Feature/ai pr analyser

.github/workflows/pr_analyzer.yml

@@ -0,0 +1,48 @@
+name: Automated V2 PR Analyzer
+
+on:
+  pull_request:
+    types: [opened, synchronize]
+
+jobs:
+  analyze_pr:
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout Repository
+        uses: actions/checkout@v4
+
+      - name: Set up Python Environment
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.10'
+
+      - name: Install System Dependencies
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y libpango-1.0-0 libpangoft2-1.0-0
+
+      - name: Install Python Libraries
+        run: |
+          python -m pip install --upgrade pip
+          # Point pip to the new subfolder
+          pip install -r tools/ai_pr_analyzer/requirements.txt
+
+      - name: Execute V2 PR Analyzer Pipeline
+        env:
+          GEMINI_API_KEY: ${{ secrets.GEMINI_API_KEY }} 
+          GITHUB_PAT: ${{ secrets.GITHUB_TOKEN }}
+          # Grab just the integer ID of the Pull Request
+          PR_NUMBER: ${{ github.event.pull_request.number }}
+        run: |
+          cd tools/ai_pr_analyzer
+          # Pass only the number to your script
+          python main.py $PR_NUMBER
+
+      - name: Upload PDF Dashboard as Artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: BioDynaMo-PR-Analysis-Report
+          # Pull the generated PDF from the new path
+          path: tools/ai_pr_analyzer/output/*.pdf
+          retention-days: 14

tools/ai_pr_analyzer/ai/llm_gemini.py

@@ -0,0 +1,153 @@
+
+GEMINI_API_KEY = ""  
+import re
+import os
+from google import genai
+from google.genai import types
+
+def generate_ai_insights(metrics, files_data, complexity_warnings="None"):
+    """Generates PR analysis text using Gemini 2.5 Flash."""
+    file_list = "\n".join(list(files_data.keys()))
+
+
+    #  THE ANALYST (Text Generation)
+    # ==================----------------===============---------------===========-------------
+    prompt_text01 = f"""
+    You are a Staff Software Engineer analyzing a Pull Request for BioDynaMo.
+
+    Quantitative Metrics:
+    - Title: {metrics.get('title', 'N/A')}
+    - Lines Added: {metrics.get('additions', 0)}
+    - Lines Removed: {metrics.get('deletions', 0)}
+
+    Files Changed:
+    {file_list}
+
+    High Complexity Warnings:
+    {complexity_warnings}
+
+    You are an expert C++ software engineer and open-source maintainer.
+    Analyze the following GitHub Pull Request and write a professional technical analysis report.
+
+    You must format your response exactly matching this 6-point rubric:
+    PURPOSE: Purpose and Functionality
+    POSITIVES: Positive Aspects.
+    CONCERNS: Potential Concerns or Weaknesses and Indicators of potential architectural issues, code smells, or spaghetti code.
+    CODE_QUALITY: Code Quality and Maintainability Considerations.
+
+
+    """
+
+    print("Generating architectural analysis text with Gemini 2.5 Flash...")
+
+    #try:
+        # Initialize the official Google GenAI client
+        # This automatically looks for the GEMINI_API_KEY environment variable
+       #client = genai.Client(api_key=GEMINI_API_KEY)
+
+
+        #response = client.models.generate_content(
+            #model='gemini-2.5-flash',
+            #contents=prompt_text01,
+            #config=types.GenerateContentConfig(
+                #temperature=0.2, # Slight creativity for writing
+           # ),
+        #)
+        #text_output01 = response.text.replace("**", "").replace("## ", "")
+
+    #except Exception as e:
+       # print(f" Gemini API Error: {e}")
+        #return "Error", "Error", "Error", "Error", "Error", "Error"
+
+    prompt_text02 = f"""
+    You are a Staff Software Engineer analyzing a Pull Request for BioDynaMo.
+
+    Quantitative Metrics:
+    - Title: {metrics.get('title', 'N/A')}
+    - Lines Added: {metrics.get('additions', 0)}
+    - Lines Removed: {metrics.get('deletions', 0)}
+
+    Files Changed:
+    {file_list}
+
+    High Complexity Warnings:
+    {complexity_warnings}
+
+    You are an expert C++ software engineer and open-source maintainer.
+    Analyze the following GitHub Pull Request and write a professional technical analysis report.
+
+    Provide your response using EXACTLY this heading:
+    IMPACT: Architectural Impact Assessment, that is Assessment of the likely functional impact and breadth of the changes.
+    ACTION_PLAN : A prioritised list of specific recommendations for improvement.
+    SUMMARY : Comprehensive explanation of the contribution overall.
+    END: ending by
+
+    """
+    prompt_text03 = f"""
+
+    You are a Staff Software Engineer analyzing a Pull Request for BioDynaMo.
+    Analyze the following GitHub Pull Request and based on the following GitHub Pull Request analysis, create a prioritized action plan for the developer.
+
+    Quantitative Metrics:
+    - Title: {metrics.get('title', 'N/A')}
+    - Lines Added: {metrics.get('additions', 0)}
+    - Lines Removed: {metrics.get('deletions', 0)}
+
+    Files Changed:
+    {file_list}
+
+    High Complexity Warnings:
+    {complexity_warnings}
+
+    Provide your response using EXACTLY this heading:
+    SUMMARY : Comprehensive explanation of the contribution overall.
+    """
+
+    def response_out(ptext):
+        client = genai.Client(api_key=GEMINI_API_KEY)
+        response_2 = client.models.generate_content(
+        model='gemini-2.5-flash',
+        contents=ptext,)
+
+        return response_2
+
+    text_output01 = response_out(prompt_text01).text.replace("**", "").replace("## ", "")
+    text_output02 =response_out(prompt_text02).text.replace("**", "").replace("## ", "")
+    text_output03 =response_out(prompt_text03).text.replace("**", "").replace("## ", "")
+
+    print(text_output01)
+    print(text_output02)
+    #print (text_output03)
+
+    # Finally, combine them for your parser, or parse them separately!
+    #text_output = text_output01 + "\n" + action_plan_text
+
+
+    # PARSING ENGINE
+    # ===================------------===========---------===========---------
+
+    def extract_section(text, start_keyword, end_keywords):
+        end_pattern = "|".join([f"{k}:?" for k in end_keywords]) + "|$"
+        pattern = rf"{start_keyword}:?\s*(.*?)(?={end_pattern})"
+        match = re.search(pattern, text, re.IGNORECASE | re.DOTALL)
+        return match.group(1).strip() if match else f"Data not generated for {start_keyword}."
+
+    try:
+        purpose = extract_section(text_output01, "PURPOSE: Purpose and Functionality", ["POSITIVES: Positive Aspects", "CONCERNS: Potential Concerns or Weaknesses and Indicators of potential architectural issues, code smells, or spaghetti code", "CODE_QUALITY: Code Quality and Maintainability Considerations", "IMPACT: Architectural Impact Assessment, that is Assessment of the likely functional impact and breadth of the changes"])
+        positives = extract_section(text_output01, "POSITIVES: Positive Aspects", ["CONCERNS: Potential Concerns or Weaknesses and Indicators of potential architectural issues, code smells, or spaghetti code", "CODE_QUALITY: Code Quality and Maintainability Considerations", "IMPACT: Architectural Impact Assessment, that is Assessment of the likely functional impact and breadth of the changes"])
+        concerns = extract_section(text_output01, "CONCERNS: Potential Concerns or Weaknesses and Indicators of potential architectural issues, code smells, or spaghetti code", ["CODE_QUALITY: Code Quality and Maintainability Considerations", "IMPACT: Architectural Impact Assessment, that is Assessment of the likely functional impact and breadth of the changes"])
+        quality = extract_section(text_output01, "CODE_QUALITY: Code Quality and Maintainability Considerations", ["IMPACT: Architectural Impact Assessment, that is Assessment of the likely functional impact and breadth of the changes"])
+        impact = extract_section(text_output02, "IMPACT: Architectural Impact Assessment, that is Assessment of the likely functional impact and breadth of the changes",["ACTION_PLAN : A prioritised list of specific recommendations for improvement","SUMMARY : Comprehensive explanation of the contribution overall","END: ending by"])
+        action_plan = extract_section(text_output02, "ACTION_PLAN : A prioritised list of specific recommendations for improvement",["SUMMARY : Comprehensive explanation of the contribution overall","END: ending by"])
+        summary = extract_section(text_output02, "SUMMARY : Comprehensive explanation of the contribution overall",["END: ending by"])
+        print(summary)
+
+    except Exception as e:
+        print(f">*< >*< >*< >*< >*< >*< >*< Parsing Error: {e}")
+        return "Error", "Error", "Error", "Error", "Error", "Error"
+
+
+    return purpose, positives, concerns, quality, impact, action_plan, summary
+
+
+    #return text_output01, ap_text_output

tools/ai_pr_analyzer/analysis/static_analyzer.py

@@ -0,0 +1,50 @@
+import os
+import base64
+import lizard
+import matplotlib.pyplot as plt
+
+
+def generate_complexity_chart(file_data):
+    #calculates cyclomatric complexity using lizard and output a Base64 PNG
+
+    filenames, complexities = [], []
+    for filename, code_string in file_data.items():
+        #feeding raw c++ code into lizard, so it can scan and find the code complexity.
+        analysis = lizard.analyze_file.analyze_source_code(filename,code_string)
+        # we are complexity of the by taking avg code complexty as the sum of number of cyclomatic_complexty by number of functions. 
+        avg_ccn = sum(f.cyclomatic_complexity for f in analysis.function_list) / len(analysis.function_list) if analysis.function_list else 0
+        filenames.append(filename.split('/')[-1])
+        complexities.append(avg_ccn)
+
+    if not filenames:
+        return None
+
+    #plotting graph for file name vs code complexity
+
+    plt.figure(figsize=(8,4))
+    colors = ['#e74c34' if c > 10 else '#f39c12' if c > 5 else '#27ae60' for c in complexities]
+    plt.bar(filenames, complexities, color = colors)
+    plt.axhline( y= 10, color='r', linestyle ='--', label='High Risk (CCN > 10)')
+    plt.title('Code Complexities Impact by File')
+    plt.ylabel('Avg Clyclomatric Complexity')
+    plt.xticks(rotation =45, ha= 'right')
+    plt.tight_layout()
+
+    chart_path = 'temp_complexity.png'
+    plt.savefig(chart_path, dpi =150)
+    plt.close()
+
+    # reopening sved plot as read binary and encodes it into standard base64 string , so that it can be weasyprint(it can read base64 sting in html)
+    #can save local memory 
+    with open(chart_path, 'rb') as f:
+        chart_b64 = base64.b64encode(f.read()).decode('utf-8')
+    os.remove(chart_path)
+    return chart_b64    
+
+
+
+
+
+
+
+

tools/ai_pr_analyzer/api/github_client.py

@@ -0,0 +1,56 @@
+import os
+import requests
+
+
+GEMINI_API_KEY =""
+GITHUB_TOKEN=""
+
+#def extract_pr_details(url):
+
+def get_pr_data(owner, repo, pr_number):
+
+    #fetching quantiataive metrics and changed file contents from the PR
+    GITHUB_TOKEN= os.getenv("")
+    headers ={"Accept": "application/vnd.github.v3+json"}  
+
+    if GITHUB_TOKEN:
+        headers["Authorization"] = f"token {GITHUB_TOKEN}"
+    else:
+        print("Warning: No GitHub Token found. Proceeding with anonymous API limits.") 
+
+    # Get PR meta and metrics
+
+    pr_url = f"https://api.github.com/repos/{owner}/{repo}/pulls/{pr_number}"
+    pr_res = requests.get(pr_url,headers=headers)
+    pr_res.raise_for_status()
+    pr_info = pr_res.json()
+
+    # extrating informations
+    metrics ={
+        "additions": pr_info.get("additions", 0),
+        "deletions": pr_info.get("deletions",0),
+        "changed_files": pr_info.get("changed_files"),
+        "total_lines": pr_info.get("additions",0) + pr_info.get("deletions",0),
+        "title": pr_info.get("title",""),
+        "description": pr_info.get("body","")
+
+    }  
+
+    #get changed files for statistical analysis
+
+    files_url = f"https://api.github.com/repos/{owner}/{repo}/pulls/{pr_number}/files"
+    files_res = requests.get(files_url,headers=headers)
+    files_res.raise_for_status()
+
+    files_data = {}
+    for f in files_res.json():
+        filename = f["filename"]
+        if filename.endswith(('.cpp','.cc','.h','.hpp'))and f['status'] != 'removed':
+            raw_url = f["raw_url"]
+            raw_res = requests.get(raw_url, headers=headers)
+            if raw_res.status_code == 200:
+                files_data[filename] =raw_res.text
+
+    return metrics, files_data            
+
+

tools/ai_pr_analyzer/main.py

@@ -0,0 +1,42 @@
+import sys
+from dotenv import load_dotenv
+
+# Load environment variables before doing anything else
+load_dotenv()
+
+# Import our custom modules
+from api.github_client import get_pr_data
+from analysis.static_analyzer import generate_complexity_chart
+from ai.llm_gemini import generate_ai_insights
+from reporting.pdf_02_generator import generate_pdf_report
+
+if __name__ == "__main__":
+    OWNER = "BioDynaMo"
+    REPO = "biodynamo"
+    PR_NUMBER = 471
+
+    if len(sys.argv) > 1:
+        # sys.argv[1] is the number we passed from the YAML file
+        PR_NUMBER = int(sys.argv[1]) 
+        print(f"CI/CD Trigger Detected! Targeting PR #{PR_NUMBER}")
+
+    print(f"Fetching data for {OWNER}/{REPO} PR #{PR_NUMBER}...")
+    metrics, files_data = get_pr_data(OWNER, REPO, PR_NUMBER)
+
+    print("*-----> Running static C++ complexity analysis...")
+    chart_b64 = generate_complexity_chart(files_data)
+
+
+    # Phase 3: AI Brain
+    print("*---> Phase 3: Generating AI insights...")
+    purpose, positives, concerns, quality, impact, action_plan, summary = generate_ai_insights(metrics, files_data)
+
+    # Phase 4: Flowchart Rendering
+    #print("🎨 Phase 4: Rendering architecture diagrams...")
+    #mermaid_svg = render_mermaid_to_svg(mermaid_code)
+
+
+    # Phase 5: PDF Assembly
+    print("*----> Phase 5: Compiling final PDF report...")
+    output_filename = f"PR_{PR_NUMBER}_Analysis.pdf"
+    generate_pdf_report(metrics, purpose, positives, concerns, quality, impact, action_plan, summary, PR_NUMBER, chart_b64, output_filename)