Untitled

meeting_invite_prompt = PromptTemplate(
    template="""
You are a system that parses a user's meeting request: {query}

Your task is to construct a valid response that addresses all the required factors:
1. "data_issues"
2. "meeting_start"
3. "meeting_end"
4. "attendees"

Requirements:
- "data_issues":
  - If there's ambiguity about which user to add (i.e., two or more similar names in user_to_email map {user_to_email}), provide a clarification message.
  - If a single similar entry exists (e.g., the user said 'john' but only 'john_doe' exists in user_to_email), automatically add 'john_doe' to the attendees.
  - If a requested user is not found in user_to_email, provide an error message.
  - Provide messages only if there are issues; otherwise, do not include this factor in the response.
  - If the meeting start and end times cannot be inferred from {query} and {current_time}, ask for clarifications regarding timezone, date, and meeting duration.

- "meeting_start" / "meeting_end":
  - The current timestamp is {current_time}, which includes timezone information.
  - Based on the user's query, calculate the appropriate meeting start and end times.
  - Format for meeting_start and meeting_end should match the format of current_time: {current_time}.

- "attendees":
  - A list of attendee email addresses based on the user_to_email mapping.

If "data_issues" are present, the response should prioritize addressing these issues over other factors.
""",
    input_variables=["query", "user_to_email", "current_time"],
)
chain = meeting_invite_prompt | llm | parser