Add Basic Level Support for Linux

.gitignore

@@ -11,3 +11,9 @@ mpnet/
 baai/
 whisper
 espeak-ng-data
+
+## LINUX IGNORE
+env/
+__pycache__/
+*.log
+*.out

README.md

@@ -63,6 +63,16 @@ Now python should be setup and running! However, there is still a few more steps
 
 Finally, download the Mistral 7B LLM from the following link and place it inside the `llm/scripts` directory alongside the python scripts used by Dot: [TheBloke/Mistral-7B-Instruct-v0.2-GGUF](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GGUF/blob/main/mistral-7b-instruct-v0.2.Q4_K_M.gguf)
 
+## Linux Installation
+```bash
+$ chmod +x setup_python.sh
+$ ./setup_python.sh
+```
+The above command will check if python is installed on your system and start the setup.
+Once the setup is done, configure all the models by specifying their features in the config file located at `llm/config.ini`
+
+Finally, run `npm start`
+
 That's it! If you follow these steps you should be able to get it all running, please let me know if you are facing any issues :)
 
 

llm/config.ini

@@ -0,0 +1,2 @@
+[EMBEDDING MODEL CONFIG]
+MODEL_NAME=sentence-transformers/all-mpnet-base-v2

llm/requirements.txt

@@ -0,0 +1,8 @@
+torch
+langchain-community
+faiss-cpu
+huggingface_hub
+llama-cpp-python -C cmake.args="-DLLAMA_BLAS=ON;-DLLAMA_BLAS_VENDOR=OpenBLAS;-DLLAMA_CUDA=ON"
+pypdf
+unstructured
+docx2txt

llm/scripts/bigdot.py

@@ -1,14 +1,15 @@
-from langchain.prompts import PromptTemplate
-from langchain.chains import LLMChain
-from langchain.llms import LlamaCpp
-from langchain.memory import ConversationBufferWindowMemory
+from langchain_core.prompts.prompt import PromptTemplate
+from langchain.chains.llm import LLMChain
+from langchain_community.llms.llamacpp import LlamaCpp
+from langchain.memory.buffer_window import ConversationBufferWindowMemory
 import sys
 import json
 import os
 
+
 def read_config(config_path):
     try:
-        with open(config_path, 'r') as file:
+        with open(config_path, "r") as file:
             config = json.load(file)
         return config
     except FileNotFoundError:
@@ -18,6 +19,7 @@ def read_config(config_path):
         print("Error decoding JSON from configuration file.")
         return {}
 
+
 if __name__ == "__main__":
     if len(sys.argv) > 1:
 
@@ -26,24 +28,27 @@ def read_config(config_path):
         folder_name = "Dot-Data"
         folder_path = os.path.join(documents_path, folder_name)
         if not os.path.exists(folder_path):
-            print('LLM NOT FOUND!')
+            print("LLM NOT FOUND!")
             os.makedirs(folder_path)
         model_path = os.path.join(folder_path, "Phi-3-mini-4k-instruct-q4.gguf")
 
         config_path = sys.argv[1].strip('"')  # Remove any surrounding quotes
         config = read_config(config_path)
-        #print("Current Configuration:", config)
+        # print("Current Configuration:", config)
 
         # Setup configuration with defaults in case some settings are missing
         n_gpu_layers = -1  # Metal set to 1 is typically enough for Apple Silicon
-        n_batch = config.get('n_batch', 256)
-        max_tokens = config.get('max_tokens', 500)
-        temperature = config.get('big_dot_temperature', 0.7)
-        n_ctx = config.get('n_ctx', 4000)
-        initial_prompt = config.get('big_dot_prompt', "You are called Dot, You are a helpful and honest assistant.")
-        if 'ggufFilePath' in config and config['ggufFilePath'] is None:
-            del config['ggufFilePath']  # Removes the key if it's explicitly None
-        llm_model = config.get('ggufFilePath', model_path)
+        n_batch = config.get("n_batch", 256)
+        max_tokens = config.get("max_tokens", 500)
+        temperature = config.get("big_dot_temperature", 0.7)
+        n_ctx = config.get("n_ctx", 4000)
+        initial_prompt = config.get(
+            "big_dot_prompt",
+            "You are called Dot, You are a helpful and honest assistant.",
+        )
+        if "ggufFilePath" in config and config["ggufFilePath"] is None:
+            del config["ggufFilePath"]  # Removes the key if it's explicitly None
+        llm_model = config.get("ggufFilePath", model_path)
 
         # Initialize the LLM with the configuration
         llm = LlamaCpp(
@@ -67,12 +72,7 @@ def read_config(config_path):
         Response:"""
         prompt = PromptTemplate.from_template(template)
         memory = ConversationBufferWindowMemory(memory_key="chat_history", k=2)
-        conversation = LLMChain(
-            llm=llm,
-            prompt=prompt,
-            verbose=False,
-            memory=memory
-        )
+        conversation = LLMChain(llm=llm, prompt=prompt, verbose=False, memory=memory)
 
         def send_response(response):
             response_json = json.dumps({"result": response})
@@ -85,9 +85,10 @@ def send_response(response):
             if not user_input:
                 break
 
-            result = conversation({"question": user_input})['text']
+            result = conversation({"question": user_input})["text"]
             max_chunk_length = 1000  # Define max length for output chunks
-            chunks = [result[i:i + max_chunk_length] for i in range(0, len(result), max_chunk_length)]
+            chunks = [
+                result[i : i + max_chunk_length]
+                for i in range(0, len(result), max_chunk_length)
+            ]
             send_response(chunks)
-
-