policy_analyser/rules.py

"""
    Rules for policy analyser
    @author : Sakshi Tantak
"""

# Imports
import json
from datetime import datetime

from policy_analyser.llm import call_openai

def prepare_payload(extraction):
    payload = {
        'Sum Insured (SI)' : 0,
        'Pre-existing diseases (PED) Waiting period' : 0,
        '30-Day Waiting Period' : False,
        'Specific Illness Waiting Period' : 0,
        'Maternity waiting period' : 0,
        'Exclusions' : [],
        'Maternity benefits' : False,
        'OPD' : 0,
        'Copay' : 0,
        'Deductible' : 0,
        'Daycare treatment' : [],
        'Free Health checkup' : False,
        'Restoration benefit' : False,
        'Sublimits' : [],
        'Room rent limit (proportionate deduction)' : 100,
        'Pre & Post Hospitalization' : False,
        'Domiciliary Cover' : False,
        'No claim bonus' : 0,
        'Ambulance cover' : 0,
        'International coverage' : False,
        'Dental treatment' : 0,
        'AYUSH treatment' : False,
        'Health incentives' : False,
        'Wellness Services' : False,
        'Consumables/ Non medical expenses' : False,
        'Hospital Cash' : False,
        'Adults' : 0,
        'Children' : 0,
        'City' : '',
        'Is Top City' : True,
        'Policy Age' : 0
    }
    num_adults, num_children, is_top_city = 0, 0, True
    today = datetime.today()
    for entity in extraction:
        if entity['entityName'] in ['Exclusions', 'Daycare treatment', 'Sublimits']:
            try:
                value = json.loads(entity['entityValue'])
                payload[entity['entityName']] = value
            except:
                pass

        if entity['entityName'] == "Policy Holder's Details":
            value = entity['entityValue']
            city = ''
            try:
                value = json.loads(value)
                if 'city' in value:
                    city = value['city']
                    try:
                        response = call_openai('Does a given city string belong to set of given cities : [Mumbai, Delhi, Bangalore, Chennai, Hyderabad, Gurgaon, Pune]. Answer in true/false only', city)
                        is_top_city = True if response == 'true' else False
                    except:
                        pass
            except:
                pass
            payload['Is Top City'] = is_top_city
            payload['City'] = city

        if entity['entityName'] == 'Insured Persons details':
            value = entity['entityValue']
            try:
                value = json.loads(value)
                for person in value:
                    if 'date_of_birth' in person:
                        dob = person['date_of_birth']
                        dob = datetime.strptime(dob, '%d/%m/%Y')
                        age = (today - dob).days / 365
                    elif 'age' in person:
                        age = person['age']

                    if age >= 18:
                            num_adults += 1
                    else:
                        num_children += 1
            except:
                num_adults = 1
            payload['Adults'] = num_adults
            payload['Children'] = num_children

        if entity['entityName'] == 'Policy Details':
            try:
                value = json.loads(entity['entityValue'])
                if 'policy_start_date' in value:
                    payload['Policy Age'] = ((today - datetime.strptime(value['policy_start_date'], '%d/%m/%Y')).days / 365) * 12
            except:
                pass

        if entity['entityName'] in ['Sum Insured (SI)', 'Pre-existing diseases (PED) Waiting period', 'Specific Illness Waiting Period',
                                    'Maternity waiting period', 'OPD', 'Copay', 'Deductible', 'No claim bonus', 'Ambulance cover',
                                    'Dental treatment', 'Room rent limit (proportionate deduction)']:
            value = entity['entityValue']
            if isinstance(value, (float, int)):
                payload[entity['entityName']] = value
            else:
                try:
                    value = float(value)
                    payload[entity['entityName']] = value
                except:
                    pass

        if entity['entityName'] in ['30-Day Waiting Period', 'Maternity benefits', 'Free Health checkup',
                                    'Restoration benefit', 'Pre & Post Hospitalization', 'Domiciliary Cover',
                                    'International coverage', 'AYUSH treatment', 'Health incentives', 'Wellness Services',
                                    'Consumables/ Non medical expenses', 'Hospital Cash']:
            value = entity['entityValue']
            if isinstance(value, bool):
                payload[entity['entityName']] = value
            else:
                payload[entity['entityName']] = True if 'true' in value else False

    return payload

def rules(payload):
    analysis = []
    if payload['Adults'] == 1:
        if payload['Is Top City']:
            if payload['Sum Insured (SI)'] >= 2500000:
                verdict, reason = 'Good', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 25L for an adult in {payload["City"]}'
            if payload['Sum Insured (SI)'] >= 1000000 and payload['Sum Insured (SI)'] < 2500000:
                verdict, reason = 'Average', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) < 25L but > 10L for an adult in {payload["City"]}'
            if payload['Sum Insured (SI)'] < 1000000:
                verdict, reason = 'Bad', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) < 10L for an adult in {payload["City"]}'
        else:
            if payload['Sum Insured (SI)'] >= 1000000:
                verdict, reason = 'Good', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 10L for an adult in {payload["City"]}'
            if payload['Sum Insured (SI)'] >= 500000 and payload['Sum Insured (SI)'] < 1000000:
                verdict, reason = 'Average', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 5L but < 10L for an adult in {payload["City"]}'
            if payload['Sum Insured (SI)'] < 500000:
                verdict, reason = 'Bad', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) < 5L for an adult in {payload["City"]}'

    if payload['Adults'] >= 2:
        if payload['Children'] == 0:
            if payload['Is Top City']:
                if payload['Sum Insured (SI)'] >= 5000000:
                    verdict, reason = 'Good', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 50L for {payload["Adults"]} adults in {payload["City"]}'
                if payload['Sum Insured (SI)'] >= 2500000 and payload['Sum Insured (SI)'] < 5000000:
                    verdict, reason = 'Average', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 25L but < 50L for {payload["Adults"]} adults in {payload["City"]}'
                if payload['Sum Insured (SI)'] < 2500000:
                    verdict, reason = 'Bad', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) < 25L for {payload["Adults"]} adults in {payload["City"]}'
            else:
                if payload['Sum Insured (SI)'] >= 2500000:
                    verdict, reason = 'Good', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 25L for {payload["Adults"]} adults in {payload["City"]}'
                if payload['Sum Insured (SI)'] >= 1000000 and payload['Sum Insured (SI)'] < 2500000:
                    verdict, reason = 'Average', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 10L but < 25L for {payload["Adults"]} adults in {payload["City"]}'
                if payload['Sum Insured (SI)'] < 1000000:
                    verdict, reason = 'Bad', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) < 10L for {payload["Adults"]} adults in {payload["City"]}'

        if payload['Children'] >= 1:
            if payload['Children'] > 1 or payload['Is Top City']:
                if payload['Sum Insured (SI)'] >= 10000000:
                    verdict, reason = 'Good', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 1 CR for {payload["Adults"]} adults & {payload["Children"]} children in {payload["City"]}'
                if payload['Sum Insured (SI)'] >= 5000000 and payload['Sum Insured (SI)'] < 10000000:
                    verdict, reason = 'Average', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 50L but < 1 CR for {payload["Adults"]} adults & {payload["Children"]} children in {payload["City"]}'
                if payload['Sum Insured (SI)'] < 5000000:
                    verdict, reason = 'Bad', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) < 50L for {payload["Adults"]} adults & {payload["Children"]} children in {payload["City"]}'
            else:
                if payload['Sum Insured (SI)'] >= 5000000:
                    verdict, reason = 'Good', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 50L for {payload["Adults"]} adults & {payload["Children"]} children in {payload["City"]}'
                if payload['Sum Insured (SI)'] >= 2500000 and payload['Sum Insured (SI)'] < 5000000:
                    verdict, reason = 'Average', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) > 25L but < 50L for {payload["Adults"]} adults & {payload["Children"]} children in {payload["City"]}'
                if payload['Sum Insured (SI)'] < 2500000:
                    verdict, reason = 'Bad', f'Sum Insured (SI) ({payload["Sum Insured (SI)"]}) < 25L for {payload["Adults"]} adults & {payload["Children"]} children in {payload["City"]}'
        analysis.append(
            {
                'factor' : 'Sum Insured (SI)',
                'verdict' : verdict,
                'reason' : reason
            }
        )
   
    if payload['Room rent limit (proportionate deduction)'] > 0:
        verdict, reason = 'Bad', f'There is cap of {payload["Room rent limit (proportionate deduction)"]} on room rent'
    else:
        verdict, reason = 'Good', 'There is no cap on room rent'
    analysis.append({'factor' : 'Room rent limit (proportionate deduction)', 'verdict' : verdict, 'reason' : reason})

    if payload['Deductible'] > 0:
        verdict, reason = 'Bad', f'There is a deductible of {payload["Deductible"]}'
    else:
        verdict, reason = 'Good', 'No deductible'
    analysis.append({'factor' : 'Deductible', 'verdict' : verdict, 'reason' : reason})

    if payload['Sublimits'] == []:
        verdict, reason = 'Good', 'There are no sublimits on any treatments or diseases'
    else:
        verdict = 'Bad'
        sublimits_str = '\n'.join([f'{sublimit["sublimit_name"]}: {sublimit["sublimit_value"]}' for sublimit in payload['Sublimits']])
        reason = f'Following sublimits were found in your policy:\n{sublimits_str}'
    analysis.append({'factor' :'Sublimits', 'verdict' : verdict, 'reason' : reason})

    if payload['Copay'] == 0 and payload['Copay'] <= 5:
        verdict, reason = 'Good', f'Copayment ({payload["Copay"]}) < 5%'
    elif payload['Copay'] > 5 and payload['Copay'] <= 10:
        verdict, reason = 'Average', f'Copayment ({payload["Copay"]}) > 5% but < 10%'
    elif payload['Copay'] > 10:
        verdict, reason = 'Bad', f'Copayment (({payload["Copay"]})) > 10%'
    analysis.append({'factor' : 'Copay', 'verdict' : verdict, 'reason' : reason})

    if payload['Pre-existing diseases (PED) Waiting period'] > 0:
        if payload['Policy Age'] > payload['Pre-existing diseases (PED) Waiting period']:
            verdict, reason = 'Good', f'Your policy has a waiting period of {payload["Pre-existing diseases (PED) Waiting period"]} months on pre-existing diseases but the waiting period has expired as of today'
        else:
            verdict, reason = 'Bad', f'Your policy has a waiting period of {payload["Pre-existing diseases (PED) Waiting period"]} months on pre-existing diseases which is yet to expire'
    else:
        verdict, reason = 'Good', f'Your policy has no waiting period on pre-existing diseases'
    analysis.append({'factor' : 'Pre-existing diseases (PED) Waiting period', 'verdict' : verdict, 'reason' : reason})

    if payload['30-Day Waiting Period']:
        if payload['Policy Age'] > 1:
            verdict, reason = 'Good', f'Your policy has a 30 day waiting period but it has expired as of today'
        else:
            verdict, reason = 'Bad', f'Your policy has a 30 day waiting period which is yet to expire'
    else:
        verdict, reason = 'Good', f'Your policy has no 30 day waiting period'
    analysis.append({'factor' : '30-Day Waiting Period', 'verdict' : verdict, 'reason' : reason})

    if payload['Specific Illness Waiting Period'] > 0:
        if payload['Policy Age'] > payload['Specific Illness Waiting Period']:
            verdict, reason = 'Good', f'Your policy has a waiting period of {payload["Specific Illness Waiting Period"]} on specific illnesses but the waiting period has expired as of today'
        else:
            verdict, reason = 'Bad', f'Your policy has a waiting period of {payload["Specific Illness Waiting Period"]} on specific illnesses which is yet to expire'
    else:
        verdict, reason = 'Good', f'Your policy has no waiting period any on specific illnesses'
    analysis.append({'factor' : 'Specific Illness Waiting Period', 'verdict' : verdict, 'reason' : reason})

    if payload['Maternity benefits']:
        analysis.append(
            {
                'factor' : 'Maternity benefits',
                'verdict' : 'Good',
                'reason' : 'Maternity benefits present, check waiting period'
            }
        )
        if payload['Maternity waiting period'] > 0:
            if payload['Policy Age'] > payload['Maternity waiting period']:
                verdict, reason = 'Good', f'Your policy has a waiting period of {payload["Maternity waiting period"]} for maternity cases but it has expired as of today'
            else:
                verdict, reason = 'Bad', f'Your policy has a waiting period of {payload["Maternity waiting period"]} for maternity cases which is yet to expire'
        else:
            verdict, reason = 'Good', f'Your policy has a no waiting period for maternity cases'
        analysis.append({'factor' : 'Maternity waiting period', 'verdict' : verdict, 'reason' : reason})
    else:
        analysis.append(
            {
                'factor' : 'Maternity benefits',
                'verdict' : 'Bad',
                'reason' : 'No maternity benefits'
            }
        )

    return analysis

if __name__ == '__main__':
    import json
    import glob
    dirpath = '/Users/sakshi.tantak/Downloads/Porting Documents/testing-data/sample/poc'
    for file in glob.glob(f'{dirpath}/*.analysis.json'):
        json_data = json.load(open(file))
        payload = prepare_payload(json_data[1]['response']['processed'])
        json_data.append({
            'stage' : 'POST_PROCESS',
            'response' : payload,
            'time' : 0
        })
        # print(json_data)
        with open(file, 'w') as f:
            json.dump(json_data, f, indent = 4)