added developers instructions

README.md

@@ -14,11 +14,16 @@ This is an early prototype of using prompting strategies to improve the LLM's re
 - [x] Perplexity (remote, requires API key)
 - [x] Groq (remote, requires API key)
 
+## Developer Resources
+
+- Instructions for adding new providers can be found in `app/utils/providers/instructions.md`
+- A skeleton provider template is available at `app/utils/providers/skeleton_provider.py`
+
 ## Work in progress
 
 - [ ] Add more providers
 - [ ] Possibly use LiteLLM instead of defining each provider
-- [ ] Create a guide to quickly add providers
+- [ ] Create a better way to add new providers for developers
 
 ## Example
 

app/utils/providers/instructions.md

@@ -0,0 +1,21 @@
+# Creating a New Provider
+
+To add a new provider to the multi1 application, follow these steps:
+
+1. Create a new file in the `app/handlers/` directory named `your_provider_handler.py`.
+
+2. Copy the contents of the `skeleton_provider.py` file into your new handler file.
+
+3. Rename the class to match your provider (e.g., `YourProviderHandler`).
+
+4. Implement the `__init__`, `_make_request`, and `_process_response` methods according to your provider's API requirements.
+
+5. Import your new handler in `app/handlers/__init__.py`.
+
+6. Update the `get_api_handler` function in `app/main.py` to include your new provider.
+
+7. Add the necessary configuration options in `app/config_menu.py`.
+
+8. Update the `README.md` file to include information about the new provider.
+
+Remember to handle API keys, rate limiting, and error responses appropriately for your provider.

app/utils/providers/skeleton_provider.py

@@ -0,0 +1,17 @@
+from api_handlers import BaseHandler
+
+class SkeletonProviderHandler(BaseHandler):
+    def __init__(self, api_key, model):
+        super().__init__()
+        self.api_key = api_key
+        self.model = model
+
+    def _make_request(self, messages, max_tokens):
+        # Implement the API request to your provider here
+        # Return the raw response from the API
+        pass
+
+    def _process_response(self, response, is_final_answer):
+        # Process the API response and return a formatted dictionary
+        # The dictionary should have 'title', 'content', and 'next_action' keys
+        pass