app.py

import streamlit as st

st.set_page_config(
    layout="wide",  # Can be "centered" or "wide". In the future also "dashboard", etc.
    initial_sidebar_state="auto",  # Can be "auto", "expanded", "collapsed"
    page_title='ZShot',  # String or None. Strings get appended with "• Streamlit".
    page_icon='./logo_zshot.png',  # String, anything supported by st.image, or None.
)

import os
import sys
import warnings

import spacy
from zshot.linker import LinkerSMXM, LinkerTARS, LinkerRegen
from zshot.utils.data_models import Entity
from zshot.mentions_extractor import MentionsExtractorSpacy
from zshot.mentions_extractor.utils import ExtractorType
from zshot import PipelineConfig, displacy

sys.path.append(os.path.abspath('./'))
import streamlit_apps_config as config

warnings.simplefilter('ignore')

## Marking down NER Style
st.markdown(config.STYLE_CONFIG, unsafe_allow_html=True)

########## To Remove the Main Menu Hamburger ########

hide_menu_style = """
        <style>
        #MainMenu {visibility: hidden;}
        </style>
        """
st.markdown(hide_menu_style, unsafe_allow_html=True)

########## Side Bar ########

## loading logo(newer version with href)
import base64


@st.cache_data()
def get_base64_of_bin_file(bin_file):
    with open(bin_file, 'rb') as f:
        data = f.read()
    return base64.b64encode(data).decode()


@st.cache_data()
def get_img_with_href(local_img_path, target_url, size='big'):
    img_format = os.path.splitext(local_img_path)[-1].replace('.', '')
    bin_str = get_base64_of_bin_file(local_img_path)
    height = '90%' if size == 'big' else '45%'
    width = '90%' if size == 'big' else '45%'
    html_code = f'''
        <a href="{target_url}" style='text-align: center;'>
            <img height="{height}" width="{width}" style='display: block; margin-left: auto; margin-right: auto;' src="data:image/{img_format};base64,{bin_str}" />
        </a>'''
    return html_code


logo_html = get_img_with_href('./logo.png', 'https://www.ibm.com/')
st.sidebar.markdown(logo_html, unsafe_allow_html=True)
logo_html = get_img_with_href('./logo_zshot.png', 'https://github.com/IBM/zshot', size='small')
st.sidebar.markdown(logo_html, unsafe_allow_html=True)

# sidebar info
linkers = ["REGEN", "SMXM", "TARS"]
st.sidebar.title("Linker to test")
selected_model = st.sidebar.selectbox("", linkers)

######## Main Page #########

if selected_model == "REGEN":
    app_title = "REGEN Linker"
    app_description = "REGEN is a T5 implementation of GENRE. It performs retrieval generating the unique entity name conditioned on the input text using constrained beam search to only generate valid identifiers."
    st.title(app_title)
    st.markdown("<h2>" + app_description + "</h2>", unsafe_allow_html=True)

elif selected_model == "SMXM":
    app_title = "SMXM Linker"
    app_description = "SMXM model uses the description of the entities to give the model information about the entities."
    st.title(app_title)
    st.markdown("<h2>" + app_description + "</h2>", unsafe_allow_html=True)

elif selected_model == "TARS":
    app_title = "TARS Linker"
    app_description = "TARS doesn't need the descriptions of the entities, so if you can't provide the descriptions of the entities maybe this is the approach you're looking for."
    st.title(app_title)
    st.markdown("<h2>" + app_description + "</h2>", unsafe_allow_html=True)

st.subheader("")

if 'entities' not in st.session_state:
    st.session_state['entities'] = [
        Entity(name="company", description="The name of a company"),
        Entity(name="location", description="A physical location"),
        Entity(name="chemical compound", description="Any substance composed of identical molecules consisting of atoms of two or more chemical elements.")
    ]

def add_ent():
    st.session_state['entities'].append(Entity(name=st.session_state["name"], description=st.session_state["description"]))
    st.session_state['name'] = ""
    st.session_state['description'] = ''
    st.write(st.session_state["name"])
    st.write(st.session_state["description"])

for i, entity in enumerate(st.session_state['entities']):
    col1, col2, col3 = st.columns([2, 5, 1])
    with col1:
        st.text(entity.name)
    with col2:
        st.text(entity.description)
    with col3:
        b = st.button('Remove', key=f"ent_{i}")
        if b:
            st.session_state['entities'].pop(i)
            st.experimental_rerun()  # This causes the app to rerun

with st.form(key="form"):
    col1, col2, col3 = st.columns([2, 5, 1])
    with col1:
        st.text_input("Entity Name", key="name")
    with col2:
        st.text_input("Entity Description", key="description")
    with col3:
        st.form_submit_button('Add', on_click=add_ent)

st.markdown("________")
text = st.text_input("Type here your text and press enter to run:",
                     value="CH2O2 is a chemical compound similar to Acetamide used in International Business "
                           "Machines Corporation (IBM) to create new materials that act like PAGs.")

def build_pipeline(model_name=selected_model):
    nlp = spacy.blank('en')
    mentions_extractor = None

    if model_name == "REGEN":
        linker = LinkerRegen()
        nlp = spacy.load('en_core_web_sm')
        mentions_extractor = MentionsExtractorSpacy(ExtractorType.NER)
    elif model_name == "TARS":
        linker = LinkerTARS()
    elif model_name == "SMXM":
        linker = LinkerSMXM()

    config = PipelineConfig(
        entities=st.session_state['entities'],
        mentions_extractor=mentions_extractor,
        linker=linker
    )
    nlp.add_pipe("zshot", config=config, last=True)

    return nlp

predict = st.button("Run ZShot")
if predict:
    # placeholder for warning
    placeholder = st.empty()
    placeholder.info("Processing...")

    nlp = build_pipeline()
    doc = nlp(text)
    placeholder.empty()

    ent_html = displacy.render(doc, style="ent", jupyter=False)  # Display the entity visualization in the browser:
    st.markdown(ent_html, unsafe_allow_html=True)

st.sidebar.info("""See more: 
- Check ZShot Github [here](https://github.com/IBM/zshot)
- Check ZShot documentation [here](https://ibm.github.io/zshot/)""")