import json
import gradio as gr
import os
import langchain_openai
import langchain_core
from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser

def enhance_subject(subject, details):
  prompt = ChatPromptTemplate.from_messages([
      ("system", "Generate a clear and concise subject, then provide additional details using descriptive language. Ensure the response is specific and avoids ambiguity or contradictions. The subject should inspire an engaging photo that tells a story. Remove any unnecessary information and don't add any punctuation at the end of the subject."),
      ("user", "The main subject is {subject} {details}.")
    ])
  output_parser = StrOutputParser()
  model = ChatOpenAI(model="gpt-3.5-turbo")
  chain = ( prompt
    | model
    | output_parser
  )
  result = chain.invoke({"subject": subject, "details": details})
  return result

def load_input_fields(filepath):
    """
    Load the input fields from a JSON file.

    Args:
    - filepath (str): The path to the JSON file containing the input fields.

    Returns:
    - dict: A dictionary containing the input fields.
    """

    with open(filepath, "r") as file:
        input_fields = json.load(file)

    return input_fields

def create_html_string(input_text, highlight_color = "green", container_style = "border: 2px solid black; padding: 2px; font-size: 16px;" ):
    """
    Create a HTML string with specific styles applied to highlighted text within square brackets.

    Args:
    - input_text (str): The input text with portions to be highlighted within square brackets.
    - optional: highlight_color (str): Color for the highlighted text (e.g., "green").
    - optional: container_style (str): Any css for inline styling (e.g,, "border: 2px solid black;")

    Returns:
    - str: A HTML string with the applied styles.
    """

    # Replace the highlighted text with HTML span elements for styling
    highlighted_text = input_text.replace("[", f'<span style="color:{highlight_color}; font-weight: bold;">[').replace("]", "]</span>")

    # Construct the full HTML string with the provided styles
    html_string = f'<p style="{container_style}">{highlighted_text}</p>'

    return html_string

def extract_names(objects):
    return [obj['name'] for obj in objects if 'name' in obj]

def clearInput():
    return ""

def format_to_markdown(objects):
    # Skip None objects
    formatted_list = [
        f"> * **{obj.get('name', 'No Name')}** - {obj.get('description', 'No Description')}"
        for obj in objects if obj is not None and obj["name"] != "None"
    ]
    return '\n'.join(formatted_list)

find_filter_by_name = lambda collection, key: next((filter for filter in collection if filter['name'] == key), None)

def display_info(collection, key):
  markdown_text = format_to_markdown([find_filter_by_name(collection, key)])
  return gr.Markdown(markdown_text, visible=True)

def enhance_pipeline(isFrog, subject, details):
  if isFrog and (subject or details):
    result = enhance_subject(subject, details)
    return [gr.Textbox(visible=False), gr.Button(visible=False), gr.TextArea(visible=True, value=result)]
  elif (subject or details) and not isFrog:
    return [gr.Textbox(visible=True), gr.Button(visible=True), gr.TextArea(visible=False)]
  else:
    return [gr.Textbox(visible=False), gr.Button(visible=False), gr.TextArea(visible=False)]

def authenticate(pwd_input, subject, details):
  if pwd_input == os.environ.get("MAGIC_WORD"):
    result = enhance_subject(subject, details)
    return [gr.TextArea(visible=True, value=result), True, gr.Textbox(visible=False), gr.Button(visible=False)]
  else:
    raise gr.Error("You are not from our pond! Use your own LLM to add some juice to your prompt.")