Closed KenjiPcx closed 9 months ago
@thatjosh I see this code that already done this
# Generate sugggestions for new topics in subsequent feynman sessions
prompt = ChatPromptTemplate.from_template("Suggest 5 topics that are {topic_type} for a student to learn about, based on the content the student explored in the session. Return just the topics as a list of strings (the output should be parsable by json.loads in python) and nothing extra. \n Questions, score and concept explored in current learning session: \n {content}")
output_parser = StrOutputParser()
chain = prompt | langchain_llm | output_parser
topics_str = chain.invoke({"content": content, "topic_type": topic_type})
topic_list = json.loads(topics_str)
logging.info(topic_list)
Can you confirm and close?
Yup yup
Issue details
Similar to #22, this is what happens when a user passes or excels their Feynman session, instead of suggesting them Feynman sessions for their knowledge gaps, we will focus on suggesting sessions for similar topics