First commit

app.py

@@ -0,0 +1,165 @@
+import streamlit as st
+import sqlite3
+import pandas as pd
+import openai
+import os
+
+from langchain.embeddings.openai import OpenAIEmbeddings
+from langchain.chat_models import ChatOpenAI
+from langchain.chains.question_answering import load_qa_chain
+from langchain.text_splitter import CharacterTextSplitter
+from langchain.vectorstores import Chroma
+
+os.environ["OPENAI_API_KEY"] = os.getenv("SECRET_KEY")
+
+def init_database():
+    conn = sqlite3.connect('GPTPromptTemplates.db')
+    cursor = conn.cursor()
+
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS USERS (
+            USER_ID INTEGER PRIMARY KEY,
+            User_Name VARCHAR(255)
+        )
+    ''')
+
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS TEMPLATES (
+            TEMPLATE_ID INTEGER PRIMARY KEY,
+            USER_ID INTEGER,
+            Prompt_Name VARCHAR(255),
+            Prompt_Text TEXT
+        )
+    ''')
+
+    cursor.execute('''
+        CREATE UNIQUE INDEX IF NOT EXISTS idx_templates_prompt_name ON TEMPLATES (USER_ID, Prompt_Name)
+    ''')
+
+    conn.commit()
+    conn.close()
+
+def insert_prompt_template(user_id, prompt_name, prompt_text):
+    conn = sqlite3.connect('GPTPromptTemplates.db')
+    cursor = conn.cursor()
+    cursor.execute('INSERT OR REPLACE INTO TEMPLATES (USER_ID, Prompt_Name, Prompt_Text) VALUES (?, ?, ?)', (user_id, prompt_name, prompt_text))
+    conn.commit()
+    conn.close()
+
+def delete_prompt_template(user_id, prompt_name):
+    conn = sqlite3.connect('GPTPromptTemplates.db')
+    cursor = conn.cursor()
+    cursor.execute('DELETE FROM TEMPLATES WHERE USER_ID = ? AND prompt_name = ?', (user_id, prompt_name))
+    conn.commit()
+    conn.close()
+
+def get_prompt(user_id, prompt_name):
+    conn = sqlite3.connect('GPTPromptTemplates.db')
+    cursor = conn.cursor()
+    cursor.execute('SELECT Prompt_Name, Prompt_Text FROM TEMPLATES WHERE Prompt_Name = ? AND USER_ID = ?', (prompt_name, user_id))
+    template = cursor.fetchone()
+    conn.close()
+
+    if template == None:
+      return '',''
+    else:
+      return template[0], template[1]
+
+def get_default_prompt(user_id):
+    conn = sqlite3.connect('GPTPromptTemplates.db')
+    cursor = conn.cursor()
+    cursor.execute('SELECT Prompt_Name, Prompt_Text FROM TEMPLATES WHERE USER_ID = ? ORDER BY Prompt_Name ASC LIMIT 1', (user_id, ))
+    template = cursor.fetchone()
+    conn.close()
+
+    if template == None:
+      return '',''
+    else:
+      return template[0], template[1]
+
+def get_prompt_list(user_id):
+    conn = sqlite3.connect('GPTPromptTemplates.db')
+    templates = pd.read_sql_query('SELECT DISTINCT Prompt_Name FROM TEMPLATES WHERE USER_ID = {} ORDER BY Prompt_Name ASC'.format(user_id), conn)
+    conn.commit()
+    conn.close()
+
+    return templates
+
+def template_change_value():
+    name, prompt = get_prompt(st.session_state.user_id, st.session_state.template_select)
+    st.session_state.name = name
+    st.session_state.prompt = prompt
+
+def template_return_value(template_name):
+    st.session_state.template_select = template_name
+    name, prompt = get_prompt(st.session_state.user_id, st.session_state.template_select)
+    st.session_state.name = name
+    st.session_state.prompt = prompt
+
+def main():
+    st.title("Working with Chat GPT with templates")
+
+    init_database()
+
+    col1, col2, col3 = st.columns([1,1,1])
+
+    user_id = 1
+    name, prompt = get_default_prompt(user_id)
+    prompt_list = get_prompt_list(user_id)
+    model_names = ['gpt-4','gpt-3.5-turbo','gpt-3.5-turbo-16k']
+
+    if not "initialized" in st.session_state:
+      st.session_state.user_id = user_id
+      st.session_state.name = name
+      st.session_state.prompt = prompt
+      st.session_state.prompt_list = prompt_list
+      st.session_state.template_select = name
+      st.session_state.output = ''
+      st.session_state.model_name = 'gpt-4'
+      st.session_state.initialized = True
+
+    with col1:
+
+      input_text = st.text_area('Please insert data for transforming', '', key="input_data", height=370)
+
+      if st.button("Apply"):
+        query = prompt
+        st.write("in progress")
+        text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
+        texts = text_splitter.split_text(input_text)
+        embeddings = OpenAIEmbeddings()
+        docsearch = Chroma.from_texts(texts, embeddings, metadatas=[{"source": str(i)} for i in range(len(texts))]).as_retriever()
+        docs = docsearch.get_relevant_documents(query)
+        chain = load_qa_chain(ChatOpenAI(model = st.session_state.model_name,temperature=0), chain_type="stuff")
+        st.session_state.output = chain.run(input_documents=docs, question=query)
+        #st.session_state["output"] = output
+        #col3.text_area('Result', value=output, key="output_data", height=450)
+        st.experimental_rerun()
+        st.success("Ready!")
+
+
+    with col2:
+      st.session_state.model_name = st.selectbox("GPT model: ",model_names, key="gpt_model")
+      template_return_value(st.selectbox("Template: ",st.session_state.prompt_list, key="prompt_template",))
+      new_name = st.text_input("Template name:",value=st.session_state.name, key="template_name")
+      input_query = st.text_area("Prompt:",value=st.session_state.prompt, key="template_text", height=200)
+
+      col4, col5 = st.columns([1,1])
+      if col4.button("Save"):
+          insert_prompt_template(user_id, new_name, input_query)
+          st.session_state.prompt_list = get_prompt_list(user_id)
+          st.success("Prompt saved!")
+          st.experimental_rerun()
+
+      if col5.button("Delete"):
+          delete_prompt_template(user_id, new_name)
+          st.session_state.prompt_list = get_prompt_list(user_id)
+          st.success("Prompt deleted!")
+          st.experimental_rerun()          
+
+    with col3:
+      txt_result = st.text_area('Result', value=st.session_state.output, key="output_data", height=450)
+
+
+if __name__ == "__main__":
+    main()

requirements.txt

@@ -0,0 +1,8 @@
+numpy
+pandas
+streamlit
+langchain 
+openai 
+chromadb 
+tiktoken
+sqlite3