Loaded from browser cache
)}
{fileMeta.type || "Unknown type"}
-
- {fileMeta.contentPreview || "(no preview)"}
-
-
+ {/* Table preview when structured data detected; otherwise show text */}
+ {tablePreview && tablePreview.origin === 'csv' ? (
+
+
+
+
+ {tablePreview.headers.map((h, idx) => (
+ | {h} |
+ ))}
+
+
+
+ {tablePreview.rows.map((r, i) => (
+
+ {r.map((c, j) => (
+ | {c} |
+ ))}
+
+ ))}
+
+
+
+ ) : (
+
+ {fileMeta.contentPreview || "(no preview)"}
+
+ )}
+
+
)}
diff --git a/frontend/nordic-privacy-ai/components/try/ChatbotPanel.tsx b/frontend/components/try/ChatbotPanel.tsx
similarity index 100%
rename from frontend/nordic-privacy-ai/components/try/ChatbotPanel.tsx
rename to frontend/components/try/ChatbotPanel.tsx
diff --git a/frontend/nordic-privacy-ai/components/try/Sidebar.tsx b/frontend/components/try/Sidebar.tsx
similarity index 100%
rename from frontend/nordic-privacy-ai/components/try/Sidebar.tsx
rename to frontend/components/try/Sidebar.tsx
diff --git a/frontend/nordic-privacy-ai/next.config.mjs b/frontend/next.config.mjs
similarity index 100%
rename from frontend/nordic-privacy-ai/next.config.mjs
rename to frontend/next.config.mjs
diff --git a/frontend/nordic-privacy-ai/README.md b/frontend/nordic-privacy-ai/README.md
deleted file mode 100644
index b424092..0000000
--- a/frontend/nordic-privacy-ai/README.md
+++ /dev/null
@@ -1,28 +0,0 @@
-# Nordic Privacy AI
-
-AI-Powered GDPR compliance & personal data protection platform tailored for Nordic ecosystems (BankID, MitID, Suomi.fi).
-
-## Tech Stack
-- Next.js (App Router, TypeScript)
-- Tailwind CSS
-
-## Getting Started
-```powershell
-npm install
-npm run dev
-```
-Visit http://localhost:3000 to view the landing page.
-
-## Scripts
-- `npm run dev` β Start dev server
-- `npm run build` β Production build
-- `npm start` β Run built app
-- `npm run lint` β ESLint
-
-## Next Steps
-- Implement /try page workflow
-- Add feature sections & agent explanations
-- Integrate backend services for data upload & scanning
-
-## License
-Internal hackathon prototype
diff --git a/frontend/nordic-privacy-ai/lib/idb.ts b/frontend/nordic-privacy-ai/lib/idb.ts
new file mode 100644
index 0000000..507b85a
--- /dev/null
+++ b/frontend/nordic-privacy-ai/lib/idb.ts
@@ -0,0 +1,113 @@
+/**
+ * IndexedDB utilities for persisting file uploads in the browser.
+ * Stores the latest uploaded file and its metadata for recovery across sessions.
+ */
+
+const DB_NAME = "NordicPrivacyAI";
+const DB_VERSION = 1;
+const STORE_NAME = "latestUpload";
+
+interface UploadedFileMeta {
+ name: string;
+ size: number;
+ type: string;
+ contentPreview: string;
+}
+
+interface LatestUploadData {
+ file: File;
+ meta: UploadedFileMeta;
+ timestamp: number;
+}
+
+/**
+ * Open or create the IndexedDB database
+ */
+function openDB(): Promise
{
+ return new Promise((resolve, reject) => {
+ const request = indexedDB.open(DB_NAME, DB_VERSION);
+
+ request.onerror = () => reject(request.error);
+ request.onsuccess = () => resolve(request.result);
+
+ request.onupgradeneeded = (event) => {
+ const db = (event.target as IDBOpenDBRequest).result;
+
+ // Create object store if it doesn't exist
+ if (!db.objectStoreNames.contains(STORE_NAME)) {
+ db.createObjectStore(STORE_NAME);
+ }
+ };
+ });
+}
+
+/**
+ * Save the latest uploaded file and its metadata to IndexedDB
+ */
+export async function saveLatestUpload(
+ file: File,
+ meta: UploadedFileMeta
+): Promise {
+ const db = await openDB();
+
+ return new Promise((resolve, reject) => {
+ const transaction = db.transaction([STORE_NAME], "readwrite");
+ const store = transaction.objectStore(STORE_NAME);
+
+ const data: LatestUploadData = {
+ file,
+ meta,
+ timestamp: Date.now(),
+ };
+
+ const request = store.put(data, "latest");
+
+ request.onerror = () => reject(request.error);
+ request.onsuccess = () => resolve();
+
+ transaction.oncomplete = () => db.close();
+ });
+}
+
+/**
+ * Retrieve the latest uploaded file and metadata from IndexedDB
+ */
+export async function getLatestUpload(): Promise {
+ const db = await openDB();
+
+ return new Promise((resolve, reject) => {
+ const transaction = db.transaction([STORE_NAME], "readonly");
+ const store = transaction.objectStore(STORE_NAME);
+ const request = store.get("latest");
+
+ request.onerror = () => reject(request.error);
+ request.onsuccess = () => {
+ const result = request.result as LatestUploadData | undefined;
+ if (result) {
+ resolve(result);
+ } else {
+ reject(new Error("No cached upload found"));
+ }
+ };
+
+ transaction.oncomplete = () => db.close();
+ });
+}
+
+/**
+ * Delete the latest upload from IndexedDB
+ */
+export async function deleteLatestUpload(): Promise {
+ const db = await openDB();
+
+ return new Promise((resolve, reject) => {
+ const transaction = db.transaction([STORE_NAME], "readwrite");
+ const store = transaction.objectStore(STORE_NAME);
+ const request = store.delete("latest");
+
+ request.onerror = () => reject(request.error);
+ request.onsuccess = () => resolve();
+
+ transaction.oncomplete = () => db.close();
+ });
+}
diff --git a/frontend/nordic-privacy-ai/next-env.d.ts b/frontend/nordic-privacy-ai/next-env.d.ts
deleted file mode 100644
index 4f11a03..0000000
--- a/frontend/nordic-privacy-ai/next-env.d.ts
+++ /dev/null
@@ -1,5 +0,0 @@
-///
-///
-
-// NOTE: This file should not be edited
-// see https://nextjs.org/docs/basic-features/typescript for more information.
diff --git a/frontend/nordic-privacy-ai/package-lock.json b/frontend/package-lock.json
similarity index 100%
rename from frontend/nordic-privacy-ai/package-lock.json
rename to frontend/package-lock.json
diff --git a/frontend/nordic-privacy-ai/package.json b/frontend/package.json
similarity index 100%
rename from frontend/nordic-privacy-ai/package.json
rename to frontend/package.json
diff --git a/frontend/nordic-privacy-ai/postcss.config.js b/frontend/postcss.config.js
similarity index 100%
rename from frontend/nordic-privacy-ai/postcss.config.js
rename to frontend/postcss.config.js
diff --git a/frontend/nordic-privacy-ai/tailwind.config.ts b/frontend/tailwind.config.ts
similarity index 100%
rename from frontend/nordic-privacy-ai/tailwind.config.ts
rename to frontend/tailwind.config.ts
diff --git a/frontend/nordic-privacy-ai/tsconfig.json b/frontend/tsconfig.json
similarity index 100%
rename from frontend/nordic-privacy-ai/tsconfig.json
rename to frontend/tsconfig.json
diff --git a/requirements.txt b/requirements.txt
index e69a34b..be42f1d 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -9,4 +9,15 @@ setuptools>=65.0.0
presidio-analyzer>=2.2.0
presidio-anonymizer>=2.2.0
spacy>=3.7.0
-# Download spaCy model with: python -m spacy download en_core_web_sm
\ No newline at end of file
+# Download spaCy model with: python -m spacy download en_core_web_sm
+
+# FastAPI Backend
+fastapi>=0.109.0
+uvicorn[standard]>=0.27.0
+python-multipart>=0.0.6
+
+# Optional: GPU Support (uncomment if you have CUDA)
+# torch>=2.0.0 --index-url https://download.pytorch.org/whl/cu121
+
+# Chatbot (WIP - not exposed in API yet)
+gpt4all>=2.0.0
\ No newline at end of file
diff --git a/src/__init__.py b/src/__init__.py
deleted file mode 100644
index fa022e1..0000000
--- a/src/__init__.py
+++ /dev/null
@@ -1,145 +0,0 @@
-"""
-AI Governance Module - Bias Detection and Risk Analysis
-"""
-
-from .data_processor import DataProcessor
-from .model_trainer import GeneralizedModelTrainer
-from .bias_analyzer import BiasAnalyzer
-from .risk_analyzer import RiskAnalyzer
-from .report_generator import ReportGenerator, NumpyEncoder
-
-import pandas as pd
-import json
-
-__version__ = '1.0.0'
-
-__all__ = [
- 'DataProcessor',
- 'GeneralizedModelTrainer',
- 'BiasAnalyzer',
- 'RiskAnalyzer',
- 'ReportGenerator',
- 'NumpyEncoder',
- 'AIGovernanceAnalyzer'
-]
-
-
-class AIGovernanceAnalyzer:
- """
- Main interface for AI Governance analysis
-
- Example:
- >>> analyzer = AIGovernanceAnalyzer()
- >>> report = analyzer.analyze('data.csv', 'target', ['gender', 'age'])
- >>> print(f"Bias Score: {report['summary']['overall_bias_score']:.3f}")
- """
-
- def __init__(self):
- """Initialize the analyzer"""
- self.processor = None
- self.trainer = None
- self.bias_analyzer = None
- self.risk_analyzer = None
- self.report_generator = None
-
- def analyze(self, data_path, target_column, protected_attributes):
- """
- Run complete AI governance analysis from file
-
- Args:
- data_path (str): Path to CSV file
- target_column (str): Name of target column
- protected_attributes (list): List of protected attribute column names
-
- Returns:
- dict: Complete analysis report
- """
- df = pd.read_csv(data_path)
- return self.analyze_dataframe(df, target_column, protected_attributes)
-
- def analyze_dataframe(self, df, target_column, protected_attributes):
- """
- Run complete AI governance analysis from DataFrame
-
- Args:
- df (pd.DataFrame): Input dataframe
- target_column (str): Name of target column
- protected_attributes (list): List of protected attribute column names
-
- Returns:
- dict: Complete analysis report
- """
- # Step 1: Process data
- self.processor = DataProcessor(df)
- self.processor.target_column = target_column
- self.processor.protected_attributes = protected_attributes
- self.processor.prepare_data()
-
- # Step 2: Train model
- self.trainer = GeneralizedModelTrainer(
- self.processor.X_train,
- self.processor.X_test,
- self.processor.y_train,
- self.processor.y_test,
- self.processor.feature_names
- )
- self.trainer.train()
- self.trainer.evaluate()
-
- # Step 3: Analyze bias
- self.bias_analyzer = BiasAnalyzer(
- self.processor.X_test,
- self.processor.y_test,
- self.trainer.y_pred,
- self.processor.df,
- self.processor.protected_attributes,
- self.processor.target_column
- )
- bias_results = self.bias_analyzer.analyze()
-
- # Step 4: Assess risks
- self.risk_analyzer = RiskAnalyzer(
- self.processor.df,
- self.trainer.results,
- bias_results,
- self.processor.protected_attributes,
- self.processor.target_column
- )
- risk_results = self.risk_analyzer.analyze()
-
- # Step 5: Generate report
- self.report_generator = ReportGenerator(
- self.trainer.results,
- bias_results,
- risk_results,
- self.processor.df
- )
-
- return self.report_generator.generate_report()
-
- def save_report(self, report, output_path):
- """
- Save report to JSON file
-
- Args:
- report (dict): Analysis report
- output_path (str): Path to save JSON file
-
- Returns:
- str: Path to saved file
- """
- with open(output_path, 'w') as f:
- json.dump(report, f, indent=2, cls=NumpyEncoder)
- return output_path
-
- def get_summary(self, report):
- """
- Get executive summary from report
-
- Args:
- report (dict): Analysis report
-
- Returns:
- dict: Summary metrics
- """
- return report.get('summary', {})
diff --git a/src/report_generator.py b/src/report_generator.py
deleted file mode 100644
index 3b41cbd..0000000
--- a/src/report_generator.py
+++ /dev/null
@@ -1,263 +0,0 @@
-"""
-Report Generator Module
-Generates comprehensive JSON reports
-"""
-
-import json
-import numpy as np
-from datetime import datetime
-
-class NumpyEncoder(json.JSONEncoder):
- """Custom JSON encoder for numpy types"""
- def default(self, obj):
- if isinstance(obj, (np.integer, np.int64, np.int32)):
- return int(obj)
- elif isinstance(obj, (np.floating, np.float64, np.float32)):
- return float(obj)
- elif isinstance(obj, (np.ndarray,)):
- return obj.tolist()
- elif isinstance(obj, (np.bool_,)):
- return bool(obj)
- return super(NumpyEncoder, self).default(obj)
-
-class ReportGenerator:
- """Generate comprehensive analysis reports"""
-
- def __init__(self, model_results, bias_results, risk_results, df):
- self.model_results = model_results
- self.bias_results = bias_results
- self.risk_results = risk_results
- self.df = df
-
- def generate_report(self):
- """Generate comprehensive JSON report"""
- report = {
- 'metadata': self._generate_metadata(),
- 'summary': self._generate_summary(),
- 'model_performance': self._format_model_results(),
- 'bias_analysis': self._format_bias_results(),
- 'risk_assessment': self._format_risk_results(),
- 'key_findings': self._extract_key_findings(),
- 'recommendations': self._compile_recommendations(),
- 'detailed_metrics': self._compile_detailed_metrics()
- }
-
- return report
-
- def _generate_metadata(self):
- """Generate report metadata"""
- return {
- 'report_id': f"AIGov_{datetime.now().strftime('%Y%m%d_%H%M%S')}",
- 'generated_at': datetime.now().isoformat(),
- 'report_version': '1.0',
- 'dataset_info': {
- 'total_records': len(self.df),
- 'total_features': len(self.df.columns),
- 'columns': list(self.df.columns)
- }
- }
-
- def _generate_summary(self):
- """Generate executive summary"""
- model_metrics = self.model_results.get('metrics', {})
-
- return {
- 'overall_bias_score': self.bias_results.get('overall_bias_score', 0.0),
- 'overall_risk_score': self.risk_results.get('overall_risk_score', 0.0),
- 'risk_level': self.risk_results.get('risk_level', 'UNKNOWN'),
- 'model_accuracy': model_metrics.get('accuracy', 0.0),
- 'fairness_violations_count': len(self.bias_results.get('fairness_violations', [])),
- 'passes_fairness_threshold': self.bias_results.get('fairness_assessment', {}).get('passes_fairness_threshold', False)
- }
-
- def _format_model_results(self):
- """Format model performance results"""
- return {
- 'model_type': self.model_results.get('model_type', 'Unknown'),
- 'metrics': self.model_results.get('metrics', {}),
- 'confusion_matrix': self.model_results.get('confusion_matrix', []),
- 'top_features': dict(list(self.model_results.get('feature_importance', {}).items())[:10])
- }
-
- def _format_bias_results(self):
- """Format bias analysis results"""
- return {
- 'overall_bias_score': self.bias_results.get('overall_bias_score', 0.0),
- 'fairness_metrics': self.bias_results.get('fairness_metrics', {}),
- 'fairness_violations': self.bias_results.get('fairness_violations', []),
- 'fairness_assessment': self.bias_results.get('fairness_assessment', {}),
- 'demographic_bias_summary': self._summarize_demographic_bias()
- }
-
- def _format_risk_results(self):
- """Format risk assessment results"""
- return {
- 'overall_risk_score': self.risk_results.get('overall_risk_score', 0.0),
- 'risk_level': self.risk_results.get('risk_level', 'UNKNOWN'),
- 'risk_categories': self.risk_results.get('risk_categories', {}),
- 'privacy_risks': self._summarize_privacy_risks(),
- 'ethical_risks': self._summarize_ethical_risks()
- }
-
- def _summarize_demographic_bias(self):
- """Summarize demographic bias"""
- demo_bias = self.bias_results.get('demographic_bias', {})
- summary = {}
-
- for attr, data in demo_bias.items():
- summary[attr] = {
- 'max_disparity': data.get('max_disparity', 0),
- 'groups_analyzed': len(data.get('approval_rates', {}))
- }
-
- return summary
-
- def _summarize_privacy_risks(self):
- """Summarize privacy risks"""
- privacy = self.risk_results.get('privacy_risks', {})
-
- return {
- 'pii_count': len(privacy.get('pii_detected', [])),
- 'anonymization_level': privacy.get('anonymization_level', 'UNKNOWN'),
- 'exposure_risk_count': len(privacy.get('exposure_risks', [])),
- 'gdpr_compliance_score': privacy.get('gdpr_compliance', {}).get('compliance_score', 0)
- }
-
- def _summarize_ethical_risks(self):
- """Summarize ethical risks"""
- ethical = self.risk_results.get('ethical_risks', {})
-
- return {
- 'fairness_issues_count': len(ethical.get('fairness_issues', [])),
- 'transparency_score': ethical.get('transparency_score', 0),
- 'bias_amplification_risk': ethical.get('bias_amplification_risk', 'UNKNOWN'),
- 'social_impact': ethical.get('social_impact_assessment', {})
- }
-
- def _extract_key_findings(self):
- """Extract key findings from analysis"""
- findings = []
-
- # Model performance findings
- accuracy = self.model_results.get('metrics', {}).get('accuracy', 0)
- if accuracy >= 0.8:
- findings.append(f"β Model achieves good accuracy ({accuracy:.2%})")
- else:
- findings.append(f"β Model accuracy is below optimal ({accuracy:.2%})")
-
- # Bias findings
- bias_score = self.bias_results.get('overall_bias_score', 0)
- if bias_score < 0.3:
- findings.append("β Low bias detected across protected attributes")
- elif bias_score < 0.5:
- findings.append("β Moderate bias detected - monitoring recommended")
- else:
- findings.append("β High bias detected - immediate action required")
-
- # Fairness violations
- violations = self.bias_results.get('fairness_violations', [])
- if violations:
- high_sev = sum(1 for v in violations if v['severity'] == 'HIGH')
- findings.append(f"β {len(violations)} fairness violations detected ({high_sev} high severity)")
- else:
- findings.append("β No fairness violations detected")
-
- # Privacy findings
- privacy = self.risk_results.get('privacy_risks', {})
- pii_count = len(privacy.get('pii_detected', []))
- if pii_count > 0:
- findings.append(f"β {pii_count} columns contain potential PII")
- else:
- findings.append("β No obvious PII detected in dataset")
-
- # Risk level
- risk_level = self.risk_results.get('risk_level', 'UNKNOWN')
- findings.append(f"Overall Risk Level: {risk_level}")
-
- return findings
-
- def _compile_recommendations(self):
- """Compile all recommendations"""
- recommendations = []
-
- # Get recommendations from each component
- privacy_recs = self.risk_results.get('privacy_risks', {}).get('recommendations', [])
- ethical_recs = self.risk_results.get('ethical_risks', {}).get('recommendations', [])
- performance_recs = self.risk_results.get('model_performance_risks', {}).get('recommendations', [])
- compliance_recs = self.risk_results.get('compliance_risks', {}).get('recommendations', [])
-
- # Prioritize recommendations
- all_recs = []
-
- # High priority (from violations and high risks)
- violations = self.bias_results.get('fairness_violations', [])
- if violations:
- all_recs.append({
- 'priority': 'HIGH',
- 'category': 'Fairness',
- 'recommendation': 'Address fairness violations in protected attributes'
- })
-
- if len(privacy_recs) > 0:
- all_recs.append({
- 'priority': 'HIGH',
- 'category': 'Privacy',
- 'recommendation': privacy_recs[0]
- })
-
- # Medium priority
- for rec in ethical_recs[:2]:
- all_recs.append({
- 'priority': 'MEDIUM',
- 'category': 'Ethics',
- 'recommendation': rec
- })
-
- # Lower priority
- for rec in performance_recs[:2]:
- all_recs.append({
- 'priority': 'MEDIUM',
- 'category': 'Performance',
- 'recommendation': rec
- })
-
- for rec in compliance_recs[:2]:
- all_recs.append({
- 'priority': 'MEDIUM',
- 'category': 'Compliance',
- 'recommendation': rec
- })
-
- # Convert to simple list with formatting
- recommendations = [
- f"[{r['priority']}] {r['category']}: {r['recommendation']}"
- for r in all_recs[:10] # Limit to top 10
- ]
-
- return recommendations
-
- def _compile_detailed_metrics(self):
- """Compile detailed metrics for analysis"""
- return {
- 'bias_metrics': {
- 'by_attribute': self.bias_results.get('fairness_metrics', {}),
- 'demographic_analysis': self.bias_results.get('demographic_bias', {})
- },
- 'risk_breakdown': {
- 'privacy': self.risk_results.get('privacy_risks', {}),
- 'ethical': self.risk_results.get('ethical_risks', {}),
- 'compliance': self.risk_results.get('compliance_risks', {}),
- 'data_quality': self.risk_results.get('data_quality_risks', {})
- },
- 'model_details': {
- 'classification_report': self.model_results.get('classification_report', {}),
- 'feature_importance': self.model_results.get('feature_importance', {})
- }
- }
-
- def save_report(self, filepath):
- """Save report to JSON file"""
- report = self.generate_report()
- with open(filepath, 'w') as f:
- json.dump(report, f, indent=2, cls=NumpyEncoder)
- return filepath
diff --git a/start_api.py b/start_api.py
new file mode 100644
index 0000000..754bcc8
--- /dev/null
+++ b/start_api.py
@@ -0,0 +1,21 @@
+"""
+Start the FastAPI server
+Run: python start_api.py
+"""
+
+import uvicorn
+
+if __name__ == "__main__":
+ print("π Starting Nordic Privacy AI API Server...")
+ print("π API will be available at: http://localhost:8000")
+ print("π Interactive docs at: http://localhost:8000/docs")
+ print("π Frontend should run at: http://localhost:3000")
+ print("\nPress CTRL+C to stop\n")
+
+ uvicorn.run(
+ "api.main:app",
+ host="0.0.0.0",
+ port=8000,
+ reload=True, # Auto-reload on code changes
+ log_level="info"
+ )
diff --git a/test_cleaning.py b/test_cleaning.py
index bcdfb91..65c9968 100644
--- a/test_cleaning.py
+++ b/test_cleaning.py
@@ -10,7 +10,7 @@ import os
# Add parent directory to path
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
-from cleaning import DataCleaner, CleaningConfig
+from data_cleaning import DataCleaner
def test_basic_cleaning():