| from langchain_core.prompts import PromptTemplate | |
| router_instruction_prompt_1=""" | |
| You are tasked with analyzing user input to identify the most relevant keys from a dataset of user information. | |
| Each key provides specific details about the user. | |
| Your goal is to understand the query and conversation context to select the appropriate keys from which the user's data can help in providing accurate, personalised and actionable solutions. | |
| 1. name: The user's full name. | |
| 2. age: The user's age. | |
| 3. gender: The user's gender. | |
| 4. preferences: These are user's preferences. | |
| 5. personalInformation: These are user's personal information details. | |
| 6. relatedDemographics: Information pertaining to the user's geographic location and other demographic factors. | |
| 7. history: A record of the user's recent activities, including search history, recent purchases, places visited, and other relevant actions or events. | |
| 8. painPoints: The user's complaints, challenges, or issues they face. | |
| 9. inefficienciesOrQualityOfLifeImprovements: Areas where the user seeks improvement, along with potential solutions to increase efficiency or enhance their quality of life. | |
| 10. additionalNotes: Any extra information that might be relevant or helpful about the user. | |
| Instructions: | |
| - Analyze the user’s query and previous message to understand its purpose, context, and specific needs. | |
| - Determine which keys (one or multiple) are most relevant to extract the required information to provide a comprehensive and actionable response. | |
| - If no additional information is needed or no relevant details are found, route to None. | |
| - Pay special attention to personalInformation if the user's financial situation or budget is relevant to the decision-making process. | |
| Query: | |
| {query} | |
| Previous Messages: | |
| {previous_messages} | |
| Format Instructions: | |
| {format_instructions} | |
| """ | |
| summary_prompt_instructions_1=""" | |
| Given the following JSON data,query and previous messages condense all the information into a summary that captures all the details that may be relevant to providing solutions for the user query. | |
| Include in the summary only those points that are relevant to the query or conversation context. If nothing was found, return "" | |
| Query: | |
| {query} | |
| Previous Messages: | |
| {previous_messages} | |
| Json Data: | |
| {data} | |
| Format Instructions: | |
| {format_instructions} | |
| """ | |
| router_instruction_prompt_2=PromptTemplate.from_template(""" | |
| You are tasked with analyzing user input to identify the most relevant keys from a dataset of user information. | |
| Each key provides specific details about the user. | |
| Your goal is to understand the query and conversation context to select the appropriate keys from which the user's data can help in providing accurate, personalised and actionable solutions. | |
| 1. name: The user's full name. | |
| 2. age: The user's age. | |
| 3. gender: The user's gender. | |
| 4. preferences: These are user's preferences. | |
| 5. personalInformation: These are user's personal information details. | |
| 6. relatedDemographics: Information pertaining to the user's geographic location and other demographic factors. | |
| 7. history: A record of the user's recent activities, including search history, recent purchases, places visited, and other relevant actions or events. | |
| 8. painPoints: The user's complaints, challenges, or issues they face. | |
| 9. inefficienciesOrQualityOfLifeImprovements: Areas where the user seeks improvement, along with potential solutions to increase efficiency or enhance their quality of life. | |
| 10. additionalNotes: Any extra information that might be relevant or helpful about the user. | |
| Instructions: | |
| - Analyze the user’s query and previous message to understand its purpose, context, and specific needs. | |
| - Determine which keys (one or multiple) are most relevant to extract the required information to provide a comprehensive, personalised and actionable response. | |
| - If no additional information is needed or no relevant details are found, route to None. | |
| - Pay special attention to personalInformation if the user's financial situation or budget is relevant to the decision-making process. | |
| Query: | |
| {query} | |
| Previous Messages: | |
| {previous_messages} | |
| """) | |
| extract_prompt_instructions_2=PromptTemplate.from_template(""" | |
| Given the following JSON data,query and previous messages extract the information that may be relevant to providing solutions for the user query. | |
| Instructions: | |
| - Include in the summary only those points that are relevant to the query or conversation context. | |
| - If nothing relevant was found, return "". | |
| - Do not provide anything in your own words. | |
| Query: | |
| {query} | |
| Previous Messages: | |
| {previous_messages} | |
| Json Data: | |
| {data} | |
| """) | |
| RESPONSE_PROMPT="""You are helpful assistant, Based on the User information and the query provide a response accordingly. | |
| Do not Fabricate or provide false details | |
| User information: | |
| {user_information} | |
| Query: | |
| {query} | |
| """ |