app.py

import gradio as gr
from transformers import pipeline
import requests
from bs4 import BeautifulSoup
import pandas as pd

# Initialize models
classification_model = pipeline("text-classification", model="models/text_classification_model", tokenizer="models/text_classification_model", top_k=5)
mask_model = pipeline("fill-mask", model="models/fill_mask_model", tokenizer="models/fill_mask_model", top_k=100)

# Load data
eunis_habitats = pd.read_excel('data/eunis_habitats.xlsx')
    
def return_habitat_image(habitat_label):
    floraveg_url = f"https://floraveg.eu/habitat/overview/{habitat_label}"
    response = requests.get(floraveg_url)
    if response.status_code == 200:
        soup = BeautifulSoup(response.text, 'html.parser')
        img_tag = soup.find('img', src=lambda x: x and x.startswith("https://files.ibot.cas.cz/cevs/images/syntaxa/thumbs/"))
        if img_tag:
            image_url = img_tag['src']
        else:
            image_url = "https://www.salonlfc.com/wp-content/uploads/2018/01/image-not-found-scaled-1150x647.png"
    else:
        image_url = "https://www.salonlfc.com/wp-content/uploads/2018/01/image-not-found-scaled-1150x647.png"
    #image_url = "https://www.commissionoceanindien.org/wp-content/uploads/2018/07/plantnet.jpg"  # While we don't have the rights
    image = gr.Image(value=image_url)
    return image

def return_species_image(species):
    species = species.capitalize()
    floraveg_url = f"https://floraveg.eu/taxon/overview/{species}"
    response = requests.get(floraveg_url)
    if response.status_code == 200:
        soup = BeautifulSoup(response.text, 'html.parser')
        img_tag = soup.find('img', src=lambda x: x and x.startswith("https://files.ibot.cas.cz/cevs/images/taxa/large/"))
        if img_tag:
            image_url = img_tag['src']
        else:
            image_url = "https://www.salonlfc.com/wp-content/uploads/2018/01/image-not-found-scaled-1150x647.png"
    else:
        image_url = "https://www.salonlfc.com/wp-content/uploads/2018/01/image-not-found-scaled-1150x647.png"
    #image_url = "https://www.commissionoceanindien.org/wp-content/uploads/2018/07/plantnet.jpg"  # While we don't have the rights
    image = gr.Image(value=image_url)
    return image

def gbif_normalization(text):
    base = "https://api.gbif.org/v1"
    api = "species"
    function = "match"
    parameter = "name"
    url = f"{base}/{api}/{function}?{parameter}="
    all_species = text.split(',')
    all_species = [species.strip() for species in all_species]
    species_gbif = []
    for species in all_species:
        url = url.replace(url.partition('name')[2], f'={species}')
        r = requests.get(url)
        r = r.json()
        if 'species' in r:
            r = r["species"]
        else:
            r = species
        species_gbif.append(r)
    text = ", ".join(species_gbif)
    text = text.lower()
    return text

def classification(text, k):
    text = gbif_normalization(text)
    result = classification_model(text)
    habitat_labels = [res['label'] for res in result[0][:k]]
    if k == 1:
        text = f"This vegetation plot probably belongs to the habitat {habitat_labels[0]}."
    else:
        text = f"This vegetation plot probably belongs to the habitat {', '.join(habitat_labels[:-1])} or {habitat_labels[-1]}."
    habitat_name = eunis_habitats[eunis_habitats['EUNIS 2020 code'] == habitat_labels[0]]['EUNIS-2021 habitat name'].values[0]
    text += f"\nThe most likely habitat is '{habitat_name}'."
    text += f"\nSee an image of this habitat (i.e., {habitat_labels[0]}) below."
    image_output = return_habitat_image(habitat_labels[0])
    return text, image_output

def masking(text, k):
    text = gbif_normalization(text)
    text_split = text.split(', ')
    
    best_predictions = []
    best_positions = []
    
    for _ in range(k):
        max_score = 0
        best_prediction = None
        best_position = None
        best_sentence = None

        for i in range(len(text_split) + 1):
            masked_text = ', '.join(text_split[:i] + ['[MASK]'] + text_split[i:])
            
            j = 0
            while True:
                prediction = mask_model(masked_text)[j]
                species = prediction['token_str']
                if species in text_split or species in best_predictions:
                    j += 1
                else:
                    break

            score = prediction['score']
            sentence = prediction['sequence']

            if score > max_score:
                max_score = score
                best_prediction = species
                best_position = i
                best_sentence = sentence
        
        best_predictions.append(best_prediction)
        best_positions.append(best_position)
        text_split.insert(best_position, best_prediction)
    if k == 1:
        text = f"The most likely missing species is {best_predictions[0]} (position {best_positions[0]})."
    else:
        text = f"The most likely missing species are {', '.join(best_predictions[:-1])} and {best_predictions[-1]} (positions {', '.join(map(str, best_positions[:-1]))} and {best_positions[-1]})."
    text += f"\nThe completed vegetation plot is '{best_sentence}'."
    text += f"\nSee an image of the most likely species (i.e., {best_predictions[0]}) below."
    image = return_species_image(best_predictions[0])
    return text, image

with gr.Blocks() as demo:

    gr.Markdown("""<h1 style="text-align: center;">Pl@ntBERT</h1>""")
    
    with gr.Tab("Vegetation plot classification"):
        gr.Markdown("""<h3 style="text-align: center;">Classification of vegetation plots!</h3>""")
        with gr.Row():
            with gr.Column():
                species_classification = gr.Textbox(lines=2, label="Species", placeholder="Enter a list of comma-separated binomial names here.")
                k_classification = gr.Slider(1, 5, value=1, step=1, label="Top-k", info="Choose the number of habitats to display.")
            with gr.Column():
                text_classification = gr.Textbox(label="Prediction")
                image_classification = gr.Image()
        button_classification = gr.Button("Classify")
        gr.Markdown("""<h5 style="text-align: center;">An example of input</h5>""")
        gr.Examples([["phragmites australis, lemna minor, typha latifolia", 3]], [species_classification, k_classification], [text_classification, image_classification], classification, True)
        
    with gr.Tab("Missing species finding"):
        gr.Markdown("""<h3 style="text-align: center;">Finding the missing species!</h3>""")
        with gr.Row():
            with gr.Column():
                species_masking = gr.Textbox(lines=2, label="Species", placeholder="Enter a list of comma-separated binomial names here.")
                k_masking = gr.Slider(1, 5, value=1, step=1, label="Top-k", info="Choose the number of missing species to find.")
            with gr.Column():
                text_masking = gr.Textbox(label="Prediction")
                image_masking = gr.Image()
        button_masking = gr.Button("Find")
        gr.Markdown("""<h5 style="text-align: center;">An example of input</h5>""")
        gr.Examples([["calamagrostis arenaria, medicago marina, pancratium maritimum, thinopyrum junceum", 1]], [species_masking, k_masking], [text_masking, image_masking], masking, True)

    button_classification.click(classification, inputs=[species_classification, k_classification], outputs=[text_classification, image_classification])
    button_masking.click(masking, inputs=[species_masking, k_masking], outputs=[text_masking, image_masking])

demo.launch()