Experience the power of AI and Speckle using pandasai, OpenAI, specklepy and streamlit. This feature was built on top of a previous web app created in a speckle video tutorial called "Create Schedule from Revit Data", which you can follow as well.
Otherwise, you can just download/clone this repository, run streamlit, add your own OpenAI key and write queries to chatSpeckle.
What was built on top of "Create Schedule from Revit Data" web app?
- Library
from pandasai import PandasAI
from pandasai.llm.openai import OpenAI- Function
def chat_speckle(df,prompt):
llm = OpenAI(api_token = OPENAI_API_KEY)
pandas_ai = PandasAI(llm, conversational=False)
result = pandas_ai.run(df, prompt=prompt)
print(result)
return result- chatSpeckle feature
col1, col2 = st.columns([1,1])
with col1:
result = st.dataframe(result_DF)
with col2:
st.info("⬇️chatSpeckle⬇️")
OPENAI_API_KEY = st.text_input('OpenAI key',"sk-...vDlY")
input_text = st.text_area('Enter your query')
if input_text is not None:
if st.button ("Send"):
st.info('Your query:'+input_text)
result = chat_with_speckle(result_DF, input_text)
st.success(result)