Upload 4 files

app.py

@@ -0,0 +1,67 @@
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import gradio as gr
+
+# --- 1. Load Roadmap and Rules ---
+def load_roadmap_and_rules(roadmap_path="roadmap.txt", rules_path="rules.txt"):
+    roadmap_content = {}
+    with open(roadmap_path, 'r') as f:
+        current_section = None
+        for line in f:
+            line = line.strip()
+            if line.startswith("#"):
+                current_section = line[1:].strip()
+                roadmap_content[current_section] = ""
+            elif current_section:
+                roadmap_content[current_section] += line + "\n"
+    with open(rules_path, 'r') as f:
+        rules_content = f.read()
+    return roadmap_content, rules_content
+
+roadmap, rules = load_roadmap_and_rules()
+
+# --- 2. Load the AI Model (Mistral 7B) ---
+model_name = "mistralai/Mistral-7B-Instruct-v0.2"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name)
+
+# --- 3. Function to Get Chatbot Response ---
+def get_chatbot_response(user_query, project_phase, roadmap, rules, model, tokenizer):
+    phase_roadmap_section = roadmap.get(project_phase, "General Guidance")
+    context = f"""
+    Project Roadmap (Phase: {project_phase}):
+    {phase_roadmap_section}
+
+    Project Rules:
+    {rules}
+
+    User Query: {user_query}
+
+    ---
+    Provide helpful guidance for this project.
+    """
+    prompt = f"<s>[INST] {context} [/INST]"
+    inputs = tokenizer(prompt, return_tensors="pt")
+    outputs = model.generate(**inputs, max_new_tokens=300)
+    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    response_start_index = response.find("[/INST]") + len("[/INST]") if "[/INST]" in response else 0
+    cleaned_response = response[response_start_index:].strip()
+    return cleaned_response
+
+# --- 4. Gradio Interface Function ---
+def chatbot_interface(user_query, project_phase):
+    return get_chatbot_response(user_query, project_phase, roadmap, rules, model, tokenizer)
+
+# --- 5. Gradio Interface Setup ---
+iface = gr.Interface(
+    fn=chatbot_interface,
+    inputs=[
+        gr.Textbox(label="Your Question"),
+        gr.Dropdown(list(roadmap.keys()), label="Project Phase", value=list(roadmap.keys())[0])
+    ],
+    outputs="text",
+    title="Project Guidance Chatbot",
+    description="Ask questions about your project phase and get guidance."
+)
+
+if __name__ == "__main__":
+    iface.launch()

requirements.txt

@@ -0,0 +1,4 @@
+transformers
+torch
+gradio
+accelerate

roadmap.txt

@@ -0,0 +1,14 @@
+# Base Model Selection
+**What to do:** Choose a pre-trained language model from Hugging Face Hub.
+**Considerations:** Project size, complexity, memory.
+**Example:** Mistral 7B is a good start.
+
+# Fine-Tuning
+**What to do:** Train model on your data.
+**Considerations:** Dataset, fine-tuning method.
+**Example:** Use LoRA for efficiency.
+
+# Deployment
+**What to do:** Make chatbot accessible.
+**Considerations:** Platform choice.
+**Example:** Hugging Face Spaces is recommended.

rules.txt

@@ -0,0 +1,3 @@
+Rule 1: Follow the project roadmap.
+Rule 2: Verify generated code.
+Rule 3: Ask if unsure.