#Code 1
kendra = boto3.client("kendra", region_name='us-east-1')

def get_kendra(query):
    
    response = kendra.query(
            QueryText = query,
            IndexId = kendra_index_id)      

    answer = ""
    for query_result in response["ResultItems"]:

        
        if query_result["Type"]=="QUESTION_ANSWER":
            answer_text = query_result["DocumentExcerpt"]["Text"]
            return answer_text

        elif query_result["Type"]=="ANSWER":
            answer_text = query_result["DocumentExcerpt"]["Text"]
            content = ""
            for line in answer_text.split("\n"):
                line = line.strip()
                if line != "":
                    content += " " + line
            answer = content
    return answer

get_kendra("Which drugs are used to treat COVID?")
get_kendra

#output: Remdesivir, Paxlovid (nirmatrelvir and ritonavir)

#Code 2
url = input("Your Doctor.ai backend url: ")
neo4j_username = input("Your Doctor.ai username: ")
neo4j_password = input("Your Doctor.ai password: ")
driver = GraphDatabase.driver(url, auth=(neo4j_username, neo4j_password))

training_text = """
#How many times did patient id_1 visit the ICU?
MATCH (p:Patient)-[:HAS_STAY]->(v:PatientUnitStay) WHERE p.patient_id =~ '(?i)id_1' RETURN COUNT(v)

#When did patient id_1 visit the ICU?
MATCH (p:Patient)-[:HAS_STAY]->(v:PatientUnitStay) WHERE p.patient_id =~ '(?i)id_1' RETURN v.hospitaldischargeyear

#What was the diagnosis of patient id_1's visit?; Why did patient id_1 visit the ICU?; What was the cause for patient id_1's visit?
MATCH (p:Patient)-[:HAS_STAY]->()-[:HAS_DIAG]->()-[:IS_DISEASE]->(d:Disease) WHERE p.patient_id =~ '(?i)id_1' RETURN d.name

[some codes are omitted here]

#"""

openai.api_key = GPT_APIKEY

def get_GPT_3 (query, temp):
  response = openai.Completion.create(
    engine="text-davinci-003",
    prompt=query,
    temperature=temp,
    max_tokens=256,
    top_p=1,
    frequency_penalty=0,
    presence_penalty=0
  )

  return response.choices[0].text.lstrip()

def get_cypher(query):
    prompt = training_text + query

    cypher = get_GPT_3(prompt, 0)

    return cypher

def get_doctorai(query):
    cypher = get_cypher(query)

    with driver.session() as session:
        result = session.run(cypher)
        return ", ".join(result.value())

answer_doctorai = get_doctorai("Which drugs are used to treat COVID-19?")
answer_doctorai

#Etesevimab, Nirmatrelvir and ritonavir, Sotrovimab, Elasomeran, 
#Tozinameran, Molnupiravir, COVID-19 vaccine, Bamlanivimab, 
#Remdesivir, Tixagevimab and cilgavimab

#Code 3
answer_gpt_3 = get_GPT_3("Which drugs are used to treat COVID?", 0.8)
answer_gpt_3

#The FDA has approved two drugs to treat COVID-19: 
#remdesivir and dexamethasone. Remdesivir is an antiviral 
#medication that works by blocking the virus from replicating 
#in the body. Dexamethasone is a steroid that reduces 
#inflammation and helps with breathing difficulties. Other 
#drugs, such as hydroxychloroquine and convalescent plasma, 
#are also being studied for their potential to be used to treat COVID-19.

#Code 4
from typing import List
import ast

def get_consensus(statements: List[str], weights: List[int], my_query):
    statements_str = "\n".join([f"{i} | {y}" for i, y in enumerate(statements)])
    #print (statements_str)

    ensemble_prompt = f"""
You have three answers to the same question.

If the answer contains multiple items, write the statement id and convert the items into a list like this
Statements:
0 | They used A, B.
1 | B was their tools.
2 | They chose C plus A.

Question:
What do they use?

Answers:
The answers are items
0 | ["A", "B"]
1 | ["B"]
2 | ["A", "C"]

If they are statements without a list of items, unify them into a coherent statement like this
Statements:
0 | This medicine XYZ is used to treat lung cancer.
1 | XYZ is developed by the company ABC.
2 | The drug XYZ is approved by the FDA.

Question:
Explain the drug XYZ.

Answers:
The answer is a statement
The company ABC has developed the drug XYZ to treat lung cancer. XYZ has reveiced the FDA approval.


Statements:
{statements_str}

Question:
{my_query}

Answers:
"""
    
    result = get_GPT_3 (ensemble_prompt, 0)
    #print (ensemble_prompt)
    #print (result)
   
    container = {}
    is_list = True
    for line in result.split("\n"):
        line = line.strip()
        if len(line) > 0:
            if line == "The answers are items":
                is_list = True
                continue
            elif line == "The answer is a statement":
                is_list = False
                continue
            
            if is_list == True:
                if "|" in line:
                    fields = line.split("|")

                    id = int(fields[0].strip())
                    items = ast.literal_eval(fields[1].strip())
                    
                    for item in items:
                        if item not in container:
                            container[item] = 0
                        container[item] += weights[id]
            else:
                container[line] = sum(weights)
    
    return container


bots = ["doctor.ai", "kendra", "GPT-3"]
statements = ["remdesivir, nirmatrelvir and ritonavir", "I used hydroxychloroquin to treat COVID", "remdesivir, nirmatrelvir and ritonavir have been successfully used against COVID-19"]
weights = [9, 6, 5]

print (get_consensus(statements, weights, "Which drugs are used to treat COVID?"))
#{
# 'remdesivir': 14, 'nirmatrelvir': 14, 'ritonavir': 14, 
# 'hydroxychloroquin': 6
#}

remdesivir, nirmatrelvir and ritonavir
I used hydroxychloroquin to treat COVID
remdesivir, nirmatrelvir and ritonavir have been successfully used against COVID-19

{'remdesivir': 14, 'nirmatrelvir': 14, 'ritonavir': 14, 'hydroxychloroquin': 6}

This medicine XYZ is used to treat lung cancer.
XYZ is developed by the company ABC.
The drug XYZ is approved by the FDA.

The company ABC has developed the drug XYZ to treat lung cancer. 
XYZ has reveiced the FDA approval.

#Code 5
my_query = "What is small cell lung cancer?"

answer_kendra = get_kendra(my_query)
answer_doctorai = get_doctorai(my_query)
answer_gpt_3 = get_GPT_3(my_query, 0.8)

bots = ["doctor.ai", "kendra", "GPT-3"]
statements = [answer_doctorai, answer_kendra, answer_gpt_3]
weights = [9, 6, 5]

for b, s in zip(bots, statements):
    print (f"{b}: {s}")

print (get_consensus(bots, statements, weights, my_query))

doctor.ai: Lung cancer is a leading cause of cancer death among men and women 
in industrialized countries. Small cell lung carcinoma (SCLC) is a highly 
aggressive neoplasm, which accounts for approximately 25% of all lung cancer 
cases. Molecular mechanisms altered in SCLC include induced expression of 
oncogene, MYC, and loss of tumorsuppressor genes, such as p53, PTEN, RB, and 
FHIT. The overexpression of MYC proteins in SCLC is largely a result of gene 
amplification. Such overexpression leads to more rapid proliferation and loss 
of terminal differentiation. Mutation or deletion of p53 or PTEN can lead to 
more rapid proliferation and reduced apoptosis. The retinoblastoma gene RB1 
encodes a nuclear phosphoprotein that helps to regulate cell-cycle progression. 
The fragile histidine triad gene FHIT encodes the enzyme diadenosine 
triphosphate hydrolase, which is thought to have an indirect role in 
proapoptosis and cell-cycle control.

kendra: Small cell lung cancer is a disease in which malignant (cancer) cells 
form in the tissues of the lung.

GPT-3: Small cell lung cancer (SCLC) is an aggressive form of lung cancer 
that affects a small number of cells in the lung. It is one of two main 
types of lung cancer, the other being non-small cell lung cancer (NSCLC). 
It is much less common than NSCLC and tends to spread quickly, often to 
other parts of the body. Symptoms of SCLC can include coughing, chest pain, 
shortness of breath, and weight loss. Treatment typically involves a 
combination of chemotherapy, radiation therapy, and surgery.

Small cell lung cancer (SCLC) is an aggressive form of lung cancer that 
affects a small number of cells in the lung. It is characterized by the 
overexpression of oncogene MYC, loss of tumor suppressor genes such as p53, 
PTEN, RB, and FHIT, and can lead to more rapid proliferation and reduced 
apoptosis. Symptoms of SCLC can include coughing, chest pain, shortness of 
breath, and weight loss. Treatment typically involves a combination of 
chemotherapy, radiation therapy, and surgery.

doctor.ai: etesevimab, nirmatrelvir and ritonavir, sotrovimab, elasomeran, 
tozinameran, molnupiravir, covid-19 vaccine, bamlanivimab, remdesivir, 
tixagevimab and cilgavimab

kendra: remdesivir, nirmatrelvir and ritonavir

GPT-3: the main drugs used to treat covid-19 are antiviral medications, such 
as remdesivir and monoclonal antibodies. other drugs being studied to treat 
severe covid-19 cases include corticosteroids, antiviral drugs, interferon, 
and anticoagulants.

{
'etesevimab': 9, 'nirmatrelvir and ritonavir': 15, 'sotrovimab': 9, 
'elasomeran': 9, 'tozinameran': 9, 'molnupiravir': 9, 'covid-19 vaccine': 9, 
'bamlanivimab': 9, 'remdesivir': 20, 'tixagevimab': 9, 'cilgavimab': 9, 
'monoclonal antibodies': 5, 'corticosteroids': 5, 'antiviral drugs': 5, 
'interferon': 5, 'anticoagulants': 5
}