DST-198: add llm selection tool and base transcript (#22)

Resolves DST-198: base code for processing transcripts: Added prompt for transcript and selection for openhermes, dolphin, gemini, gpt 3.5 and gpt 4

04-call-summaries/.gitignore

@@ -0,0 +1,2 @@
+.env
+__pycache__

04-call-summaries/llm.py

@@ -0,0 +1,68 @@
+import dotenv
+
+from langchain_community.llms.ollama import Ollama
+import google.generativeai as genai
+from openai import OpenAI
+import os
+
+dotenv.load_dotenv()
+
+
+def get_transcript(file_path="./transcript.txt"):
+ file = open(file_path)
+ content = file.read()
+ return content
+
+
+def ollama_client(
+ model_name=None,
+ prompt=None,
+ callbacks=None,
+ settings=None,
+):
+ if not settings:
+ settings = {
+ # "temperature": 0.1,
+ # "system": "",
+ # "template": "",
+ # See https://github.com/langchain-ai/langchain/blob/master/libs/community/langchain_community/llms/ollama.py
+ "stop": None
+ }
+
+ print("LLM settings:", model_name, settings)
+ # To connect via another URL: Ollama(base_url='http://localhost:11434', ...)
+ return Ollama(model=model_name, callbacks=callbacks, **settings).invoke(prompt)
+
+
+def google_gemini_client(
+ model_name="gemini-pro",
+ prompt=None,
+ settings=None,
+):
+ # Get a Google API key by following the steps after clicking on Get an API key button
+ # at https://ai.google.dev/tutorials/setup
+ GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY")
+
+ print("LLM settings:", model_name, settings)
+
+ genai.configure(api_key=GOOGLE_API_KEY)
+ if settings:
+ genai.GenerationConfig(**settings)
+ model = genai.GenerativeModel(model_name)
+ return model.generate_content(prompt)
+
+
+def gpt3_5(prompt, model="gpt-3.5-turbo"):
+ OPEN_AI_API_KEY = os.environ.get("OPEN_AI_API_KEY")
+ openai_client = OpenAI(api_key=OPEN_AI_API_KEY) # Uses OPENAI_API_KEY
+ return (
+ openai_client.chat.completions.create(
+ model=model, messages=[{"role": "user", "content": prompt}]
+ )
+ .choices[0]
+ .message.content
+ )
+
+
+def gpt_4_turbo(prompt):
+ return gpt3_5(prompt, model="gpt-4-turbo")

04-call-summaries/requirements.txt

@@ -0,0 +1,5 @@
+google-generativeai
+tokenizers
+langchain
+langchain_community
+openai

04-call-summaries/run.py

@@ -0,0 +1,76 @@
+from llm import google_gemini_client, ollama_client, gpt3_5, gpt_4_turbo
+from langchain_core.prompts import PromptTemplate
+
+
+# download transcripts from https://drive.google.com/drive/folders/19r6x3Zep4N9Rl_x4n4H6RpWkXviwbxyw?usp=sharing
+def get_transcript(file_path="./transcript.txt"):
+ file = open(file_path)
+ content = file.read()
+ file.close()
+ return content
+
+
+prompt = """
+You are a helpful AI assistant who will summarize this transcript {transcript}, using the following template:
+---------------------
+Caller Information (Name, contact information, availability, household information)
+
+Reason/Type of Call (Applying for benefits, Follow-Ups)
+
+Previous Benefits History (Applied for, Receives, Denied, etc)
+
+Put # in front of the benefit discussed (i.e. #SNAP, LIHEAP)
+
+Discussion Points (Key information points)
+
+Documents Needed (Income, Housing, etc)
+
+Next Steps for Client
+
+Next Steps for Agent
+---------------------
+
+"""
+
+print("""
+ Select an llm
+ 1. openhermes (default)
+ 2. dolphin
+ 3. gemini
+ 4. gpt 3.5
+ 5. gpt 4
+ """)
+
+
+transcript = get_transcript()
+llm = input() or "1"
+prompt_template = PromptTemplate.from_template(prompt)
+formatted_prompt = prompt_template.format(transcript=transcript)
+
+if llm == "2":
+ test = ollama_client(model_name="dolphin-mistral", prompt=formatted_prompt)
+ print("""----------
+ Dolphin
+ """)
+elif llm == "3":
+ test = google_gemini_client(prompt=formatted_prompt).text
+ print("""----------
+ Gemini
+ """)
+elif llm == "4":
+ print("""----------
+ GPT 3.5
+ """)
+ test = gpt3_5(prompt=formatted_prompt)
+elif llm == "5":
+ print("""----------
+ GPT 4
+ """)
+ test = gpt_4_turbo(prompt=formatted_prompt)
+else:
+ test = ollama_client(model_name="openhermes", prompt=formatted_prompt)
+ print("""
+ Openhermes
+ """)
+
+print(test)