self
+        if df is not None:
+            self.df = self._define_dataframe(df)
+        # Configure LLM:
+        self.configure_llm(use_chat=True)
+        # Configure VectorStore if enabled:
+        if self._use_vector:
+            self.configure_store()
+        # Conversation History:
+        self.memory = self.get_memory(input_key="input", output_key="output")
+        # 1. Initialize the Agent (as the base for RunnableMultiActionAgent)
+        self.agent = self.pandas_agent()
+        # 2. Create Agent Executor - This is where we typically run the agent.
+        self._agent = self.agent
+        # 3. When agent is correctly created, caching the data:
+        await self._cache_data(
+            self.chatbot_id,
+            self.df,
+            cache_expiration=24
+        )
+
+    def mimefromext(self, ext: str) -> str:
+        """Get the mime type from the file extension."""
+        mime_types = {
+            '.csv': 'text/csv',
+            '.xlsx': 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet',
+            '.xls': 'application/vnd.ms-excel',
+            '.json': 'application/json',
+            '.txt': 'text/plain',
+            '.html': 'text/html',
+            '.htm': 'text/html',
+            '.pdf': 'application/pdf',
+            '.png': 'image/png',
+            '.jpg': 'image/jpeg',
+            '.jpeg': 'image/jpeg',
+            '.gif': 'image/gif',
+            '.svg': 'image/svg+xml',
+            '.md': 'text/markdown',
+            '.ogg': 'audio/ogg',
+            '.wav': 'audio/wav',
+            '.mp3': 'audio/mpeg',
+            '.mp4': 'video/mp4',
+        }
+        return mime_types.get(ext, None)
+
+    def extract_filenames(self, response: AgentResponse) -> List[Path]:
+        """Extract filenames from the content."""
+        # Split the content by lines
+        output_lines = response.output.splitlines()
+        current_filename = ""
+        filenames = {}
+        for line in output_lines:
+            if 'filename:' in line:
+                current_filename = line.split('filename:')[1].strip()
+                if current_filename:
+                    try:
+                        filename_path = Path(current_filename).resolve()
+                        if filename_path.is_file():
+                            content_type = self.mimefromext(filename_path.suffix)
+                            url = str(filename_path).replace(str(self._static_path), BASE_STATIC_URL)
+                            filenames[filename_path.name] = {
+                                'content_type': content_type,
+                                'file_path': filename_path,
+                                'filename': filename_path.name,
+                                'url': url
+                            }
+                        continue
+                    except AttributeError:
+                        pass
+        if filenames:
+            response.filename = filenames
+
+    async def invoke(
+        self,
+        query: str, llm: Optional[Any] = None,
+        llm_params: Optional[Dict[str, Any]] = None
+    ):
+        """Invoke the agent with optional LLM override.
+
+        Args:
+            query (str): The query to ask the chatbot.
+            llm (Optional[Any]): Optional LLM to use for this specific invocation.
+            llm_params (Optional[Dict[str, Any]]): Optional parameters to modify LLM behavior
+                                                (temperature, max_tokens, etc.)
+
+        Returns:
+            str: The response from the chatbot.
+
+        """
+        original_agent = None
+        original_llm = None
+        try:
+            # If a different LLM or parameters are provided, create a temporary agent
+            if llm is not None:
+                # Get the current LLM if we're just updating parameters
+                current_llm = llm if llm is not None else self._llm
+                # Store original LLM for reference
+                original_llm = self._llm
+                # Store original agent for reference
+                original_agent = self._agent
+                # Temporarily update the instance LLM
+                self._llm = current_llm
+                # Create a new agent with the updated LLM
+                self._agent = self.pandas_agent()
+            # Invoke the agent with the query
+            result = await self._agent.ainvoke(
+                {"input": query}
+            )
+        except Exception as e:
+            return None, e
+        try:
+            # Parse tool outputs if present
+            try:
+                if isinstance(result, dict):
+                    output = result.get('output', '')
+                    if '```tool_outputs' in output:
+                        tool_json_match = re.search(r'```tool_outputs\n(.*?)\n```', output, re.DOTALL)
+                        if tool_json_match:
+                            tool_json = tool_json_match.group(1)
+                            # Parse the JSON
+                            tool_data = json_decoder(tool_json)
+                            # Get the actual content
+                            if "python_repl_ast_response" in tool_data:
+                                python_response = tool_data["python_repl_ast_response"]
+                                if isinstance(python_response, dict) and "content" in python_response:
+                                    # Replace the output with just the content
+                                    result['output'] = python_response["content"].strip()
+            except Exception as parse_error:
+                self.logger.error(
+                    f"Error parsing tool output: {parse_error}"
+                )
+            response = AgentResponse(question=query, **result)
+            # check if return is a file:
+            try:
+                self.extract_filenames(response)
+            except Exception as exc:
+                self.logger.error(
+                    f"Unable to extract filenames: {exc}"
+                )
+            # Restore the original agent if any:
+            if original_llm is not None:
+                self._llm = original_llm
+                self._agent = original_agent
+            try:
+                return self.as_markdown(
+                    response
+                ), response
+            except Exception as exc:
+                self.logger.exception(
+                    f"Error on response: {exc}"
+                )
+                return result.get('output', None), None
+        except Exception as e:
+            return result, e
+
+    def _configure_python_tool(self, df_locals: dict, **kwargs) -> PythonAstREPLTool:
+        """Configure the Python tool."""
+        # Create the Python tool with the given locals and globals
+        df_locals['execution_results'] = {}
+        # Create the Python REPL tool
+        PythonAstREPLTool_init = PythonAstREPLTool.__init__
+
+        def PythonAstREPLTool_init_wrapper(self, *args, **kwargs):
+            PythonAstREPLTool_init(self, *args, **kwargs)
+            self.globals = self.locals
+
+        PythonAstREPLTool.__init__ = PythonAstREPLTool_init_wrapper
+        python_tool = PythonAstREPLTool(
+            locals=df_locals,
+            globals=kwargs.get('globals', {}),
+            verbose=True,
+            **kwargs
+        )
+
+        # Add essential library imports and helper functions
+        setup_code = """
+# Ensure essential libraries are imported
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import seaborn as sns
+from collections import Counter, defaultdict
+from parrot.bots.tools import quick_eda, generate_eda_report, list_available_dataframes, create_plot, generate_pdf_from_html
+
+# Set plotting style
+plt.style.use('seaborn-v0_8-whitegrid')
+sns.set_palette('Set2')
+
+# Verify pandas is loaded correctly
+print(f"Pandas version: {pd.__version__}")
+"""
+        try:
+            python_tool.run(setup_code)
+        except Exception as e:
+            self.logger.error(
+                f"Error setting up python tool: {e}"
+            )
+        print(':: PYTHON TOOL > ', python_tool)
+        return python_tool
+
+    def _metrics_guide(self, df_key: str, df_name: str, columns: list) -> str:
+        """Generate a guide for the dataframe columns."""
+        # Create a markdown table with column category, column name, type and with dataframe is present:
+        table = "\n| Category | Column Name | Type | DataFrame | Dataframe Name |\n"
+        table += "|------------------|-------------|------|-----------|\n"
+        for column in columns:
+            # Get the column name
+            column_name = column
+            # split by "_" and first element is the category (if any):
+            # Get the column category
+            try:
+                column_category = column.split('_')[0]
+            except IndexError:
+                column_category = df_name
+            # Get the type of the column
+            column_type = str(self.df[df_name][column].dtype)
+            # Add the row to the table
+            table += f"| {column_category} | {column_name} | {column_type} | {df_key} | {df_name} |\n\n\n"
+        # Add a note about the dataframe
+        table += f"\nNote: {df_key} is also available as {df_name}\n"
+        return table
+
+    def define_prompt(self, prompt, **kwargs):
+        now = datetime.now(timezone.utc).strftime("%Y-%m-%d")
+        self.agent_report_dir = self._static_path.joinpath(str(self.chatbot_id))
+        if self.agent_report_dir.exists() is False:
+            self.agent_report_dir.mkdir(parents=True, exist_ok=True)
+        # Add dataframe information
+        num_dfs = len(self.df)
+        self.df_locals['agent_report_dir'] = self.agent_report_dir
+        df_info = ''
+        for i, (df_name, df) in enumerate(self.df.items()):
+            df_key = f"df{i + 1}"
+            self.df_locals[df_name] = df
+            self.df_locals[df_key] = df
+            row_count = len(df)
+            self.df_locals[f"{df_key}_row_count"] = row_count
+            # Get basic dataframe info
+            df_shape = f"DataFrame Shape: {df.shape[0]} rows × {df.shape[1]} columns"
+            df_columns = self._metrics_guide(df_key, df_name, df.columns.tolist())
+            # Generate summary statistics
+            summary_stats = brace_escape(df.describe(include='all').to_markdown())
+            df_head = brace_escape(df.head(4).to_markdown())
+            # Create df_info block
+            if self.agent_type == "tool-calling":
+                df_info += f"""
+                ## Dataframe Name: {df_key}:
+                ## DataFrame: {df_key} (also accessible as {df_name} in Python code)
+
+                **Shape {df_key}**: {df_shape}
+
+                **Column Details (Name, Type, Category)**:
+                {df_columns}
+
+                **First 4 Rows (`print({df_key}.head(4).to_markdown())`)**:
+                {df_head}
+
+                **Summary Statistics (`print({df_key}.describe(include='all').to_markdown())`)**:
+                {summary_stats}
+                """
+            else:
+                df_info += f"""
+                ## DataFrame {df_key}:
+
+                ### {df_key} Shape:
+                {df_shape}
+
+                ### {df_key} Info:
+                {df_columns}
+
+                ### {df_key} Summary Statistics:
+                {summary_stats}
+                """
+        # Configure Python tool:
+        python_tool = self._configure_python_tool(
+            df_locals=self.df_locals,
+            **kwargs
+        )
+        # Add the Python tool to the tools list
+        self.tools.append(python_tool)
+        # List of Tools:
+        list_of_tools = ""
+        for tool in self.tools:
+            name = tool.name
+            description = tool.description  # noqa  pylint: disable=E1101
+            list_of_tools += f'- {name}: {description}\n'
+        list_of_tools += "\n"
+        tools_names = [tool.name for tool in self.tools]
+        capabilities = ''
+        if self._capabilities:
+            capabilities = "**Your Capabilities:**\n"
+            capabilities += self.sanitize_prompt_text(self._capabilities) + "\n"
+        # Create the prompt
+        sanitized_backstory = ''
+        if self.backstory:
+            sanitized_backstory = self.sanitize_prompt_text(self.backstory)
+        tmpl = Template(self._prompt_template)
+        self._prompt_prefix = tmpl.safe_substitute(
+            name=self.name,
+            description=self.description,
+            list_of_tools=list_of_tools,
+            backstory=sanitized_backstory,
+            capabilities=capabilities,
+            today_date=now,
+            system_prompt_base=prompt,
+            tools=", ".join(tools_names),
+            format_instructions=self._format_instructions.format(
+                tool_names=", ".join(tools_names)),
+            df_info=df_info,
+            num_dfs=num_dfs,
+            rationale=self.rationale,
+            agent_report_dir=self.agent_report_dir,
+            **kwargs
+        )
+        if self.agent_type == "tool-calling":
+            tmpl = Template(TOOL_CALLING_PROMPT_SUFFIX)
+            self._prompt_suffix = tmpl.safe_substitute(
+                df_info=df_info,
+                num_dfs=num_dfs
+            )
+        else:
+            self._prompt_suffix = AGENT_PROMPT_SUFFIX
+
+    def default_backstory(self) -> str:
+        return "You are a helpful assistant built to provide comprehensive guidance and support on data calculations and data analysis working with pandas dataframes."
+
+    @staticmethod
+    async def call_qs(queries: list) -> Dict[str, pd.DataFrame]:
+        """
+        call_qs.
+        description: Call the QuerySource queries.
+
+        This method is used to execute multiple queries and files on the QueryObject.
+        It returns a dictionary with the results.
+        """
+        dfs = {}
+        for query in queries:
+            if not isinstance(query, str):
+                raise ValueError(
+                    f"Query {query} is not a string."
+                )
+            # now, the only query accepted is a slug:
+            try:
+                qy = QS(
+                    slug=query
+                )
+                df, error = await qy.query(output_format='pandas')
+                if error:
+                    raise ValueError(
+                        f"Query {query} fail with error {error}."
+                    )
+                if not isinstance(df, pd.DataFrame):
+                    raise ValueError(
+                        f"Query {query} is not returning a dataframe."
+                    )
+                dfs[query] = df
+            except ValueError:
+                raise
+            except Exception as e:
+                raise ValueError(
+                    f"Error executing Query {query}: {e}"
+                )
+        return dfs
+
+    @staticmethod
+    async def call_multiquery(query: dict) -> Dict[str, pd.DataFrame]:
+        """
+        call_multiquery.
+        description: Call the MultiQuery queries.
+
+        This method is used to execute multiple queries and files on the QueryObject.
+        It returns a dictionary with the results.
+        """
+        data = {}
+        _queries = query.pop('queries', {})
+        _files = query.pop('files', {})
+        if not _queries and not _files:
+            raise ValueError(
+                "Queries or files are required."
+            )
+        try:
+            ## Step 1: Running all Queries and Files on QueryObject
+            qs = MultiQS(
+                slug=[],
+                queries=_queries,
+                files=_files,
+                query=query,
+                conditions=data,
+                return_all=True
+            )
+            result, _ = await qs.execute()
+        except Exception as e:
+            raise ValueError(
+                f"Error executing MultiQuery: {e}"
+            )
+        if not isinstance(result, dict):
+            raise ValueError(
+                "MultiQuery is not returning a dictionary."
+            )
+        # MultiQuery returns a dictionary with the results
+        return result
+
+    @classmethod
+    async def gen_data(
+        cls,
+        query: Union[list, dict],
+        agent_name: Optional[str] = None,
+        refresh: bool = False,
+        cache_expiration: int = 48,
+        no_cache: bool = False
+    ) -> Dict[str, pd.DataFrame]:
+        """
+        gen_data.
+
+        Generate the dataframes required for the agent to work, with Redis caching support.
+
+        Parameters:
+        -----------
+        query : Union[list, dict]
+            The query or queries to execute to generate dataframes.
+        refresh : bool
+            If True, forces regeneration of dataframes even if cached versions exist.
+        cache_expiration_hours : int
+            Number of hours to keep the cached dataframes (default: 48).
+        no_cache : bool
+            If True, disables caching even if no agent_name is provided.
+
+        Returns:
+        --------
+        Dict[str, pd.DataFrame]
+            A Dictionary of named pandas DataFrames generated from the queries.
+        """
+        # If agent_name is provided, we'll use Redis caching
+        if not agent_name:
+            agent_name = cls.chatbot_id
+
+        if not refresh:
+            # Try to get cached dataframes
+            cached_dfs = await cls._get_cached_data(agent_name)
+            if cached_dfs:
+                return cached_dfs
+
+        # Generate dataframes from query if no cache exists or refresh is True
+        dfs = await cls._execute_query(query)
+
+        # If agent_name is provided, cache the generated dataframes
+        if no_cache is False:
+            await cls._cache_data(agent_name, dfs, cache_expiration)
+        return dfs
+
+    @classmethod
+    async def _execute_query(cls, query: Union[list, dict]) -> Dict[str, pd.DataFrame]:
+        """Execute the query and return the generated dataframes."""
+        if isinstance(query, dict):
+            # is a MultiQuery execution, use the MultiQS class engine to do it:
+            try:
+                return await cls.call_multiquery(query)
+            except ValueError as e:
+                raise ValueError(f"Error creating Query For Agent: {e}")
+        elif isinstance(query, (str, list)):
+            if isinstance(query, str):
+                query = [query]
+            try:
+                return await cls.call_qs(query)
+            except ValueError as e:
+                raise ValueError(f"Error creating Query For Agent: {e}")
+        else:
+            raise ValueError(
+                f"Expected a list of queries or a dictionary, got {type(query)}"
+            )
+
+    @classmethod
+    async def _get_redis_connection(cls):
+        """Get a connection to Redis."""
+        # You should adjust these parameters according to your Redis configuration
+        # Consider using environment variables for these settings
+        return await aioredis.Redis.from_url(
+            REDIS_HISTORY_URL,
+            decode_responses=True
+        )
+
+    @classmethod
+    async def _get_cached_data(cls, agent_name: str) -> Optional[Dict[str, pd.DataFrame]]:
+        """
+        Retrieve cached data from Redis if they exist.
+
+        Returns None if no cache exists or on error.
+        """
+        try:
+            redis_conn = await cls._get_redis_connection()
+            # Check if the agent key exists
+            key = f"agent_{agent_name}"
+            if not await redis_conn.exists(key):
+                await redis_conn.close()
+                return None
+
+            # Get all dataframe keys stored for this agent
+            df_keys = await redis_conn.hkeys(key)
+            if not df_keys:
+                await redis_conn.close()
+                return None
+
+            # Retrieve and convert each dataframe
+            dataframes = {}
+            for df_key in df_keys:
+                df_json = await redis_conn.hget(key, df_key)
+                if df_json:
+                    # Convert from JSON to dataframe
+                    df_data = json_decoder(df_json)
+                    df = pd.DataFrame.from_records(df_data)
+                    dataframes[df_key] = df
+
+            await redis_conn.close()
+            return dataframes if dataframes else None
+
+        except Exception as e:
+            # Log the error but continue execution without cache
+            print(f"Error retrieving cache: {e}")
+            return None
+
+    @classmethod
+    async def _cache_data(
+        cls,
+        agent_name: str,
+        dataframes: Dict[str, pd.DataFrame],
+        cache_expiration: int
+    ) -> None:
+        """
+        Cache the given dataframes in Redis.
+
+        The dataframes are stored as JSON records under a hash key named after the agent.
+        """
+        try:
+            if not dataframes:
+                return
+
+            redis_conn = await cls._get_redis_connection()
+            key = f"agent_{agent_name}"
+
+            # Delete any existing cache for this agent
+            await redis_conn.delete(key)
+            hkeys = await redis_conn.hkeys(key)
+            if hkeys:
+                await redis_conn.hdel(key, *hkeys)
+
+            # Store each dataframe under the agent's hash
+            for df_key, df in dataframes.items():
+                # Convert DataFrame to JSON
+                df_json = json_encoder(df.to_dict(orient='records'))
+                await redis_conn.hset(key, df_key, df_json)
+
+            # Set expiration time
+            expiration = timedelta(hours=cache_expiration)
+            await redis_conn.expire(key, int(expiration.total_seconds()))
+
+            logging.info(
+                f"Data was cached for agent {agent_name} with expiration of {cache_expiration} hours"
+            )
+
+            await redis_conn.close()
+
+        except Exception as e:
+            # Log the error but continue execution
+            print(f"Error caching dataframes: {e}")