Shell GPT using Langchain

shell-gpt.py

from langchain.llms import OpenAI
import os
from langchain import LLMChain, PromptTemplate
import sys
import platform
import distro
from langchain.memory import ConversationBufferMemory

def get_llm():
    os.environ["OPENAI_API_KEY"] = "sk-abc"
    llm = OpenAI(model_name="gpt-3.5-turbo", temperature=0)    
    return llm

def get_shell():
    if os.name == 'nt':  # For Windows
        shell = os.environ.get('COMSPEC', 'Unknown shell')
    elif os.name == 'posix':
        shell = os.environ.get('SHELL', 'Unknown shell')
    else:
        shell = 'Bash'
    return shell

def get_os():
    os_name = platform.system()
    if os_name == 'Windows':
        return 'Windows'
    elif os_name == 'Darwin':
        return 'Apple macOS'
    elif os_name == 'Linux':
        return distro.name(pretty=True)
    else:
        return 'Linux'
llm = get_llm()
PROMPT_TEMPLATE = """If someone asks you to perform a task, your job is to come up with a series of shell commands that will perform the task. 
Ensure the commands are for the OS {os_name} and the shell {shell_name}
Make sure to reason step by step, using this format:

Question: "copy the files in the directory named 'target' into a new directory at the same level as target called 'myNewDirectory'"

I need to take the following actions:
- List all files in the directory
- Create a new directory
- Copy the files from the first directory into the second directory
```Command:
ls
mkdir myNewDirectory
cp -r target/* myNewDirectory
```

That is the format. Begin!

{chat_history}
Question: {question}
"""



PROMPT = PromptTemplate(
    input_variables=["question", "os_name", "shell_name", "chat_history"],
    template=PROMPT_TEMPLATE,
)

memory = ConversationBufferMemory(memory_key="chat_history", input_key="question")
llm_chain = LLMChain(prompt=PROMPT, llm=llm, verbose=False, memory=memory)

shell_name = get_shell()
os_name = get_os()
while True:
    question = input("Please provide a question: ")
    # print(PROMPT_TEMPLATE.format(question = question, os_name = os_name, shell_name =  shell_name, chat_history = ""))
    print(llm_chain.run({"question": question, "shell_name": shell_name, "os_name": os_name}))