from openai import OpenAI
client = OpenAI()

import json

ASSISTANT_ID = os.getenv("NIC_KAL_GPT")

def wait_on_run(run, thread):
    while run.status == "queued" or run.status == "in_progress":
        run = client.beta.threads.runs.retrieve(
            thread_id=thread.id,
            run_id=run.id,
        )
        time.sleep(0.5)
    return run


user_prompt = f"""
write the following:

- a cold email, using known frameworks, and in my style
- 3 key questions to ask on a cold call
- a pesonalised Linkedin connection request message

to:

{contact_data}

Write in:
    - German, if contact is in Germany or Austria
    - French, if contact is in France
    - English, otherwise, or if unsure

Never start emails with "I hope this email finds you well" or "I hope you are doing well", or "Quick question". 
Instead, start with a question or a statement that shows you know something about the person or their company.
Use the informal way to refer to the company's name, as if you were talking to a friend (eg "BMW" not "BMW Group").
When using client references from Kaltura, choose relevant ones that are likely to be known by the recipient, so either in the same industry or well-known Enterprises.

Only return the content of the email, including Subject line, but do not add any extra comments to your answer.

Return as JSON following this format:

{
    "email_subject": "Subject Line",
    "email_body": "Email Body",
    "questions": ["Question 1", "Question 2", "Question 3"],
    "linkedin": "Message",
}

Do not return anything else, as I will parse your response programmatically.
Remove any Markdown formatting from your response.
"""

thread = client.beta.threads.create()

message = client.beta.threads.messages.create(
    thread_id=thread.id,
    role="user",
    content=user_prompt,
)

run = client.beta.threads.runs.create(
    thread_id=thread.id,
    assistant_id=ASSISTANT_ID,
)

run = wait_on_run(run, thread)

messages = client.beta.threads.messages.list(thread_id=thread.id)

message = messages.data[0].content[0].text.value

print("\n\nmessage:")
print()
print(message)

print("\n\n")

# Parse the JSON data
data = json.loads(message)

# Accessing individual fields
email_subject = data["email_subject"]
email_body = data["email_body"]
questions = data["questions"]
linkedin_message = data["linkedin"]

# Printing extracted data
# print(f"\n\nOutbound Suggestions for {x.first} {x.last} at {x.company} located in {x.country}:\n")
print("Email Subject:\n", email_subject)
print("\nEmail Body:\n", email_body)
print("\n\nQuestions for cold call:")
for question in questions:
    print("-", question)
print("\n\nLinkedIn Connect Message:\n", linkedin_message)