import os
import openai

os.environ["TOKENIZERS_PARALLELISM"] = "false"
os.environ["OPENAI_API_KEY"] 
def save_file(input_file):
    import shutil
    import os

    destination_dir = "/home/user/app/file/"
    os.makedirs(destination_dir, exist_ok=True)

    output_dir="/home/user/app/file/"

    for file in input_file:
      shutil.copy(file.name, output_dir)

    return "File(s) saved successfully!"

def process_file():
    from langchain.document_loaders import PyPDFLoader
    from langchain.document_loaders import DirectoryLoader
    from langchain.document_loaders import TextLoader
    from langchain.document_loaders import Docx2txtLoader
    from langchain.vectorstores import FAISS
    from langchain.embeddings.openai import OpenAIEmbeddings
    from langchain.text_splitter import RecursiveCharacterTextSplitter
    import openai

    loader1 = DirectoryLoader('/home/user/app/file/', glob="./*.pdf", loader_cls=PyPDFLoader)
    document1 = loader1.load()

    loader2 = DirectoryLoader('/home/user/app/file/', glob="./*.txt", loader_cls=TextLoader)
    document2 = loader2.load()

    loader3 = DirectoryLoader('/home/user/app/file/', glob="./*.docx", loader_cls=Docx2txtLoader)
    document3 = loader3.load()

    document1.extend(document2)
    document1.extend(document3)

    text_splitter = RecursiveCharacterTextSplitter(
        chunk_size=1000,
        chunk_overlap=200,
        length_function=len)

    docs = text_splitter.split_documents(document1)
    embeddings = OpenAIEmbeddings()

    file_db = FAISS.from_documents(docs, embeddings)
    file_db.save_local("/home/user/app/file_db/")

    return "File(s) processed successfully!"

def formatted_response(docs, response):
    formatted_output = response + "\n\nSources"

    for i, doc in enumerate(docs):
        source_info = doc.metadata.get('source', 'Unknown source')
        page_info = doc.metadata.get('page', None)

        file_name = source_info.split('/')[-1].strip()

        if page_info is not None:
            formatted_output += f"\n{file_name}\tpage no {page_info}"
        else:
            formatted_output += f"\n{file_name}"

    return formatted_output

def search_file(question):
    from langchain.embeddings.openai import OpenAIEmbeddings
    from langchain.vectorstores import FAISS
    from langchain.chains.question_answering import load_qa_chain
    from langchain.callbacks import get_openai_callback
    from langchain.llms import OpenAI
    import openai
    from langchain.chat_models import ChatOpenAI

    embeddings = OpenAIEmbeddings()
    file_db = FAISS.load_local("/home/user/app/file_db/", embeddings)
    docs = file_db.similarity_search(question)

    llm = ChatOpenAI(model_name='gpt-3.5-turbo')
    chain = load_qa_chain(llm, chain_type="stuff")

    with get_openai_callback() as cb:
        response = chain.run(input_documents=docs, question=question)
        print(cb)

    return formatted_response(docs, response)

def local_search(question):
    from langchain.embeddings.openai import OpenAIEmbeddings
    from langchain.vectorstores import FAISS
    from langchain.chains.question_answering import load_qa_chain
    from langchain.callbacks import get_openai_callback
    from langchain.llms import OpenAI
    import openai
    from langchain.chat_models import ChatOpenAI

    embeddings = OpenAIEmbeddings()
    file_db = FAISS.load_local("/home/user/app/local_db/", embeddings)
    docs = file_db.similarity_search(question)

    llm = ChatOpenAI(model_name='gpt-3.5-turbo')
    chain = load_qa_chain(llm, chain_type="stuff")

    with get_openai_callback() as cb:
        response = chain.run(input_documents=docs, question=question)
        print(cb)

    return formatted_response(docs, response)

def delete_file():

    import shutil

    path1 = "/home/user/app/file/"
    path2 = "/home/user/app/file_db/"

    try:
        shutil.rmtree(path1)
        shutil.rmtree(path2)
        return "Deleted Successfully"

    except:
        return "Already Deleted"

def soap_refresh():
    import os
    import gradio as gr

    destination_folder = "/home/user/app/soap_docs/"
    if not os.path.exists(destination_folder):
        os.makedirs(destination_folder)

    directory = '/home/user/app/soap_docs/'
    file_list = []

    for root, dirs, files in os.walk(directory):
        for file in files:
            file_list.append(file)
    return gr.Dropdown.update(choices=file_list)

def sbar_refresh():
    import os
    import gradio as gr

    destination_folder = "/home/user/app/sbar_docs/"
    if not os.path.exists(destination_folder):
        os.makedirs(destination_folder)

    directory = '/home/user/app/sbar_docs/'
    file_list = []

    for root, dirs, files in os.walk(directory):
        for file in files:
            file_list.append(file)
    return gr.Dropdown.update(choices=file_list)

def ask_soap(doc_name, question):
    from langchain.llms import OpenAI
    from langchain import PromptTemplate, LLMChain
    from langchain.chat_models import ChatOpenAI
    import openai
    import docx

    docx_path = "/home/user/app/soap_docs/" + doc_name

    doc = docx.Document(docx_path)
    extracted_text = "Extracted text:\n\n\n"

    for paragraph in doc.paragraphs:
        extracted_text += paragraph.text + "\n"

    question = (
        "\n\nUse the 'Extracted text' to answer the following question:\n" + question
    )
    extracted_text += question

    if extracted_text:
        print(extracted_text)
    else:
        print("failed")

    template = """Question: {question}

    Answer: Let's think step by step."""

    prompt = PromptTemplate(template=template, input_variables=["question"])

    llm = ChatOpenAI(model_name="gpt-3.5-turbo")
    llm_chain = LLMChain(prompt=prompt, llm=llm)
    response = llm_chain.run(extracted_text)

    return response

def ask_sbar(doc_name, question):
    from langchain.llms import OpenAI
    from langchain import PromptTemplate, LLMChain
    from langchain.chat_models import ChatOpenAI
    import openai
    import docx

    docx_path = "/home/user/app/sbar_docs/" + doc_name

    doc = docx.Document(docx_path)
    extracted_text = "Extracted text:\n\n\n"

    for paragraph in doc.paragraphs:
        extracted_text += paragraph.text + "\n"

    question = (
        "\n\nUse the 'Extracted text' to answer the following question:\n" + question
    )
    extracted_text += question

    if extracted_text:
        print(extracted_text)
    else:
        print("failed")

    template = """Question: {question}

    Answer: Let's think step by step."""

    prompt = PromptTemplate(template=template, input_variables=["question"])

    llm = ChatOpenAI(model_name="gpt-3.5-turbo")
    llm_chain = LLMChain(prompt=prompt, llm=llm)
    response = llm_chain.run(extracted_text)

    return response

def search_gpt(question):
    from langchain.llms import OpenAI
    from langchain import PromptTemplate, LLMChain
    from langchain.chat_models import ChatOpenAI

    template = """Question: {question}

    Answer: Let's think step by step."""

    prompt = PromptTemplate(template=template, input_variables=["question"])

    llm = ChatOpenAI(model_name="gpt-3.5-turbo")
    llm_chain = LLMChain(prompt=prompt, llm=llm)
    response = llm_chain.run(question)

    return response

def local_gpt(question):
    from langchain.llms import OpenAI
    from langchain import PromptTemplate, LLMChain
    from langchain.chat_models import ChatOpenAI

    template = """Question: {question}

    Answer: Let's think step by step."""

    prompt = PromptTemplate(template=template, input_variables=["question"])

    llm = ChatOpenAI(model_name="gpt-3.5-turbo")
    llm_chain = LLMChain(prompt=prompt, llm=llm)
    response = llm_chain.run(question)

    return response

global output

def audio_text(filepath):
    import openai
    global output

    audio = open(filepath, "rb")
    transcript = openai.Audio.transcribe("whisper-1", audio)
    output = transcript["text"]

    return output

global soap_response
global sbar_response

def transcript_soap(text):
    from langchain.llms import OpenAI
    from langchain import PromptTemplate, LLMChain
    from langchain.chat_models import ChatOpenAI

    global soap_response

    question = (
        "Use the following context given below to generate a detailed SOAP Report:\n\n"
    )
    question += text
    print(question)

    template = """Question: {question}

    Answer: Let's think step by step."""

    word_count = len(text.split())
    prompt = PromptTemplate(template=template, input_variables=["question"])

    if word_count < 2000:
        llm = ChatOpenAI(model="gpt-3.5-turbo")
    elif word_count < 5000:
        llm = ChatOpenAI(model="gpt-4")
    else:
        llm = ChatOpenAI(model="gpt-4-32k")

    llm_chain = LLMChain(prompt=prompt, llm=llm)
    soap_response = llm_chain.run(question)

    return soap_response

def transcript_sbar(text):
    from langchain.llms import OpenAI
    from langchain import PromptTemplate, LLMChain
    from langchain.chat_models import ChatOpenAI

    global sbar_response

    question = (
        "Use the following context given below to generate a detailed SBAR Report:\n\n"
    )
    question += text
    print(question)

    template = """Question: {question}

    Answer: Let's think step by step."""

    word_count = len(text.split())
    prompt = PromptTemplate(template=template, input_variables=["question"])

    if word_count < 2000:
        llm = ChatOpenAI(model="gpt-3.5-turbo")
    elif word_count < 5000:
        llm = ChatOpenAI(model="gpt-4")
    else:
        llm = ChatOpenAI(model="gpt-4-32k")

    llm_chain = LLMChain(prompt=prompt, llm=llm)
    sbar_response = llm_chain.run(question)

    return sbar_response

def text_soap():
    from langchain.llms import OpenAI
    from langchain import PromptTemplate, LLMChain
    from langchain.chat_models import ChatOpenAI

    global output
    global soap_response
    output = output

    question = (
        "Use the following context given below to generate a detailed SOAP Report:\n\n"
    )
    question += output
    print(question)

    template = """Question: {question}

    Answer: Let's think step by step."""

    word_count = len(output.split())
    prompt = PromptTemplate(template=template, input_variables=["question"])

    if word_count < 2000:
        llm = ChatOpenAI(model="gpt-3.5-turbo")
    elif word_count < 5000:
        llm = ChatOpenAI(model="gpt-4")
    else:
        llm = ChatOpenAI(model="gpt-4-32k")

    llm_chain = LLMChain(prompt=prompt, llm=llm)
    soap_response = llm_chain.run(question)

    return soap_response

def text_sbar():
    from langchain.llms import OpenAI
    from langchain import PromptTemplate, LLMChain
    from langchain.chat_models import ChatOpenAI

    global output
    global sbar_response
    output = output

    question = (
        "Use the following context given below to generate a detailed SBAR Report:\n\n"
    )
    question += output
    print(question)

    template = """Question: {question}

    Answer: Let's think step by step."""

    word_count = len(output.split())
    prompt = PromptTemplate(template=template, input_variables=["question"])

    if word_count < 2000:
        llm = ChatOpenAI(model="gpt-3.5-turbo")
    elif word_count < 5000:
        llm = ChatOpenAI(model="gpt-4")
    else:
        llm = ChatOpenAI(model="gpt-4-32k")

    llm_chain = LLMChain(prompt=prompt, llm=llm)
    sbar_response = llm_chain.run(question)

    return sbar_response

def soap_docx(name):
    global soap_response
    soap_response = soap_response
    import docx
    import os

    destination_folder = "/home/user/app/soap_docs/"
    if not os.path.exists(destination_folder):
        os.makedirs(destination_folder)

    path = f"/home/user/app/soap_docs/SOAP_{name}.docx"

    doc = docx.Document()
    doc.add_paragraph(soap_response)
    doc.save(path)

    return "Successfully saved .docx File"

def sbar_docx(name):
    global sbar_response
    sbar_response = sbar_response
    import docx
    import os

    destination_folder = "/home/user/app/sbar_docs/"
    if not os.path.exists(destination_folder):
        os.makedirs(destination_folder)

    path = f"/home/user/app/sbar_docs/SBAR_{name}.docx"

    doc = docx.Document()
    doc.add_paragraph(sbar_response)
    doc.save(path)

    return "Successfully saved .docx File"

import gradio as gr

css = """
.col{
    max-width: 50%;
    margin: 0 auto;
    display: flex;
    flex-direction: column;
    justify-content: center;
    align-items: center;
}
"""

with gr.Blocks(css=css) as demo:
    gr.Markdown("## <center>Medical App</center>")

    with gr.Tab("SOAP and SBAR Note Creation"):
      with gr.Column(elem_classes="col"):

        with gr.Tab("From Recorded Audio"):
          with gr.Column():

            mic_audio_input = gr.Audio(type="filepath", label="Speak to the Microphone")
            mic_audio_button = gr.Button("Generate Transcript")
            mic_audio_output = gr.Textbox(label="Output")

            mic_text_soap_button = gr.Button("Generate SOAP Report")
            mic_text_soap_output = gr.Textbox(label="Output")
            mic_text_sbar_button = gr.Button("Generate SBAR Report")
            mic_text_sbar_output = gr.Textbox(label="Output")

            mic_docx_input = gr.Textbox(label="Enter the name of .docx File")
            mic_soap_docx_button = gr.Button("Save SOAP .docx File")
            mic_soap_docx_output = gr.Textbox(label="Output")
            mic_sbar_docx_button = gr.Button("Save SBAR .docx File")
            mic_sbar_docx_output = gr.Textbox(label="Output")

        with gr.Tab("From Uploaded Audio"):
          with gr.Column():

            upload_audio_input = gr.Audio(type="filepath", label="Upload Audio File here")
            upload_audio_button = gr.Button("Generate Transcript")
            upload_audio_output = gr.Textbox(label="Output")

            upload_text_soap_button = gr.Button("Generate SOAP Report")
            upload_text_soap_output = gr.Textbox(label="Output")
            upload_text_sbar_button = gr.Button("Generate SBAR Report")
            upload_text_sbar_output = gr.Textbox(label="Output")

            upload_docx_input = gr.Textbox(label="Enter the name of .docx File")
            upload_soap_docx_button = gr.Button("Save SOAP .docx File")
            upload_soap_docx_output = gr.Textbox(label="Output")
            upload_sbar_docx_button = gr.Button("Save SBAR .docx File")
            upload_sbar_docx_output = gr.Textbox(label="Output")

        with gr.Tab("From Text Transcript"):
          with gr.Column():

            text_transcript_input = gr.Textbox(label="Enter your Transcript here")

            text_text_soap_button = gr.Button("Generate SOAP Report")
            text_text_soap_output = gr.Textbox(label="Output")
            text_text_sbar_button = gr.Button("Generate SBAR Report")
            text_text_sbar_output = gr.Textbox(label="Output")

            text_docx_input = gr.Textbox(label="Enter the name of .docx File")
            text_soap_docx_button = gr.Button("Save SOAP .docx File")
            text_soap_docx_output = gr.Textbox(label="Output")
            text_sbar_docx_button = gr.Button("Save SBAR .docx File")
            text_sbar_docx_output = gr.Textbox(label="Output")

    with gr.Tab("SOAP and SBAR Queries"):
      with gr.Column(elem_classes="col"):

        with gr.Tab("Query SOAP Reports"):
          with gr.Column():

            soap_refresh_button = gr.Button("Refresh")
            ask_soap_input = gr.Dropdown(label="Choose File")

            ask_soap_question = gr.Textbox(label="Enter Question here")
            ask_soap_button = gr.Button("Submit")
            ask_soap_output = gr.Textbox(label="Output")

        with gr.Tab("Query SBAR Reports"):
          with gr.Column():

            sbar_refresh_button = gr.Button("Refresh")
            ask_sbar_input = gr.Dropdown(label="Choose File")

            ask_sbar_question = gr.Textbox(label="Enter Question here")
            ask_sbar_button = gr.Button("Submit")
            ask_sbar_output = gr.Textbox(label="Output")

    with gr.Tab("All Queries"):
      with gr.Column(elem_classes="col"):

        local_search_input = gr.Textbox(label="Enter Question here")
        local_search_button = gr.Button("Search")
        local_search_output = gr.Textbox(label="Output")

        local_gpt_button = gr.Button("Ask ChatGPT")
        local_gpt_output = gr.Textbox(label="Output")


    with gr.Tab("Documents Queries"):
      with gr.Column(elem_classes="col"):

        with gr.Tab("Upload and Process Documents"):
          with gr.Column():

            file_upload_input = gr.Files(label="Upload File(s) here")
            file_upload_button = gr.Button("Upload")
            file_upload_output = gr.Textbox(label="Output")

            file_process_button = gr.Button("Process")
            file_process_output = gr.Textbox(label="Output")

        with gr.Tab("Query Documents"):
          with gr.Column():

            file_search_input = gr.Textbox(label="Enter Question here")
            file_search_button = gr.Button("Search")
            file_search_output = gr.Textbox(label="Output")

            search_gpt_button = gr.Button("Ask ChatGPT")
            search_gpt_output = gr.Textbox(label="Output")

            file_delete_button = gr.Button("Delete")
            file_delete_output = gr.Textbox(label="Output")

    ######################################################################################################
    file_upload_button.click(save_file, inputs=file_upload_input, outputs=file_upload_output)
    file_process_button.click(process_file, inputs=None, outputs=file_process_output)

    file_search_button.click(search_file, inputs=file_search_input, outputs=file_search_output)
    search_gpt_button.click(search_gpt, inputs=file_search_input, outputs=search_gpt_output)

    file_delete_button.click(delete_file, inputs=None, outputs=file_delete_output)

    ######################################################################################################
    local_search_button.click(local_search, inputs=local_search_input, outputs=local_search_output)
    local_gpt_button.click(local_gpt, inputs=local_search_input, outputs=local_gpt_output)

    #######################################################################################################
    soap_refresh_button.click(soap_refresh, inputs=None, outputs=ask_soap_input)
    ask_soap_button.click(ask_soap, inputs=[ask_soap_input, ask_soap_question], outputs=ask_soap_output)

    sbar_refresh_button.click(sbar_refresh, inputs=None, outputs=ask_sbar_input)
    ask_sbar_button.click(ask_sbar, inputs=[ask_sbar_input, ask_sbar_question], outputs=ask_sbar_output)

    ####################################################################################################
    mic_audio_button.click(audio_text, inputs=mic_audio_input, outputs=mic_audio_output)

    mic_text_soap_button.click(text_soap, inputs=None, outputs=mic_text_soap_output)
    mic_text_sbar_button.click(text_sbar, inputs=None, outputs=mic_text_sbar_output)

    mic_soap_docx_button.click(soap_docx, inputs=mic_docx_input, outputs=mic_soap_docx_output)
    mic_sbar_docx_button.click(sbar_docx, inputs=mic_docx_input, outputs=mic_sbar_docx_output)
    ####################################################################################################
    upload_audio_button.click(audio_text, inputs=upload_audio_input, outputs=upload_audio_output)

    upload_text_soap_button.click(text_soap, inputs=None, outputs=upload_text_soap_output)
    upload_text_sbar_button.click(text_sbar, inputs=None, outputs=upload_text_sbar_output)

    upload_soap_docx_button.click(soap_docx, inputs=upload_docx_input, outputs=upload_soap_docx_output)
    upload_sbar_docx_button.click(sbar_docx, inputs=upload_docx_input, outputs=upload_sbar_docx_output)
    ###########################################################################################################
    text_text_soap_button.click(transcript_soap, inputs=text_transcript_input, outputs=text_text_soap_output)
    text_text_sbar_button.click(transcript_sbar, inputs=text_transcript_input, outputs=text_text_sbar_output)

    text_soap_docx_button.click(soap_docx, inputs=text_docx_input, outputs=text_soap_docx_output)
    text_sbar_docx_button.click(sbar_docx, inputs=text_docx_input, outputs=text_sbar_docx_output)
    #############################################################################################################

demo.queue()
demo.launch()