E-sim Chatbot Semester Long Internship autumn 2025 ( Chatbot Enhancements)

installation_guide.md

@@ -0,0 +1,170 @@
+# eSim AI Chatbot - Generalized Installation and Setup Guide
+
+## 1. Overview
+
+Before you begin, ensure you have:
+
+- **OS:** Windows 10/11 with WSL2, macOS, or Ubuntu Linux
+- **RAM:** Minimum 8 GB
+- **Disk Space:** At least 15 GB free
+- **Internet:** Required for initial downloads
+
+---
+
+## 2. Install Required Dependencies (OS-Specific)
+
+### Linux (Ubuntu/Debian)
+
+```bash
+sudo apt update && sudo apt upgrade -y
+sudo apt install -y python3 python3-venv python3-pip git curl \
+portaudio19-dev python3-pyaudio \
+libgl1 libglib2.0-0 \
+libxcb-xinerama0 libxkbcommon-x11-0 libgl1-mesa-glx
+```
+
+### macOS
+
+```bash
+brew install python3 portaudio git
+```
+
+### Windows (WSL2)
+
+Open **PowerShell as Administrator** and run:
+
+```powershell
+wsl --install
+```
+
+Then restart your system and follow the Linux steps above inside WSL.
+
+---
+
+## 3. Install Ollama
+
+Download and install Ollama from:
+
+https://ollama.com
+
+Start Ollama:
+
+```bash
+ollama serve
+```
+
+---
+
+## 4. Download AI Models
+
+Run the following commands:
+
+```bash
+ollama pull qwen2.5-coder:3b
+ollama pull qwen2.5-coder:1.5b
+ollama pull nomic-embed-text
+ollama pull minicpm-v
+ollama pull tinyllama:1.1b
+```
+
+Verify installation:
+
+```bash
+ollama list
+```
+
+---
+
+## 5. Clone the Repository
+
+```bash
+git clone https://github.com/username/Repository_name
+
+cd Repository_name
+```
+
+---
+
+## 6. Set Up Python Environment
+
+### Linux/macOS
+
+```bash
+python3 -m venv venv
+
+source venv/bin/activate
+```
+
+### Windows
+
+```powershell
+python -m venv venv
+
+venv\Scripts\activate
+```
+
+---
+
+## 7. Install Python Packages
+
+```bash
+pip install --upgrade pip
+
+pip install -r requirements.txt
+```
+
+---
+
+## 8. Run the Chatbot
+
+```bash
+PYTHONPATH=src python3 -m frontEnd.Application
+```
+
+---
+
+## 9. Daily Usage
+
+```bash
+cd esim-chatbot-pr34
+
+source venv/bin/activate
+
+ollama serve &
+
+PYTHONPATH=src python3 -m frontEnd.Application
+```
+
+---
+## 10. Voice output 
+
+```bash
+ sudo apt install -y espeak 
+```
+
+## 11. Troubleshooting
+
+### GUI Won't Open
+
+- Restart WSL
+- Check display settings
+
+### Missing Models
+
+Run:
+
+```bash
+ollama pull <model-name>
+```
+
+### Microphone Issues
+
+```bash
+pip install pyaudio
+```
+
+### Qt Plugin Errors
+
+- Verify Qt libraries are properly installed
+
+---

src/chatbot/image_handler.py

@@ -28,7 +28,6 @@
 except Exception as e:
     HAS_PADDLE = False
     print(f"[INIT] PaddleOCR init failed: {e}")
-    print("[INIT] Vision analysis unavailable. Text and netlist analysis still work.")
 
 
 def encode_image(image_path: str) -> str:

src/chatbot/ollama_runner.py

@@ -1,90 +1,40 @@
 import os
 import ollama
-import json
-import time
+import json,time
 
-# ==================== CLIENT ====================
+# Model configuration
+VISION_MODELS = {"primary": "minicpm-v:latest"}
+TEXT_MODELS = {"default": "qwen2.5:3b"}
+EMBED_MODEL = "nomic-embed-text" 
 
 ollama_client = ollama.Client(
     host="http://localhost:11434",
-    timeout=300.0,
+    timeout=300.0,  
 )
 
-# ==================== SETTINGS ====================
-
-_SETTINGS_DIR = os.path.join(
-    os.path.expanduser("~"), ".local", "share", "esim-copilot"
-)
-_SETTINGS_PATH = os.path.join(_SETTINGS_DIR, "settings.json")
-
-_DEFAULT_TEXT_MODEL = "qwen2.5:3b"
-_DEFAULT_VISION_MODEL = "minicpm-v:latest"
-EMBED_MODEL = "nomic-embed-text"
-
-
-def load_model_settings() -> dict:
-    """Load persisted model preferences from disk."""
-    try:
-        with open(_SETTINGS_PATH, "r", encoding="utf-8") as f:
-            return json.load(f)
-    except Exception:
-        return {}
-
-
-def save_model_settings(text_model: str, vision_model: str) -> None:
-    """Persist model preferences to disk."""
-    os.makedirs(_SETTINGS_DIR, exist_ok=True)
-    try:
-        with open(_SETTINGS_PATH, "w", encoding="utf-8") as f:
-            json.dump({"text_model": text_model, "vision_model": vision_model}, f, indent=2)
-    except Exception as e:
-        print(f"[SETTINGS] Failed to save: {e}")
-
-
-def list_available_models() -> list:
-    """Query Ollama for installed models. Returns list of model name strings."""
-    try:
-        resp = ollama_client.list()
-        names = [m["name"] for m in resp.get("models", [])]
-        return names if names else [_DEFAULT_TEXT_MODEL, _DEFAULT_VISION_MODEL]
-    except Exception:
-        return [_DEFAULT_TEXT_MODEL, _DEFAULT_VISION_MODEL, EMBED_MODEL]
-
-
-# Load settings and initialise model dicts
-_settings = load_model_settings()
-
-VISION_MODELS = {"primary": _settings.get("vision_model", _DEFAULT_VISION_MODEL)}
-TEXT_MODELS   = {"default": _settings.get("text_model",   _DEFAULT_TEXT_MODEL)}
-
-
-def reload_model_settings() -> None:
-    """Re-read settings from disk and update running dicts (called after save)."""
-    s = load_model_settings()
-    VISION_MODELS["primary"] = s.get("vision_model", _DEFAULT_VISION_MODEL)
-    TEXT_MODELS["default"]   = s.get("text_model",   _DEFAULT_TEXT_MODEL)
-
-
-# ==================== VISION ====================
-
-def run_ollama_vision(prompt: str, image_input) -> str:
-    """Call vision model with Chain-of-Thought for better accuracy."""
+def run_ollama_vision(prompt: str, image_input: str | bytes) -> str:
+    """Call minicpm-v:latest with Chain-of-Thought for better accuracy."""
     model = VISION_MODELS["primary"]
-
+    
     try:
         import base64
-
+        
         image_b64 = ""
+
+
         if isinstance(image_input, bytes):
             image_b64 = base64.b64encode(image_input).decode("utf-8")
-        elif isinstance(image_input, str) and os.path.isfile(image_input):
+
+        elif os.path.isfile(image_input):
             with open(image_input, "rb") as f:
                 image_b64 = base64.b64encode(f.read()).decode("utf-8")
+
         elif isinstance(image_input, str) and len(image_input) > 100:
-            image_b64 = image_input
+             image_b64 = image_input
         else:
-            raise ValueError("Invalid image input format")
+             raise ValueError("Invalid image input format")
 
+        # === CHAIN OF THOUGHT ===
         system_prompt = (
             "You are an expert Electronics Engineer using eSim.\n"
             "Analyze the schematic image carefully.\n\n"
@@ -115,7 +65,7 @@ def run_ollama_vision(prompt: str, image_input) -> str:
                 {
                     "role": "user",
                     "content": prompt,
-                    "images": [image_b64],
+                    "images": [image_b64], # <--- MUST BE LIST OF BASE64 STRINGS
                 },
             ],
             options={
@@ -126,17 +76,18 @@ def run_ollama_vision(prompt: str, image_input) -> str:
         )
 
         content = resp["message"]["content"]
-
+
+        # === PARSE JSON FROM MIXED OUTPUT ===
         import re
         json_match = re.search(r'```json\s*(\{.*?\})\s*```', content, re.DOTALL)
         if json_match:
             return json_match.group(1)
-
+        
         start = content.find('{')
         end = content.rfind('}') + 1
         if start != -1 and end != -1:
             return content[start:end]
-
+            
         return "{}"
 
     except Exception as e:
@@ -145,45 +96,44 @@ def run_ollama_vision(prompt: str, image_input) -> str:
             "vision_summary": f"Vision failed: {str(e)[:50]}",
             "component_counts": {},
             "circuit_analysis": {"circuit_type": "Error", "design_errors": [], "design_warnings": []},
-            "components": [],
-            "values": {},
+            "components": [], "values": {}
         })
 
-
-# ==================== TEXT ====================
-
 def run_ollama(prompt: str, mode: str = "default") -> str:
-    """Run text model with focused parameters."""
+    """
+    OPTIMIZED: Run text model with focused parameters.
+    """
     model = TEXT_MODELS.get(mode, TEXT_MODELS["default"])
-
+    
     try:
         resp = ollama_client.chat(
             model=model,
             messages=[
                 {
                     "role": "system",
-                    "content": "You are an eSim and electronics expert. Be concise, accurate, and practical.",
+                    "content": "You are an eSim and electronics expert. Be concise, accurate, and practical."
                 },
                 {"role": "user", "content": prompt},
             ],
             options={
-                "temperature": 0.05,
-                "num_ctx": 2048,
-                "num_predict": 400,
+                "temperature": 0.05, 
+                "num_ctx": 2048,      
+                "num_predict": 400,   
                 "top_p": 0.9,
-                "repeat_penalty": 1.1,
+                "repeat_penalty": 1.1,  
             },
         )
+
         return resp["message"]["content"].strip()
-
+    
     except Exception as e:
         return f"[Error] {str(e)}"
 
 
-# ==================== EMBEDDINGS ====================
-
 def get_embedding(text: str):
-    """Get text embeddings for RAG."""
+    """
+    OPTIMIZED: Get text embeddings for RAG.
+    """
     try:
         r = ollama_client.embeddings(model=EMBED_MODEL, prompt=text)
         return r["embedding"]