Merge pull request #51 from Deeptechia/main

Geppetto v0.2.0

.github/dependabot.yml

@@ -0,0 +1,11 @@
+# To get started with Dependabot version updates, you'll need to specify which
+# package ecosystems to update and where the package manifests are located.
+# Please see the documentation for all configuration options:
+# https://docs.github.com/code-security/dependabot/dependabot-version-updates/configuration-options-for-the-dependabot.yml-file
+
+version: 2
+updates:
+  - package-ecosystem: "cargo" # See documentation for possible values
+    directory: "/" # Location of package manifests
+    schedule:
+      interval: "weekly"

.github/workflows/Sync-Github.yml

@@ -1,15 +1,16 @@
 name: GitlabSync
 
 on:
-  - push
-  - delete
+  push:
+    branches:
+      - main
 
 jobs:
   sync:
     runs-on: ubuntu-latest
     name: Git Repo Sync
     steps:
-    - uses: actions/checkout@v2
+    - uses: actions/checkout@v4
       with:
         fetch-depth: 0
     - uses: wangchucheng/[email protected]

.github/workflows/deploy.yml

@@ -14,7 +14,7 @@ jobs:
 
     steps:
       - name: Checkout Repository
-        uses: actions/checkout@v2
+        uses: actions/checkout@v4
 
       - name: Install SSH Client
         run: sudo apt-get update && sudo apt-get install -y openssh-client

.github/workflows/tests-python.yml

@@ -1,15 +1,19 @@
 
 name: Test Geppetto
 
-on: [ push, pull_request]
+on:
+  push:
+    branches:
+      - main
+      - develop
 
 jobs:
   Test-Python3:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
 
-      - uses: actions/setup-python@v4
+      - uses: actions/setup-python@v5
         with:
           python-version: '3.x'
 

README.md

@@ -7,16 +7,19 @@
   <img src="./assets/GeppettoMini.png" alt="Geppetto Logo"/>
 </p>
 
-Geppetto is a Slack bot for teams to easily interact with ChatGPT. It integrates with OpenAI's ChatGPT-4 and DALL-E-3 models. This project is brought to you by [DeepTechia](https://deeptechia.io/), where the future of technology meets today’s business needs.
+Geppetto is a sophisticated Slack bot that facilitates seamless interaction with multiple AI models, including OpenAI's ChatGPT-4, DALL-E-3, and Google's Gemini model. This versatility allows for a variety of AI-driven interactions tailored to team requirements. This project is brought to you by [DeepTechia](https://deeptechia.io/), where the future of technology meets today’s business needs.
 
 ## Features
 
-1. **Interaction with ChatGPT-4:**
-   - You can send direct messages to the application and receive responses from ChatGPT-4.
-   - Each message generates a conversation thread, and the application uses the message history to formulate coherent responses.
+1. **Flexible AI Model Integration and System Management:**
+   - Users can seamlessly switch between ChatGPT-4-turbo and Gemini to suit their specific interaction needs. ChatGPT-4-turbo is set as the default model.
+   - You can send direct messages to the application and receive responses from Geppetto. Each message generates a conversation thread, and the application uses the message history to formulate coherent responses.
+   - The newly introduced LLM controller component allows the user to manage multiple AI models.
+   - Simplified installation and management process, facilitated by Docker deployment.
+
+2. **Advanced Image Generation with DALL-E-3:**
+   - Leverage DALL-E-3 to generate creative and contextually relevant images directly within Slack conversations.
 
-2. **Image Generation with DALL-E-3:**
-   - The application uses DALL-E-3 to generate an image based on the message.
 
 ![Geppetto](/assets/Geppetto.gif)
 
@@ -72,6 +75,8 @@ Before running the application, copy the `.configuration/.env.example` file into
 - `SIGNING_SECRET`: Your Signing secret to verify Slack requests (from your Slack App Credentials).
 - `DALLE_MODEL`: The OpenAI DALL-E-3 model.
 - `CHATGPT_MODEL`: The OpenAI ChatGPT-4 model.
+- `GEMINI_MODEL`: The Gemini model.
+- `GOOGLE_API_KEY`: The Google Gemini API key.
 
 ## Deployment
 
@@ -89,11 +94,11 @@ Follow these steps to deploy Geppetto:
 Enjoy interacting with ChatGPT-4 and DALL-E-3 on Slack!
 
 ## Docker
-To run geppetto in a docker container, when you have Docker & Docker compose installed:
-1. Move docker-compose.example.yml to docker-compose.yml with customizing where your config folder resides
-2. Change the config values in config/.env
-3. Run docker compose build
-4. Run docker compose up -d
+To run geppetto in a docker container, when you have Docker and Docker compose installed:
+1. Move `docker-compose.example.yml` to `docker-compose.yml`, specifying where your config folder resides.
+2. Change the config values in `config/.env`.
+3. Run `docker compose build`.
+4. Run `docker compose up -d`.
 
 ## Tests
 
@@ -104,7 +109,7 @@ or `python -m unittest -v` for a verbose more specific output
 
 ## About DeepTechia
 
-We are DeepTechia, where the future of technology meets today’s business needs. As pioneers in the digital realm, we’ve made it our mission to bridge the gap between innovation and practicality, ensuring that businesses not only survive but thrive in an ever-evolving technological landscape.
+We are [DeepTechia](https://deeptechia.io/), where the future of technology meets today’s business needs. As pioneers in the digital realm, we’ve made it our mission to bridge the gap between innovation and practicality, ensuring that businesses not only survive but thrive in an ever-evolving technological landscape.
 
 Born from a passion for cutting-edge technology and a vision for a digitally integrated future, DeepTechia was established to be more than just a tech consultancy. We are visionaries, strategists, and implementers, dedicated to pushing the boundaries of what’s possible while ensuring real-world applicability.
 

config/.env.example

@@ -1,6 +1,8 @@
 SLACK_BOT_TOKEN = "YOUR_TOKEN"
 SLACK_APP_TOKEN = "YOUR_TOKEN"
 OPENAI_API_KEY  = "YOUR_TOKEN"
-CHATGPT_MODEL   = "gpt-4"
+CHATGPT_MODEL   = "gpt-4-turbo"
 DALLE_MODEL     = "dall-e-3"
 SIGNING_SECRET  = "YOUR_SECRET"
+GOOGLE_API_KEY  = "YOUR_TOKEN"
+GEMINI_MODEL    = "gemini-pro"

config/allowed-slack-ids.json

@@ -1,4 +1,5 @@
 {
+  "*":"*",
   "User A": "#MemberIDUserA",
   "User B": "#MemberIDUSerB"
 }

config/default_responses.json

@@ -2,8 +2,7 @@
   "features": {
     "personality": "You are Geppetto, a general intelligence bot created by DeepTechia."
   },
-  "dalle": { "preparing_image": "Preparing image.." },
   "user": {
     "permission_denied": "The requesting user does not belong to the list of allowed users. Request permission to use the app"
   }
-}
+}

geppetto/exceptions.py

@@ -0,0 +1,10 @@
+# Geppetto Exceptions
+
+
+class InvalidThreadFormatError(KeyError):
+    """Invalid thread format.
+
+    Raise if the submitted thread format doesn't have the expected layout.
+    Since the UIs and the underlying LLM engines must meet an interface,
+    some validations have to be undertaken to assure key fields.
+    """

geppetto/gemini_handler.py

@@ -0,0 +1,61 @@
+from urllib.request import urlopen
+import logging
+
+from .exceptions import InvalidThreadFormatError
+from .llm_api_handler import LLMHandler
+from dotenv import load_dotenv
+from typing import List, Dict
+import os
+import textwrap
+import google.generativeai as genai
+from IPython.display import display
+from IPython.display import Markdown
+
+load_dotenv(os.path.join("config", ".env"))
+
+GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY")
+GEMINI_MODEL=os.getenv("GEMINI_MODEL", "gemini-pro")
+MSG_FIELD = "parts"
+MSG_INPUT_FIELD = "content"
+
+def to_markdown(text):
+    text = text.replace('•', '  *')
+    return Markdown(textwrap.indent(text, '> ', predicate=lambda _: True))
+
+class GeminiHandler(LLMHandler):
+
+    def __init__(
+        self,
+        personality,
+    ):
+        super().__init__(
+            'Gemini',
+            GEMINI_MODEL,
+            genai.GenerativeModel(GEMINI_MODEL),
+        )
+        self.personality = personality
+        self.system_role = "system"
+        self.assistant_role = "model"
+        self.user_role = "user"
+        genai.configure(api_key=GOOGLE_API_KEY)
+
+    def llm_generate_content(self, user_prompt, status_callback=None, *status_callback_args):
+        logging.info("Sending msg to gemini: %s" % user_prompt)
+        if len(user_prompt) >= 2 and user_prompt[0].get('role') == 'user' and user_prompt[1].get('role') == 'user':
+            merged_prompt = {
+                'role': 'user',
+                'parts': [msg['parts'][0] for msg in user_prompt[:2]]
+            }
+            user_prompt = [merged_prompt] + user_prompt[2:]
+        response= self.client.generate_content(user_prompt)
+        markdown_response = to_markdown(response.text)
+        return str(markdown_response.data)
+
+    def get_prompt_from_thread(self, thread: List[Dict], assistant_tag: str, user_tag: str):
+        prompt = super().get_prompt_from_thread(thread, assistant_tag, user_tag)
+        for msg in prompt:
+            if MSG_INPUT_FIELD in msg:
+                msg[MSG_FIELD] = [msg.pop(MSG_INPUT_FIELD)]
+            else:
+                raise InvalidThreadFormatError("The input thread doesn't have the field %s" % MSG_INPUT_FIELD) 
+        return prompt

geppetto/llm_api_handler.py

@@ -0,0 +1,30 @@
+from abc import ABC, abstractmethod
+from typing import List, Dict, Callable
+from .exceptions import InvalidThreadFormatError
+
+ROLE_FIELD = "role"
+
+class LLMHandler(ABC):
+    def __init__(self, name, model, client):
+        self.name = name
+        self.model = model
+        self.client = client
+
+    def get_info(self):
+        return f"Name: {self.name} - Model: {self.model}"
+
+    @abstractmethod
+    def llm_generate_content(self, prompt: str, callback: Callable, *callback_args):
+        pass
+
+    def get_prompt_from_thread(self, thread: List[Dict], assistant_tag: str, user_tag: str):
+        prompt = []
+        for msg in thread:
+            formatted_msg = dict(msg)
+            if ROLE_FIELD in formatted_msg:
+                formatted_msg[ROLE_FIELD] = formatted_msg[ROLE_FIELD].replace(assistant_tag, self.assistant_role)
+                formatted_msg[ROLE_FIELD] = formatted_msg[ROLE_FIELD].replace(user_tag, self.user_role)
+                prompt.append(formatted_msg)
+            else:
+                raise InvalidThreadFormatError("The input thread doesn't have the field %s" % ROLE_FIELD)
+        return prompt

geppetto/llm_controller.py

@@ -0,0 +1,36 @@
+from typing import List, Type, TypedDict, Dict
+from .llm_api_handler import LLMHandler
+
+
+class LLMCfgRec(TypedDict):
+    name: str
+    handler: Type[LLMHandler]
+    handler_args: Dict
+
+
+LLMCfgs = List[LLMCfgRec]
+
+
+class LLMController:
+
+    def __init__(self, llm_cfgs: LLMCfgs):
+        self.llm_cfgs = llm_cfgs
+        self.handlers = {}
+
+    def init_controller(self):
+        for llm in self.llm_cfgs:
+            name = llm['name']
+            self.handlers[name] = self.get_handler(name)
+
+    def list_llms(self):
+        return [x['name'] for x in self.llm_cfgs]
+
+    def get_llm_cfg(self, name):
+        for llm in self.llm_cfgs:
+            if llm['name'] == name:
+                return llm
+        raise ValueError("LLM configuration not found for name: %s" % name)
+
+    def get_handler(self, name):
+        llm_cfg = self.get_llm_cfg(name)
+        return llm_cfg['handler'](**llm_cfg['handler_args'])

geppetto/main.py

@@ -1,33 +1,59 @@
 import os
 import logging
 from dotenv import load_dotenv
+
+from .llm_controller import LLMController
 from .slack_handler import SlackHandler
+from .openai_handler import OpenAIHandler
+from .gemini_handler import GeminiHandler
 from slack_bolt.adapter.socket_mode import SocketModeHandler
 from .utils import load_json
 
 load_dotenv(os.path.join("config", ".env"))
 
-OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
-DALLE_MODEL = os.getenv("DALLE_MODEL")
-CHATGPT_MODEL = os.getenv("CHATGPT_MODEL")
+
 SLACK_BOT_TOKEN = os.getenv("SLACK_BOT_TOKEN_TEST")
 SLACK_APP_TOKEN = os.getenv("SLACK_APP_TOKEN_TEST")
 SIGNING_SECRET = os.getenv("SIGNING_SECRET_TEST")
 
+DEFAULT_RESPONSES = load_json("default_responses.json")
+
 # Initialize logging
 # TODO: log to a file
 logging.basicConfig(level=logging.INFO)
 
 
+def initialized_llm_controller():
+    controller = LLMController(
+        [
+            {
+                "name": "OpenAI",
+                "handler": OpenAIHandler,
+                "handler_args": {
+                    "personality": DEFAULT_RESPONSES["features"]["personality"]
+                }
+            },
+            {
+                "name": "Gemini",
+                "handler": GeminiHandler,
+                "handler_args": {
+                    "personality": DEFAULT_RESPONSES["features"]["personality"]
+                }
+            }
+        ]
+    )
+    controller.init_controller()
+    return controller
+
+
+
 def main():
     Slack_Handler = SlackHandler(
         load_json("allowed-slack-ids.json"),
-        load_json("default_responses.json"),
+        DEFAULT_RESPONSES,
         SLACK_BOT_TOKEN,
         SIGNING_SECRET,
-        OPENAI_API_KEY,
-        DALLE_MODEL,
-        CHATGPT_MODEL,
+        initialized_llm_controller()
     )
     SocketModeHandler(Slack_Handler.app, SLACK_APP_TOKEN).start()