fix: improved --help output

README.md

@@ -130,8 +130,11 @@ Usage: gptme [OPTIONS] [PROMPTS]...
     /undo        Undo the last action.
     /log         Show the conversation log.
     /edit        Edit previous messages.
+    /rename      Rename the conversation.
+    /fork        Create a copy of the conversation with a new name.
     /summarize   Summarize the conversation so far.
     /load        Load a file.
+    /save        Save the most recent code block to a file.
     /shell       Execute a shell command.
     /python      Execute a Python command.
     /replay      Re-execute past commands in the conversation (does not store output in log).
@@ -145,9 +148,8 @@ Options:
   --name TEXT                     Name of conversation. Defaults to generating
                                   a random name. Pass 'ask' to be prompted for
                                   a name.
-  --llm [openai|llama]            LLM to use.
-  --model [gpt-4|gpt-3.5-turbo|wizardcoder-...]
-                                  Model to use (gpt-3.5 not recommended)
+  --llm [openai|local]            LLM to use.
+  --model TEXT                    Model to use.
   --stream / --no-stream          Stream responses
   -v, --verbose                   Verbose output.
   -y, --no-confirm                Skips all confirmation prompts.
@@ -157,6 +159,7 @@ Options:
                                   used in testing.
   --show-hidden                   Show hidden system messages.
   --version                       Show version.
+  --server                        Run as server.
   --help                          Show this message and exit.
 ```
 

gptme/cli.py

@@ -51,7 +51,7 @@
 logger = logging.getLogger(__name__)
 print_builtin = __builtins__["print"]  # type: ignore
 
-LLMChoice = Literal["openai", "llama"]
+LLMChoice = Literal["openai", "local"]
 ModelChoice = Literal["gpt-3.5-turbo", "gpt4"]
 
 
@@ -86,12 +86,12 @@
     "--llm",
     default="openai",
     help="LLM to use.",
-    type=click.Choice(["openai", "llama"]),
+    type=click.Choice(["openai", "local"]),
 )
 @click.option(
     "--model",
     default="gpt-4",
-    help="Model to use (gpt-3.5 not recommended). Can be: gpt-4, gpt-3.5-turbo, wizardcoder-..., etc.",
+    help="Model to use.",
 )
 @click.option(
     "--stream/--no-stream",

gptme/llm.py

@@ -34,7 +34,7 @@ def init_llm(llm: str):
             print("Error: OPENAI_API_KEY not set in env or config, see README.")
             sys.exit(1)
         openai.api_key = api_key
-    elif llm in ["local", "llama"]:
+    elif llm == "local":
         if "OPENAI_API_BASE" in os.environ:
             api_base = os.environ["OPENAI_API_BASE"]
         elif api_base := config["env"].get("OPENAI_API_BASE", None):
@@ -58,17 +58,6 @@ def reply(messages: list[Message], model: str, stream: bool = False) -> Message:
         return Message("assistant", response)
 
 
-def _complete(prompt: str, model: str) -> str:
-    print(prompt)
-    response = openai.Completion.create(  # type: ignore
-        model=model,
-        prompt=prompt,
-        temperature=temperature,
-        top_p=top_p,
-    )
-    return response.choices[0].text
-
-
 def _chat_complete(messages: list[Message], model: str) -> str:
     # This will generate code and such, so we need appropriate temperature and top_p params
     # top_p controls diversity, temperature controls randomness