Initial commit

.gitignore

+.venv
+__pycache__
+secret_key.py
\ No newline at end of file

README.md

+# Prerequisits
+- python3 installed
+- [openAI API Key](https://auth.openai.com/)
+
+# Install
+```
+python3 -m venv .venv
+source .venv/bin/activate
+pip install-r requirements.txt
+cp ./secret_key.py.template ./secret_key.py
+```
+
+# Configure
+populate the `secret_key.py` file with proper information
+
+# Start
+Each segment of the file `basics.py` represents a step.
+Each step is meant to run independently. So you uncomment a single segment, and run `python3 basics.py`
+
+# Sources
+- [YT: LangChain Crash Course For Beginners | LangChain Tutorial ](https://www.youtube.com/watch?v=nAmC7SoVLd8)

basics.py

+##############################################
+# PREPARE FOR FOLLOWING STEP(S)
+##############################################
+# import openai api key from ./secret_key.py
+from secret_key import openapi_key
+import os
+os.environ['OPENAI_API_KEY'] = openapi_key
+
+# define llm
+from langchain_openai import OpenAI
+# higher temperature allows llm to be more creative (value between 0.0 and 1.0)
+llm = OpenAI(temperature=0)
+
+##############################################
+#  SIMPLE PROMPT TO OPENAI
+##############################################
+name = llm("I want to open a restaurant for Indian food. Suggest a fancy name for this.")
+print(name)
+
+##############################################
+# CREATE A PROMPT TEMPLATE
+##############################################
+# from langchain.prompts import PromptTemplate
+
+# # set the template
+# prompt_template_name = PromptTemplate(
+#   input_variables=['cuisine'],
+#   template = "I want to open a restaurant for {cuisine} food. Suggest a fancy name for this."
+# )
+
+# # pipe the value
+# prompt = prompt_template_name.format(cuisine="Mexican")
+# print(prompt)
+
+
+##############################################
+# CREATE A CHAIN
+############################################## 
+# from langchain.chains import LLMChain
+# from langchain.prompts import PromptTemplate
+
+# # set the template
+# prompt_template_name = PromptTemplate(
+#   input_variables=['cuisine'],
+#   template = "I want to open a restaurant for {cuisine} food. Suggest a fancy name for this."
+# )
+# # create the chain link
+# restaurant_name_chain = LLMChain(llm=llm, prompt=prompt_template_name)
+# answer = restaurant_name_chain.invoke("American")
+# print('answer:',answer)
+
+##############################################
+# CREATE A SIMPLE SEQUENTIAL CHAIN
+# a chain is aware of the steps/links and uses them to sharpen the output of the following chain links
+############################################## 
+# from langchain.chains import LLMChain
+# from langchain.prompts import PromptTemplate
+
+# # set the template 1
+# prompt_template_name = PromptTemplate(
+#   input_variables=['cuisine'],
+#   template = "I want to open a restaurant for {cuisine} food. Suggest a fancy name for this."
+# )
+# # create the chain link 1
+# restaurant_name_link = LLMChain(llm=llm, prompt=prompt_template_name)
+
+# # set the template 2
+# prompt_template_items = PromptTemplate(
+#   input_variables=['restaurant_name'],
+#   template = "Suggest ten menu items for {restaurant_name}. Return it as comma separated list."
+# )
+# # create the chain link 2
+# food_items_link = LLMChain(llm=llm, prompt=prompt_template_items)
+
+# # chaining the links
+# from langchain.chains import SimpleSequentialChain
+# chain = SimpleSequentialChain(chains = [restaurant_name_link, food_items_link])
+# response = chain.invoke("Korean")
+# # but the output only provides the answer to the very last prompt
+# # to avoid this, use a squential chain (not a SIMPLE sequential chain)
+# print(response)
+
+##############################################
+# CREATE A SEQUENTIAL CHAIN
+# can have mutliple in- and outputs
+############################################## 
+# from langchain.chains import LLMChain
+# from langchain.prompts import PromptTemplate
+
+# # set the template 1
+# prompt_template_name = PromptTemplate(
+#   input_variables=['cuisine'],
+#   template = "I want to open a restaurant for {cuisine} food. Suggest a fancy name for this."
+# )
+# # create the chain link 1
+# restaurant_name_link = LLMChain(llm=llm, prompt=prompt_template_name, output_key='restaurant_name')
+
+# # set the template 2
+# prompt_template_items = PromptTemplate(
+#   input_variables=['restaurant_name'],
+#   template = "Suggest ten menu items for {restaurant_name}. Return it as comma separated list."
+# )
+# # create the chain link 2
+# food_items_link = LLMChain(llm=llm, prompt=prompt_template_items, output_key='menu_items')
+
+# # chaining the links
+# from langchain.chains import SequentialChain
+# chain = SequentialChain(
+#   chains = [restaurant_name_link, food_items_link],
+#   input_variables=['cuisine'],
+#   output_variables=['restaurant_name', 'menu_items'],
+#   )
+# response = chain({'cuisine': 'Greek'})
+# print(response)
+
+##############################################
+# AGENTS
+# agents are a way to use external functions & sources
+############################################## 
+# from langchain.agents import AgentType, initialize_agent, load_tools
+# from secret_key import serpapi_api_key
+
+# os.environ['SERPAPI_API_KEY'] = serpapi_api_key
+
+# # load an agent
+# # before using serpapi run 'pip install google-search-results numexpr'
+# tools= load_tools(['serpapi','llm-math'], llm=llm)
+# agent = initialize_agent(tools, llm, agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
+# response = agent.invoke("What was the GDP of the US in 2022?")
+# print(response)
+
+
+##############################################
+# MEMORY
+# keeps awareness of the whole conversation
+##############################################
+# from langchain.memory import ConversationBufferMemory
+# from langchain.chains import LLMChain
+# from langchain.prompts import PromptTemplate
+
+# # initiate memory instance
+# memory = ConversationBufferMemory()
+# # set the template
+# prompt_template_name = PromptTemplate(
+#   input_variables=['cuisine'],
+#   template = "I want to open a restaurant for {cuisine} food. Suggest a fancy name for this."
+# )
+# # create the chain link
+# restaurant_name_chain = LLMChain(llm=llm, prompt=prompt_template_name, memory=memory)
+# answer = restaurant_name_chain.invoke("American")
+# print('answer:',answer)
+
+# # run once more
+# answer = restaurant_name_chain.invoke("German")
+# print('answer:',answer)
+
+# # look at memory now
+# print(restaurant_name_chain.memory.buffer)
+
+##############################################
+# CONVERSATION CHAIN
+# appends whole history to each request
+##############################################
+# from langchain.chains import ConversationChain
+# # initiate
+# conv = ConversationChain(llm=llm)
+# # check default template
+# # print(conv.prompt.template)
+
+# answer = conv.invoke("Who won the first cricket world cup?")
+# print(answer)
+# answer = conv.invoke("What's 5+5?")
+# print(answer)
+# answer = conv.invoke("Who was the captain of the winning team?")
+# print(answer)
+# #all the history is send on each request. Thi can cost a lot of tokens (and money)
+
+##############################################
+# CONVERSATION CHAIN WITH BUFFER WINDOW MEMORY
+# appends a defined amount of history to each request
+##############################################
+# from langchain.chains import ConversationChain
+# from langchain.memory import ConversationBufferWindowMemory
+# # initiate (k=count of how many historic conversations shall be send on request)
+# memory = ConversationBufferWindowMemory(k=1)
+# conv = ConversationChain(llm=llm, memory=memory)
+
+# answer = conv.invoke("Who won the first cricket world cup?")
+# print(answer)
+# answer = conv.invoke("What's 5+5?")
+# print(answer)
+# answer = conv.invoke("Who was the captain of the winning team?")
+# print(answer)
+# # all the history is send on each request. Thi can cost a lot of tokens (and money)
\ No newline at end of file

requirements.txt

+aiohttp==3.9.4
+aiosignal==1.3.1
+annotated-types==0.6.0
+anyio==4.3.0
+attrs==23.2.0
+certifi==2024.2.2
+charset-normalizer==3.3.2
+dataclasses-json==0.6.4
+distro==1.9.0
+frozenlist==1.4.1
+google-search-results==2.4.2
+greenlet==3.0.3
+h11==0.14.0
+httpcore==1.0.5
+httpx==0.27.0
+idna==3.7
+jsonpatch==1.33
+jsonpointer==2.4
+langchain==0.1.16
+langchain-community==0.0.32
+langchain-core==0.1.42
+langchain-openai==0.1.3
+langchain-text-splitters==0.0.1
+langsmith==0.1.47
+marshmallow==3.21.1
+multidict==6.0.5
+mypy-extensions==1.0.0
+numexpr==2.10.0
+numpy==1.26.4
+openai==1.17.1
+orjson==3.10.0
+packaging==23.2
+pydantic==2.7.0
+pydantic_core==2.18.1
+PyYAML==6.0.1
+regex==2023.12.25
+requests==2.31.0
+sniffio==1.3.1
+SQLAlchemy==2.0.29
+tenacity==8.2.3
+tiktoken==0.6.0
+tqdm==4.66.2
+typing-inspect==0.9.0
+typing_extensions==4.11.0
+urllib3==2.2.1
+yarl==1.9.4

secret_key.py.template

+openapi_key="sk-hj8dfs8f9dsfsdgibberish"
\ No newline at end of file