main.py

import logfire
import streamlit as st

from app import agent, models

logfire.configure(
    service_name="streamlit", scrubbing=False, send_to_logfire="if-token-present"
)
logfire.instrument_pydantic_ai()


def format_output(statements: list[models.Statement]) -> str:
    sentences = []
    citations = {}

    for statement in statements:
        sentence = statement.text

        if sentence.startswith(("*", "-")):
            # Bullet points should be on a newline.
            sentence = f"\n{sentence}"

        if statement.sources:
            citation_ids = []
            for source in statement.sources:
                if not (citation_id := citations.get(source.citation)):
                    citation_id = len(citations) + 1
                    citations[source.citation] = citation_id
                citation_ids.append(citation_id)
            sentence += " ".join(f"[^{i}]" for i in sorted(citation_ids))

        sentences.append(sentence)

    answer = " ".join(sentences)
    footnotes = "\n".join(f"[^{id}]: {citation}" for citation, id in citations.items())
    return f"{answer}\n\n{footnotes}"


def main():
    st.title("Elna")
    with st.form("search", border=False):
        query = st.text_input("Your medical question")
        submit = st.form_submit_button("Ask")

        if submit:
            with st.spinner("Thinking...", show_time=True):
                output = agent.agent.run_sync(query).output
                answer = format_output(output.statements)
            with st.expander("Thinking Process"):
                st.markdown(output.thoughts)
            st.markdown(answer)


if __name__ == "__main__":
    main()