Spaces:
Sleeping
Sleeping
| import openai | |
| import gradio as gr | |
| import json | |
| import os | |
| os.environ['OPENAI_API_KEY'] = os.getenv("openaikey") | |
| def get_completion_messages(messages, model="gpt-4o-mini", temperature=0, max_tokens =500, presence_penalty=0, seed =None, stream=False): | |
| response = openai.chat.completions.create( | |
| model=model, | |
| messages=messages, | |
| temperature=temperature, | |
| max_tokens=max_tokens, | |
| presence_penalty=presence_penalty, | |
| seed=seed, | |
| stream = stream | |
| ) | |
| return response.choices[0].message.content | |
| delimiter = "####" | |
| AV_system_message = f""" | |
| You are Anu Madan, the learning solutions Manager at Analytics Vidhya, an ed-tech platform for Data Science, AI, and Generative AI.\ | |
| You focus on business development for clients in the US. | |
| You will be provided with email chains between you and potential customers. \ | |
| The emails will be delimited with \ | |
| {delimiter} characters. | |
| You have to perform the following tasks by going through the emails.\ | |
| 1. Classify the email into a category. | |
| 2. Share information about the prospective client such as name, designation, and company. | |
| 3. Mention the date of last communication with the client | |
| 4. Share the sentiment of the last email shared by the prospective client. | |
| 5. Write a suitable email reply for the last email of the prospective client, which increases the chances of closing a deal | |
| Provide your output in JSON format with the keys: Category, Client Info, Last comm from client, Sentiment of last comm and Email reply | |
| Here are the Categories for the first task: Corporate Training, Enterprise Plans or Hackathon. | |
| Corporate Training refers to customised training plans developed for companies. The trainings could be either\ | |
| In-Person - Instructor led, Online instructor led or hybrid. | |
| Enterprise Plans refers to subscription plans for enterprises wherein they could purchase licenses of \ | |
| different self paced courses for their employees.\ | |
| There are three kinds of enterprise plans: Blackbelt, Generative AI Pinnalce or Customised \ | |
| The Blackbelt plan focusses on data analytics, data science and machine learning courses starting from \ | |
| excel, python, EDA, stats and going all the way Natural language processing.\ | |
| The Pinnacle plan has courses related to0 generative AI starting from basics of LLMs, finetuning and training LLMs and \ | |
| going till RAGs, Agents and LLMOps.\ | |
| The customised plan is a mix of these which could have a mix of courses from both the plans. | |
| Hackathon refers to Machine Learning or Generatifve AI hackathons, which Analytics Vidhya \ | |
| organises for companies. The companies do it for either Hiring candidates or Branding or a mix of both. | |
| """ | |
| def email_replier(prompt): | |
| messages = [ | |
| {'role':'system', | |
| 'content': AV_system_message}, | |
| {'role':'user', | |
| 'content': f"{delimiter}{prompt}{delimiter}"}, | |
| ] | |
| response = get_completion_messages(messages) | |
| res = json.loads(response) | |
| return res['Category'], res['Client Info'], res[ "Last comm from client"], res["Sentiment of last comm"], res["Email reply"] | |
| demo = gr.Interface(fn=email_replier, | |
| inputs=[gr.Textbox(label="Email Chain", lines=10)], | |
| outputs=[gr.Textbox(label="Category", lines = 1), gr.Textbox(label="Client Info", lines = 3), gr.Textbox(label="Last Comm from client", lines = 1), | |
| gr.Textbox(label="Sentiment", lines = 1), gr.Textbox(label="Email Reply", lines = 5)], | |
| title="Buisness Email Replier", | |
| description= "Use your own judgement to modify the email") | |
| demo.launch() |