Closed adnanbhanji closed 6 months ago
For this issue, we are planning on adding a flow control mechanism to decide whether to proceed with report generation or consult LLaMA-2 for an answer. We have to incorporate logic to handle the "start report" command specifically, while other inputs are processed through LLaMA-2.
I am pleased to report that the necessary setup procedures have been completed on my computer, enabling seamless integration between 'llama2' and WhatsApp. As a result, the initial steps of the conversation flow have been initiated successfully.
Modified msgrcvd function so that when a message is received, and there's an ongoing conversation (indicated by the presence of the user's number in user_states), it treats the message as an answer to the current question and processes the state transition accordingly.
Managed to ask a question during the conversation flow Modified the msgrcvd function: Added a new function to call the llama model when a question is asked:
Modified the code to output a summary of the conversation flow at the end:
Modified existing process_answer function tu handle the summary generation:
Created a function to prepare the summary:
modified funciton to create a pdf: pdf generated:
Great work, passes tests and achieves purpose.
Modify the conversation flows in our chatbot to accommodate LLama's response format and input requirements.