vaxtrust_covid_setup.py

#!/usr/bin/env python3
"""
VaxTrust - COVID-19 Vaccine Transparency Platform
Hackathon MVP - India Focus

Core Concept:
- Physical QR codes at vaccination centers
- Patients scan after vaccination
- See real feedback: side effects, recovery time, local trust scores
- AI detects unusual patterns/anomalies
- Build confidence in vaccines through transparency
"""

import os
import requests
import pandas as pd
import json
from datetime import datetime, timedelta
import subprocess

class VaxTrustCOVIDDataManager:
    """COVID-19 specific data pipeline for VaxTrust MVP"""

    def __init__(self, download_dir="./vaxtrust_covid_data"):
        self.download_dir = download_dir
        os.makedirs(download_dir, exist_ok=True)
        self.batch_data = {}
        self.aefi_data = None
        self.vaccination_data = None
        print(f"[✓] VaxTrust COVID-19 Data Manager initialized")
        print(f"[✓] Data directory: {download_dir}")

    # ============================================================
    # DATASET 1: India COVID Vaccination Numbers (OWID)
    # ============================================================
    def download_covid_vaccination_data(self):
        """
        Downloads India COVID-19 vaccination data from Our World in Data
        - Daily vaccination numbers by vaccine type
        - Covers Jan 2021 - Present
        - 2.2 billion+ doses
        """
        print("\n" + "="*60)
        print("[DATASET 1] Downloading COVID-19 Vaccination Data")
        print("="*60)

        url = "https://github.com/owid/covid-19-data/raw/master/public/data/vaccinations/country_data/India.csv"
        filepath = os.path.join(self.download_dir, "covid_vaccination_india.csv")

        try:
            print("[→] Fetching from Our World in Data...")
            response = requests.get(url, timeout=30)
            response.raise_for_status()

            with open(filepath, 'w') as f:
                f.write(response.text)

            df = pd.read_csv(filepath)
            self.vaccination_data = df

            # Summary
            print(f"[✓] Downloaded {len(df)} daily records")
            print(f"[✓] Date range: {df['date'].min()} to {df['date'].max()}")
            print(f"[✓] Vaccines tracked: {df['vaccine'].unique().tolist()}")
            print(f"[✓] File size: {os.path.getsize(filepath) / 1024:.2f} KB")

            # Key stats
            latest = df.iloc[-1]
            print(f"[✓] Latest (as of {latest['date']}):")
            print(f"    - Total doses: {latest['total_vaccinations']:,.0f}")
            print(f"    - Fully vaccinated: {latest['people_fully_vaccinated']:,.0f}")

            return filepath
        except Exception as e:
            print(f"[✗] Error: {e}")
            return None

    # ============================================================
    # DATASET 2: AEFI Adverse Events (Ministry of Health)
    # ============================================================
    def generate_covid_aefi_data(self, num_records=600):
        """
        Generates synthetic COVID-19 AEFI data based on MOHFW patterns

        Real data source: Ministry of Health & Family Welfare
        - 2,708+ serious COVID AEFI cases reported
        - Causality assessed by WHO guidelines
        - Available in MOHFW PDF reports

        For MVP: Synthetic data matching real structure
        """
        print("\n" + "="*60)
        print("[DATASET 2] Generating COVID-19 AEFI Data")
        print("="*60)

        import random

        # COVID vaccines available in India
        covid_vaccines = ['COVISHIELD', 'COVAXIN', 'SPUTNIK_V']

        # Indian states
        states = [
            'AndhraPradesh', 'Karnataka', 'Maharashtra', 'Delhi', 'UttarPradesh',
            'WestBengal', 'Gujarat', 'Rajasthan', 'Punjab', 'Odisha',
            'Haryana', 'Madhya Pradesh', 'Telangana', 'Bihar', 'Jharkhand'
        ]

        # COVID-specific outcomes
        covid_outcomes = [
            'HOSPITALIZED_RECOVERED', 'DEATH', 'RECOVERED_AT_HOME',
            'ONGOING_TREATMENT', 'RECOVERED_WITH_SEQUELAE'
        ]

        # COVID-specific diagnoses (actual side effects reported)
        covid_diagnoses = [
            'FEVER', 'MYALGIA', 'HEADACHE', 'FATIGUE',
            'THROMBOEMBOLISM', 'MYOCARDITIS', 'PERICARDITIS',
            'BELL_PALSY', 'GUILLAIN_BARRE_SYNDROME',
            'ANAPHYLAXIS', 'ALLERGIC_REACTION',
            'LYMPHADENOPATHY', 'SEIZURE', 'SYNCOPE'
        ]

        # Classification by AEFI committee
        # A = Causal (vaccine caused it)
        # B = Indeterminate
        # C = Coincidental (not vaccine related)
        classifications = [
            'A1', 'A1', 'A1',  # More weight to A1 (product-related)
            'A4', 'A4',  # Anxiety reactions
            'B1', 'B1', 'B2',  # Indeterminate
            'C', 'C', 'C', 'C', 'C'  # Most are coincidental (5:1 ratio)
        ]

        print(f"[→] Generating {num_records} COVID-19 AEFI records...")
        print(f"    Vaccines: {covid_vaccines}")
        print(f"    States: {len(states)} states")

        data = []
        start_date = datetime(2021, 1, 16)  # India COVID vaccination start date

        for i in range(num_records):
            # Realistic date distribution
            days_offset = random.randint(0, 1095)  # 3 years
            vax_date = start_date + timedelta(days=days_offset)

            record = {
                'CASE_ID': f'IND_COVID_AEFI_{i:06d}',
                'VACCINATION_DATE': vax_date.strftime('%Y-%m-%d'),
                'VACCINE': random.choice(covid_vaccines),
                'DOSE_NUMBER': random.choice([1, 2, 3]),  # 1st, 2nd, or booster
                'STATE': random.choice(states),
                'AGE': random.randint(18, 85),
                'GENDER': random.choice(['M', 'F']),
                'SYMPTOM_ONSET_DAYS': random.randint(0, 30),  # Days after vaccination
                'DIAGNOSIS': random.choice(covid_diagnoses),
                'OUTCOME': random.choice(covid_outcomes),
                'HOSPITALIZED': random.choice(['YES', 'NO', 'YES', 'NO']),
                'DAYS_HOSPITALIZED': random.randint(0, 45) if random.random() > 0.7 else 0,
                'CLASSIFICATION': random.choice(classifications),
                'APPROVED_DATE': datetime.now().strftime('%Y-%m-%d'),
                'NOTES': 'Synthetic data for MVP - based on MOHFW patterns'
            }
            data.append(record)

        df = pd.DataFrame(data)
        self.aefi_data = df
        filepath = os.path.join(self.download_dir, "covid_aefi_data.csv")
        df.to_csv(filepath, index=False)

        # Summary stats
        print(f"[✓] Generated {len(df)} records")
        print(f"[✓] Vaccine breakdown:")
        print(df['VACCINE'].value_counts().to_string())
        print(f"[✓] Classification breakdown:")
        print(df['CLASSIFICATION'].value_counts().to_string())
        print(f"[✓] Death rate: {(df['OUTCOME'] == 'DEATH').sum() / len(df) * 100:.3f}%")
        print(f"[✓] Hospitalization rate: {(df['HOSPITALIZED'] == 'YES').sum() / len(df) * 100:.1f}%")

        return filepath

    # ============================================================
    # DATASET 3: Vaccine Batch QR Data (For Vaccination Centers)
    # ============================================================
    def generate_covid_vaccine_batch_qr_data(self, num_batches=150):
        """
        Generates QR code data for COVID vaccine batches

        Each vaccination center gets batch QR codes
        Patients scan after vaccination to:
        1. Verify batch details
        2. Report side effects
        3. See local trust score
        4. Read others' experiences
        """
        print("\n" + "="*60)
        print("[DATASET 3] Generating COVID Vaccine Batch QR Data")
        print("="*60)

        import random

        covid_vaccines = ['COVISHIELD', 'COVAXIN', 'SPUTNIK_V']
        states = [
            'AndhraPradesh', 'Karnataka', 'Maharashtra', 'Delhi', 'UttarPradesh',
            'WestBengal', 'Gujarat', 'Rajasthan', 'Punjab', 'Odisha'
        ]

        print(f"[→] Creating {num_batches} batch records...")

        batch_data = []

        for i in range(num_batches):
            vaccine = covid_vaccines[i % len(covid_vaccines)]
            state = states[i % len(states)]

            # Create meaningful batch ID
            batch_num = str(i).zfill(5)
            batch_date = (datetime(2021, 1, 1) + timedelta(days=random.randint(0, 1095))).strftime('%Y%m')
            batch_id = f"{vaccine[:3]}_{state[:2].upper()}_{batch_date}_{batch_num}"

            qr_record = {
                "batch_id": batch_id,
                "vaccine_name": vaccine,
                "vaccine_full_name": self._get_vaccine_full_name(vaccine),
                "manufacturer": self._get_manufacturer(vaccine),
                "batch_number": batch_num,
                "state": state,
                "allocated_to": f"{state} State Immunization Officer",
                "manufacturing_date": (datetime(2020, 6, 1) + timedelta(days=random.randint(0, 600))).strftime('%Y-%m-%d'),
                "expiry_date": (datetime(2021, 6, 1) + timedelta(days=random.randint(400, 800))).strftime('%Y-%m-%d'),
                "doses_in_batch": random.choice([100, 500, 1000, 5000, 10000]),
                "received_date": (datetime(2021, 1, 1) + timedelta(days=random.randint(0, 1000))).strftime('%Y-%m-%d'),
                "qr_generation_date": datetime.now().strftime('%Y-%m-%d'),
                "vaccine_cold_chain": "2-8°C",
                "storage_location": f"{state} Cold Storage - Unit {random.randint(1, 5)}"
            }
            batch_data.append(qr_record)
            self.batch_data[batch_id] = qr_record

        filepath = os.path.join(self.download_dir, "covid_batch_qr_data.json")
        with open(filepath, 'w') as f:
            json.dump(batch_data, f, indent=2)

        print(f"[✓] Created {len(batch_data)} batch records")
        print(f"[✓] Vaccine distribution:")
        vaccine_counts = {}
        for b in batch_data:
            v = b['vaccine_name']
            vaccine_counts[v] = vaccine_counts.get(v, 0) + 1
        for v, c in vaccine_counts.items():
            print(f"    - {v}: {c} batches")
        print(f"[✓] State distribution: {len(set(b['state'] for b in batch_data))} states")

        return filepath

    # ============================================================
    # DATASET 4: Calculate Trust Scores (Core VaxTrust Logic)
    # ============================================================
    def calculate_covid_trust_scores(self):
        """
        Calculates vaccine trust scores based on AEFI data

        Formula:
        Trust Score = (1 - (Serious_Causal_AEFIs / Total_Doses)) 
                    × (1 - (Deaths / Total_Doses) × 100)

        Result: 0-1 score where 1.0 = completely safe
        """
        print("\n" + "="*60)
        print("[ANALYSIS] Calculating COVID Vaccine Trust Scores")
        print("="*60)

        if self.aefi_data is None or self.vaccination_data is None:
            print("[✗] Missing data - load datasets first")
            return None

        print("[→] Analyzing AEFI patterns by vaccine...")

        trust_scores = []

        for vaccine in self.aefi_data['VACCINE'].unique():
            vaccine_aefi = self.aefi_data[self.aefi_data['VACCINE'] == vaccine]

            # Count causal cases (A1-A4 are vaccine-caused)
            causal_cases = len(vaccine_aefi[vaccine_aefi['CLASSIFICATION'].isin(['A1', 'A2', 'A3', 'A4'])])
            total_aefi = len(vaccine_aefi)
            deaths = len(vaccine_aefi[vaccine_aefi['OUTCOME'] == 'DEATH'])
            hospitalized = len(vaccine_aefi[vaccine_aefi['HOSPITALIZED'] == 'YES'])

            # Get vaccination numbers
            vax_records = self.vaccination_data[self.vaccination_data['vaccine'].str.contains(vaccine, case=False, na=False)]
            total_doses = vax_records['total_vaccinations'].max() if not vax_records.empty else 50000000

            # Calculate trust score
            death_rate = (deaths / total_doses * 100000) if total_doses > 0 else 0
            causal_rate = (causal_cases / total_doses * 100000) if total_doses > 0 else 0

            trust_score = max(0, min(1, 1 - (causal_cases / total_doses)))

            score_record = {
                'vaccine': vaccine,
                'trust_score': round(trust_score, 4),
                'severity': self._get_severity_rating(trust_score),
                'total_aefi_reported': total_aefi,
                'causal_cases': causal_cases,
                'deaths': deaths,
                'hospitalized': hospitalized,
                'total_doses_administered': total_doses,
                'death_rate_per_100k': round(death_rate, 2),
                'causal_rate_per_100k': round(causal_rate, 2),
                'safe_percentage': round((1 - (causal_cases / total_doses)) * 100, 2) if total_doses > 0 else 99.9
            }
            trust_scores.append(score_record)

        trust_df = pd.DataFrame(trust_scores)
        filepath = os.path.join(self.download_dir, "covid_trust_scores.csv")
        trust_df.to_csv(filepath, index=False)

        print("[✓] Trust Scores Calculated:")
        print(trust_df.to_string(index=False))

        return filepath

    # ============================================================
    # DATASET 5: State-wise Trust Score (For Public Dashboard)
    # ============================================================
    def calculate_state_wise_trust_scores(self):
        """
        Breaks down trust scores by state
        Useful for: Geographic hotspot detection, local dashboards
        """
        print("\n" + "="*60)
        print("[ANALYSIS] State-wise Trust Score Analysis")
        print("="*60)

        if self.aefi_data is None:
            print("[✗] AEFI data not loaded")
            return None

        state_scores = []

        for state in self.aefi_data['STATE'].unique():
            state_aefi = self.aefi_data[self.aefi_data['STATE'] == state]

            causal = len(state_aefi[state_aefi['CLASSIFICATION'].isin(['A1', 'A2', 'A3', 'A4'])])
            total = len(state_aefi)
            deaths = len(state_aefi[state_aefi['OUTCOME'] == 'DEATH'])
            hospitalized = len(state_aefi[state_aefi['HOSPITALIZED'] == 'YES'])

            # Estimate doses per state (rough calculation)
            doses_per_state = 50000000 / 15  # Simplified

            trust = 1 - (causal / doses_per_state) if doses_per_state > 0 else 0.99
            trust = max(0, min(1, trust))

            state_scores.append({
                'state': state,
                'trust_score': round(trust, 4),
                'total_aefi': total,
                'causal_cases': causal,
                'deaths': deaths,
                'hospitalized_count': hospitalized,
                'hospitals_affected': len(state_aefi['STATE'].unique())
            })

        state_df = pd.DataFrame(state_scores).sort_values('trust_score', ascending=False)
        filepath = os.path.join(self.download_dir, "state_wise_trust_scores.csv")
        state_df.to_csv(filepath, index=False)

        print("[✓] State Trust Scores (Top 10):")
        print(state_df.head(10).to_string(index=False))

        return filepath

    # ============================================================
    # Helper Methods
    # ============================================================
    def _get_vaccine_full_name(self, vaccine_code):
        mapping = {
            'COVISHIELD': 'Oxford/AstraZeneca (Manufactured by SII)',
            'COVAXIN': 'Inactivated COVID-19 vaccine (Bharat Biotech)',
            'SPUTNIK_V': 'Sputnik V (Russian vaccine)'
        }
        return mapping.get(vaccine_code, vaccine_code)

    def _get_manufacturer(self, vaccine_code):
        mapping = {
            'COVISHIELD': 'Serum Institute of India',
            'COVAXIN': 'Bharat Biotech International Limited',
            'SPUTNIK_V': 'Gamaleya Institute'
        }
        return mapping.get(vaccine_code, 'Unknown')

    def _get_severity_rating(self, trust_score):
        if trust_score >= 0.99:
            return "VERY_SAFE"
        elif trust_score >= 0.98:
            return "SAFE"
        elif trust_score >= 0.95:
            return "GENERALLY_SAFE"
        else:
            return "REQUIRES_REVIEW"

    # ============================================================
    # Final Report Generation
    # ============================================================
    def generate_mvp_report(self):
        """
        Generates a complete MVP readiness report
        """
        print("\n" + "="*60)
        print("[REPORT] VaxTrust COVID-19 MVP - Data Summary")
        print("="*60)

        report = f"""
╔═══════════════════════════════════════════════════════════════╗
║          VaxTrust COVID-19 Vaccine Transparency MVP          ║
║                  India Focus - Data Summary                  ║
║              Generated: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}               ║
╚═══════════════════════════════════════════════════════════════╝

📊 DATASET SUMMARY
─────────────────────────────────────────────────────────────────
✓ COVID Vaccination Data (OWID)
  - Coverage: India, Jan 2021 - Present
  - Total Doses: 2.2 billion+
  - Daily Updates: Automatic

✓ AEFI Adverse Events (Synthetic - MOHFW Pattern)
  - Records: 600 causality-assessed events
  - Vaccines: Covishield, Covaxin, Sputnik V
  - States: 15 major states
  - Based on real MOHFW data structure

✓ Vaccine Batch QR Codes
  - Total Batches: 150 batch records
  - Distribution: 3 vaccines × 50 states
  - Use: Scan at vaccination centers

✓ Trust Scores
  - Vaccine-wise: Overall safety rating per vaccine
  - State-wise: Geographic trust variations
  - Calculation: Based on AEFI causality assessment

🎯 CORE MVP FEATURES
─────────────────────────────────────────────────────────────────
1. QR Scanning
   - Vaccination center prints batch QR code
   - Patient scans after vaccination
   - Shows: Batch details + trust score

2. Side Effect Reporting
   - Anonymous feedback form
   - What symptoms, recovery time
   - Helps future patients decide

3. Local Trust Score
   - Public: Anyone can see
   - Real: Based on verified AEFI data
   - Transparent: No hidden algorithms

4. Public Dashboard
   - Health authorities monitor
   - AI flags unusual patterns
   - Early intervention capability

📁 FILES READY FOR INTEGRATION
─────────────────────────────────────────────────────────────────
✓ covid_vaccination_india.csv         (Real vaccination data)
✓ covid_aefi_data.csv                (Adverse event records)
✓ covid_batch_qr_data.json           (Batch information)
✓ covid_trust_scores.csv             (Vaccine trust scores)
✓ state_wise_trust_scores.csv        (Geographic breakdown)

⚙️  IMPLEMENTATION CHECKLIST
─────────────────────────────────────────────────────────────────
Frontend (React):
  [ ] QR Scanner page (use qr-scanner library)
  [ ] Batch detail display
  [ ] Trust score visualization
  [ ] Side effect form
  [ ] State-wise dashboard

Backend (Python/Flask):
  [ ] /api/batch/<batch_id> - Get batch trust score
  [ ] /api/report-symptom - Accept anonymous reports
  [ ] /api/trust-scores - Get all vaccine scores
  [ ] /api/state/<state>/trust - State-wise data

Database/Storage:
  [ ] Store batch data (JSON or PostgreSQL)
  [ ] Store anonymous reports (encrypted)
  [ ] Cache trust scores (update daily)
  [ ] Audit logs (for transparency)

AI/Anomaly Detection:
  [ ] Train on AEFI data (included)
  [ ] Flag unusual symptom clusters
  [ ] Detect geographic hotspots
  [ ] Alert health authorities

📈 SUCCESS METRICS FOR DEMO
─────────────────────────────────────────────────────────────────
✓ Users can scan COVID vaccine QR codes
✓ System shows real batch + trust score data
✓ Users can report side effects anonymously
✓ Dashboard shows local trust trends
✓ AI identifies unusual patterns (if exists)
✓ Data is transparent and verifiable

🎬 DEMO SCENARIO (36-48 hours)
─────────────────────────────────────────────────────────────────
1. Mock vaccination center with printed QR codes
2. Patient scans QR after taking COVID vaccine
3. App shows:
   - Batch details (vaccine, manufacturer, batch #)
   - Trust score: 0.9956 (Very Safe)
   - "2,134 people took this batch"
   - "238 reported side effects"
   - "Most common: Fever (2%), Myalgia (1.8%)"
   - "98.2% recovered within 2 days"
4. Patient fills: "Felt fever 6hrs, gone by next day"
5. Dashboard updates (shows new feedback)
6. Health officer sees: "Batch COVISLD_TN_202301 - All normal"

🏆 COMPETITIVE ADVANTAGE
─────────────────────────────────────────────────────────────────
✓ Transparency: Real data, real stories
✓ Simplicity: No accounts needed, just scan
✓ Trust: Community-verified, not government-mandated
✓ Speed: Real-time feedback (not quarterly reports)
✓ Privacy: Anonymous reporting
✓ Scalability: Works across all states

🚀 NEXT STEPS (Post-MVP)
─────────────────────────────────────────────────────────────────
1. Pilot with 5-10 vaccination centers
2. Collect 1,000+ real user reports
3. Integrate live CoWIN API
4. Connect SAFE-VAC surveillance
5. Expand to other vaccines (Polio, Rotavirus, etc.)
6. Multi-language support (Hindi, Regional)
7. Mobile app for iOS/Android

═══════════════════════════════════════════════════════════════════
Ready to Build! 🚀 All data in: ./vaxtrust_covid_data/
═══════════════════════════════════════════════════════════════════
"""

        print(report)

        # Save report
        report_path = os.path.join(self.download_dir, "MVP_REPORT.txt")
        with open(report_path, 'w') as f:
            f.write(report)

        return report_path


def main():
    """Main execution - Complete VaxTrust COVID-19 MVP setup"""

    print("""
╔═══════════════════════════════════════════════════════════════╗
║   VaxTrust - COVID-19 Vaccine Transparency MVP Setup         ║
║           India Focus | Hackathon Edition                    ║
╚═══════════════════════════════════════════════════════════════╝
    """)

    # Initialize
    manager = VaxTrustCOVIDDataManager()

    # Step 1: Download real vaccination data
    print("\n[STEP 1/5] Loading Real COVID-19 Vaccination Data")
    print("─" * 60)
    vax_path = manager.download_covid_vaccination_data()

    # Step 2: Generate AEFI data
    print("\n[STEP 2/5] Generating COVID-19 AEFI Data")
    print("─" * 60)
    aefi_path = manager.generate_covid_aefi_data(num_records=600)

    # Step 3: Generate batch QR data
    print("\n[STEP 3/5] Creating Vaccine Batch QR Data")
    print("─" * 60)
    batch_path = manager.generate_covid_vaccine_batch_qr_data(num_batches=150)

    # Step 4: Calculate trust scores
    print("\n[STEP 4/5] Calculating Trust Scores")
    print("─" * 60)
    trust_path = manager.calculate_covid_trust_scores()
    state_trust_path = manager.calculate_state_wise_trust_scores()

    # Step 5: Generate report
    print("\n[STEP 5/5] Generating MVP Report")
    print("─" * 60)
    report_path = manager.generate_mvp_report()

    print(f"""
╔═══════════════════════════════════════════════════════════════╗
║              ✓ Setup Complete! Ready to Build                ║
║                                                               ║
║  All data files in: ./vaxtrust_covid_data/                   ║
║  Start with: Python/React integration                        ║
║  Demo time: Let patients scan COVID vaccine QR codes! 🎯     ║
╚═══════════════════════════════════════════════════════════════╝
    """)


if __name__ == "__main__":
    main()