app.py

import gradio as gr
import torch
import itertools
import pandas as pd
import spaces
import random
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, AutoModel
from sklearn.metrics import pairwise_distances
from collections import Counter
from itertools import chain
from nltk.translate.bleu_score import sentence_bleu, SmoothingFunction
import math
import markdown
from src.text import doctree_from_url, get_selectors_for_class, split_by_heading, DocTree
from src.optimization import ngrams, count_ngrams, self_bleu, dist_n, perplexity, js_divergence


model_name = 'philipp-zettl/t5-small-long-qa'
qa_model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
model_name = 'philipp-zettl/t5-small-qg'
qg_model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained('google/flan-t5-small')

embedding_model = AutoModel.from_pretrained('sentence-transformers/paraphrase-MiniLM-L6-v2')
embedding_tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/paraphrase-MiniLM-L6-v2')

# Move only the student model to GPU if available
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
qa_model = qa_model.to(device)
qg_model = qg_model.to(device)
embedding_model = embedding_model.to(device)

max_questions = 1
max_answers = 1
max_elem_value = 100


def embedding_similarity(inputs, outputs):
    global embedding_model, embedding_tokenizer, device
    def embed(texts):
        inputs = embedding_tokenizer(texts, return_tensors='pt', padding=True, truncation=True).to(device)
        with torch.no_grad():
            outputs = embedding_model(**inputs)
        return outputs.last_hidden_state.mean(dim=1).cpu().numpy()

    input_embeddings = embed(inputs)
    output_embeddings = embed(outputs)

    similarities = pairwise_distances(input_embeddings, output_embeddings, metric='cosine')
    return sum(similarities) / len(similarities)


def evaluate_model(num_beams, num_beam_groups, model, tokenizer, eval_data, max_length=85):
    generated_outputs = []

    for input_text in eval_data:
        input_ids = tokenizer.encode(input_text, return_tensors="pt").to(device)
        outputs = model.generate(
            input_ids, 
            num_beams=num_beams, 
            num_beam_groups=num_beam_groups, 
            diversity_penalty=1.0,
            max_new_tokens=max_length,
        )
        decoded_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
        generated_outputs.append(decoded_text.split())

    # Self-BLEU for diversity
    diversity_score = self_bleu(generated_outputs)

    # Dist-1 and Dist-2 for diversity
    dist1 = dist_n(generated_outputs, 1)
    dist2 = dist_n(generated_outputs, 2)

    # Perplexity for fluency and relevance
    fluency_score = perplexity(model, tokenizer, [" ".join(output) for output in generated_outputs])

    # Embedding similarity for contextual relevance
    contextual_score = embedding_similarity(eval_data, [" ".join(output) for output in generated_outputs])

    # Jensen-Shannon Divergence for distribution similarity
    generated_ngrams = count_ngrams(list(chain(*generated_outputs)), 4)
    reference_ngrams = count_ngrams(list(chain(*[tokenizer.tokenize(text) for text in eval_data])), 4)
    all_ngrams = set(generated_ngrams.keys()).union(set(reference_ngrams.keys()))
    p = [generated_ngrams[ngram] for ngram in all_ngrams]
    q = [reference_ngrams[ngram] for ngram in all_ngrams]
    jsd_score = js_divergence(p, q)

    return {
        "diversity_score": diversity_score,
        "dist1": dist1,
        "dist2": dist2,
        "fluency_score": fluency_score,
        "contextual_score": contextual_score,
        "jsd_score": jsd_score
}


def find_best_parameters(eval_data, model, tokenizer, max_length=85):

    # Parameter ranges
    parameter_map = {
        2: [2],
        4: [2],
        6: [2], # 6x3 == 4x2
        8: [2], # 8x4 == 6x3 == 4x2
        9: [3],
        10: [2], # 10x5 == 8x4 == 6x3 == 4x2
    }

    # Find the best parameters
    best_score = -float('inf')
    best_params = None

    for num_beams in parameter_map.keys():
        for num_beam_groups in parameter_map[num_beams]:
            if num_beam_groups > num_beams:
                continue  # num_beam_groups should not be greater than num_beams

            scores = evaluate_model(num_beams, num_beam_groups, model, tokenizer, eval_data, max_length=max_length)
            # Combine scores to determine the best parameters
            combined_score = (scores['dist1'] + scores['dist2'] - scores['fluency_score'] + scores['contextual_score'] - scores['jsd_score']).mean()
            print(f"num_beams={num_beams}, num_beam_groups={num_beam_groups}, avg combined score={combined_score}")
            
            if combined_score > best_score:
                best_score = combined_score
                best_params = (num_beams, num_beam_groups)

    print(f"Best parameters: num_beams={best_params[0]}, num_beam_groups={best_params[1]} with combined score={best_score}")
    return best_params


def run_model(inputs, tokenizer, model, num_beams=2, num_beam_groups=2, temperature=0.5, num_return_sequences=1, max_length=85, seed=42069):
    all_outputs = []
    torch.manual_seed(seed)
    for input_text in inputs:
        model_inputs = tokenizer([input_text], max_length=512, padding=True, truncation=True)
        input_ids = torch.tensor(model_inputs['input_ids']).to(device)
        for sample in input_ids:
            sample_outputs = []
            with torch.no_grad():
                sample_output = model.generate(
                    input_ids[:1],
                    max_length=max_length,
                    num_return_sequences=num_return_sequences,
                    low_memory=True,
                    use_cache=True,
                    # Diverse Beam search decoding
                    num_beams=max(2, num_return_sequences),
                    num_beam_groups=max(2, num_return_sequences),
                    diversity_penalty=temperature,

                )
                for i, sample_output in enumerate(sample_output):
                    sample_output = sample_output.unsqueeze(0)
                    sample_output = tokenizer.decode(sample_output[0], skip_special_tokens=True)
                    sample_outputs.append(sample_output)

            all_outputs.append(sample_outputs)
    return all_outputs


@spaces.GPU
def gen(content, temperature_qg=0.5, temperature_qa=0.75, num_return_sequences_qg=1, num_return_sequences_qa=1, max_length=85, seed=42069, optimize_questions=False):
    inputs = [
        f'context: {content}'
    ]
    question = run_model(
        inputs, 
        tokenizer, 
        qg_model, 
        num_beams=num_return_sequences_qg, 
        num_beam_groups=num_return_sequences_qg, 
        temperature=temperature_qg, 
        num_return_sequences=num_return_sequences_qg, 
        max_length=max_length,
        seed=seed
    )

    if optimize_questions:
        q_params = find_best_parameters(
            list(chain.from_iterable(question)), qg_model, tokenizer, max_length=max_length
        )

        question = run_model(
            inputs, 
            tokenizer, 
            qg_model, 
            num_beams=q_params[0],
            num_beam_groups=q_params[1],
            temperature=temperature_qg, 
            num_return_sequences=num_return_sequences_qg, 
            max_length=max_length,
            seed=seed
        )

    inputs = list(chain.from_iterable([
        [f'question: {q} context: {content}' for q in q_set] for q_set in question
    ]))
    answer = run_model(
        inputs,
        tokenizer, 
        qa_model, 
        num_beams=num_return_sequences_qa,
        num_beam_groups=num_return_sequences_qa,
        temperature=temperature_qa,
        num_return_sequences=num_return_sequences_qa,
        max_length=max_length,
        seed=seed
    )

    questions = list(chain.from_iterable(question))
    answers = list(chain.from_iterable(answer))

    results = []
    for idx, ans in enumerate(answers):
        results.append({'question': questions[idx % num_return_sequences_qg], 'answer': ans})
    return results


def variable_outputs(k, max_elems=10):
    global max_elem_value
    k = int(k)
    return [gr.Text(visible=True)] * k + [gr.Text(visible=False)] * (max(max_elems, max_elem_value)- k)


def set_outputs(content, max_elems=10):
    c = eval(content)
    print('received content: ', c)
    return [gr.Text(value=t, visible=True) for t in c] + [gr.Text(visible=False)] * (max(max_elems, 10) - len(c))


def create_file_download(qnas):
    with open('qnas.tsv', 'w') as f:
        for idx, qna in qnas.iterrows():
            f.write(qna['Question'] + '\t' + qna['Answer'])
            if idx < len(qnas) - 1:
                f.write('\n')
    return 'qnas.tsv'


def main():
    with gr.Tab(label='QA Generator'):
        with gr.Tab(label='Explanation'):
            gr.Markdown(
                '''
                # QA Generator
                This tab allows you to generate questions and answers from a given piece of text content.

                ## How to use
                1. Enter the text content you want to generate questions and answers from.
                2. Adjust the diversity penalty for question generation and answer generation.
                3. Set the maximum length of the generated questions and answers.
                4. Choose the number of questions and answers you want to generate.
                5. Click on the "Generate" button.

                The next section will give you insights into the generated questions and answers.

                If you're satisfied with the generated questions and answers, you can download them as a TSV file.
                '''
            )
            with gr.Accordion(label='Optimization', open=False):
                gr.Markdown("""
                For optimization of the question generation we apply the following combined score:

                $$\\text{combined} = \\text{dist1} + \\text{dist2} - \\text{fluency} + \\text{contextual} - \\text{jsd}$$

                Here's a brief explanation of each component:

                1. **dist1 and dist2**: These represent the diversity of the generated outputs. dist1 measures the ratio of unique unigrams to total unigrams, and dist2 measures the ratio of unique bigrams to total bigrams. <u>**Higher values indicate more diverse outputs.**</u>

                2. **fluency**: This is the perplexity of the generated outputs, which measures how well the outputs match the language model's expectations. <u>**Lower values indicate better fluency.**</u>

                3. **contextual**: This measures the similarity between the input and generated outputs using embedding similarity. <u>**Higher values indicate better contextual relevance.**</u>

                4. **jsd**: This is the Jensen-Shannon Divergence between the n-gram distributions of the generated outputs and the reference data. <u>**Lower values indicate greater similarity between distributions.**</u>
                """, latex_delimiters=[{'display': False, 'left': '$$', 'right': '$$'}])
        with gr.Tab(label='Generate QA'):
            with gr.Row(equal_height=True):
                with gr.Group("Content"):
                    content = gr.Textbox(label='Content', lines=15, placeholder='Enter text here', max_lines=10_000)
                with gr.Group("Settings"):
                    temperature_qg = gr.Slider(label='Diversity Penalty QG', value=0.2, minimum=0, maximum=1, step=0.01)
                    temperature_qa = gr.Slider(label='Diversity Penalty QA', value=0.5, minimum=0, maximum=1, step=0.01)
                    max_length = gr.Number(label='Max Length', value=85, minimum=1, step=1, maximum=512)
                    num_return_sequences_qg = gr.Number(label='Number Questions', value=max_questions, minimum=1, step=1, maximum=max(max_questions, max_elem_value))
                    num_return_sequences_qa = gr.Number(label="Number Answers", value=max_answers, minimum=1, step=1, maximum=max(max_questions, max_elem_value))
                    seed = gr.Number(label="seed", value=42069)
                    optimize_questions = gr.Checkbox(label="Optimize questions?", value=False)

            with gr.Row():
                gen_btn = gr.Button("Generate")

            @gr.render(
                inputs=[
                    content, temperature_qg, temperature_qa, num_return_sequences_qg, num_return_sequences_qa,
                    max_length, seed, optimize_questions
                ],
                triggers=[gen_btn.click]
            )
            def render_results(content, temperature_qg, temperature_qa, num_return_sequences_qg, num_return_sequences_qa, max_length, seed, optimize_questions):
                if not content.strip():
                    raise gr.Error('Please enter some content to generate questions and answers.')
                qnas = gen(
                    content, temperature_qg, temperature_qa, num_return_sequences_qg, num_return_sequences_qa,
                    max_length, seed, optimize_questions
                )
                df = gr.Dataframe(
                    value=[u.values() for u in qnas],
                    headers=['Question', 'Answer'],
                    col_count=2,
                    wrap=True
                )
                pd_df = pd.DataFrame([u.values() for u in qnas], columns=['Question', 'Answer'])

                download = gr.DownloadButton(label='Download (without headers)', value=create_file_download(pd_df))

            content.change(lambda x: x.strip(), content)


def new_main():
    with gr.Tab('Content extraction from URL'):
        with gr.Tab(label='Explanation'):
            gr.Markdown(
                '''
                # Content extraction from URL
                This tab allows you to extract content from a URL and chunk it into sections.

                ## How to use
                1. Enter the URL of the webpage you want to extract content from.
                2. Select the element class and class name of the HTML element you want to extract content from.
                3. Click on the "Extract content" button.

                The next section will give you insights into the extracted content.

                This was done to give you the possibility to look at the extracted content, as well as manipulate it further.

                Once you extract the content, you can choose the depth level to chunk the content into sections.
                1. Enter the depth level you want to chunk the content into. **Note: <u>This is based on the HTML structure of the webpage, we're utilizing heading tags for this purpose</u>**
                2. Click on the "Chunk content" button.
                '''
            )
        with gr.Tab(label='Extract content'):
            url = gr.Textbox(label='URL', placeholder='Enter URL here', lines=1, max_lines=1)
            elem_class = gr.Dropdown(label='CSS element class', choices=['div', 'p', 'span'], value='div')
            class_name = gr.Dropdown(label='CSS class name', choices=[], allow_custom_value=True)

            extract_btn = gr.Button('Extract content')

            with gr.Group():
                content_state = gr.State(None)
                final_content = gr.Textbox(value='', show_copy_button=True, label='Final content', interactive=True)
                with gr.Accordion('Reveal original input', open=False):
                    og_content = gr.Textbox(value='', label='OG HTML content')

            with gr.Group(visible=False) as step_2_group:
                depth_level = gr.Number(label='Depth level', value=1, minimum=0, step=1, maximum=6)
                continue_btn = gr.Button('Chunk content')

            def render_results(url, elem_class_, class_name_):
                if not url.strip():
                    raise gr.Error('Please enter a URL to extract content.')
                content = doctree_from_url(url, elem_class_, class_name_)
                return [
                    content,
                    content.content,
                    content.as_markdown(content.merge_sections(content.get_sections(0))),
                    gr.Group(visible=True)
                ]

            def get_class_options(url, elem_class):
                if not url.strip():
                    raise gr.Error('Please enter a URL to extract content.')

                return gr.Dropdown(label='CSS class name', choices=list(set(get_selectors_for_class(url, elem_class))))

            def update_content_state_on_final_change(final_content):
                html_content = markdown.markdown(final_content)
                return DocTree(split_by_heading(html_content, 1))

            @gr.render(inputs=[content_state, depth_level], triggers=[continue_btn.click])
            def select_content(content, depth_level):
                if not content:
                    raise gr.Error('Please extract content first.')

                sections = content.get_sections_by_depth(depth_level)
                print(f'Found {len(sections)} sections')
                ds = []
                for idx, section in enumerate(sections):
                    ds.append([idx, content.as_markdown(content.merge_sections(section))])
                gr.Dataframe(value=ds, headers=['Section #', 'Content'], interactive=True, wrap=True)

            url.change(
                get_class_options,
                inputs=[url, elem_class],
                outputs=[class_name]
            )

            extract_btn.click(
                render_results,
                inputs=[
                    url, elem_class, class_name,
                ],
                outputs=[
                    content_state, og_content, final_content, step_2_group
                ]
            )
            final_content.change(update_content_state_on_final_change, inputs=[final_content], outputs=[content_state])


with gr.Blocks() as demo:
    gr.Markdown(
        '''
        # QA-Generator
        A tool to build FAQs or QnAs from a given piece of text content.

        ## How to use
        We provide you two major functionalities:
        1. **Content extraction from URL**: Extract content from a URL and chunk it into sections.
        2. **QA Generator**: Generate questions and answers from a given text content.

        Select the tab you want to use and follow the instructions.
        '''
    )
    new_main()
    main()


demo.queue()
demo.launch()