feat: Init

Dockerfile

@@ -0,0 +1,17 @@
+# read the doc: https://huggingface.co/docs/hub/spaces-sdks-docker
+# you will also find guides on how best to write your Dockerfile
+
+FROM python:3.9
+
+WORKDIR /code
+
+COPY ./requirements.txt /code/requirements.txt
+
+RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
+
+COPY . .
+
+ENV H2O_WAVE_LISTEN=":7860"
+ENV H2O_WAVE_ADDRESS='http://127.0.0.1:7860'
+
+CMD ["wave", "run", "app", "--no-reload"]

README.md

@@ -1,11 +1,54 @@
----
-title: Ner Annotation
-emoji: 🐠
-colorFrom: gray
-colorTo: yellow
-sdk: docker
-pinned: false
-license: apache-2.0
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+<div align='center'>
+
+<h1>WaveTon</h1>
+💯 Wave applications
+
+<br>
+<br>
+
+[![License](https://img.shields.io/badge/license-Apache%202.0-blue.svg?logo=apache)](https://github.com/vopani/waveton/blob/master/LICENSE)
+[![GitHub](https://img.shields.io/github/stars/vopani/waveton?color=yellowgreen&logo=github)](https://img.shields.io/github/stars/vopani/waveton?color=yellowgreen&logo=github)
+[![Twitter](https://img.shields.io/twitter/follow/vopani)](https://twitter.com/vopani)
+
+</div>
+
+## NER Annotation 🖥️
+
+Annotate entities for Named-Entity Recognition tasks.
+
+## Setup ⚙️
+
+1. Check the version of Python, must be Python 3.9+ but recommended to use Python 3.10+ for best experience
+
+```bash
+python3 --version
+```
+
+2. Clone the repository
+
+```bash
+git clone https://github.com/vopani/waveton.git
+```
+
+3. Create a virtual environment
+
+```bash
+cd waveton/apps/data_apps/ner_annotation
+python3 -m venv venv
+source venv/bin/activate
+```
+
+4. Install the packages
+
+```bash
+python3 -m pip install -U pip
+python3 -m pip install -r requirements.txt
+```
+
+5. Run the application
+
+```bash
+wave run app
+```
+
+6. View the application on your local browser: [http://localhost:10101](http://localhost:10101)

app.py

@@ -0,0 +1,299 @@
+import logging
+from random import randint
+
+from h2o_wave import Q, main, app, copy_expando, handle_on, on
+
+import cards
+import constants
+
+# Set up logging
+logging.basicConfig(format='%(levelname)s:\t[%(asctime)s]\t%(message)s', level=logging.INFO)
+
+
+@app('/')
+async def serve(q: Q):
+    """
+    Main entry point. All queries pass through this function.
+    """
+
+    try:
+        # Initialize the app if not already
+        if not q.app.initialized:
+            await initialize_app(q)
+
+        # Initialize the client (browser tab) if not already
+        if not q.client.initialized:
+            await initialize_client(q)
+
+        # Update theme if toggled
+        elif q.args.theme_dark is not None and q.args.theme_dark != q.client.theme_dark:
+            await update_theme(q)
+
+        # Delegate query to query handlers
+        elif await handle_on(q):
+            pass
+
+        # Adding this condition to help in identifying bugs (instead of seeing a blank page in the browser)
+        else:
+            await handle_fallback(q)
+
+    except Exception as error:
+        await show_error(q, error=str(error))
+
+
+async def initialize_app(q: Q):
+    """
+    Initialize the app.
+    """
+
+    logging.info('Initializing app')
+
+    # Set initial argument values
+    q.app.cards = ['ner_entities', 'ner_annotator', 'error']
+
+    q.app.initialized = True
+
+
+async def initialize_client(q: Q):
+    """
+    Initialize the client (browser tab).
+    """
+
+    logging.info('Initializing client')
+
+    # Set initial argument values
+    q.client.theme_dark = True
+
+    q.client.ner_tags = constants.NER_TAGS
+    q.client.ner_data = constants.NER_DATA
+    q.client.ner_index = 0
+    q.client.disable_next = False
+    q.client.disable_previous = True
+
+    # Add layouts, header and footer
+    q.page['meta'] = cards.meta
+    q.page['header'] = cards.header
+    q.page['footer'] = cards.footer
+
+    # Add cards for main page
+    q.page['ner_entities'] = cards.ner_entities(ner_tags=q.client.ner_tags)
+    q.page['ner_annotator'] = cards.ner_annotator(
+        ner_tags=q.client.ner_tags,
+        ner_items=q.client.ner_data[q.client.ner_index],
+        disable_next=q.client.disable_next,
+        disable_previous=q.client.disable_previous
+    )
+
+    q.client.initialized = True
+
+    await q.page.save()
+
+
+async def update_theme(q: Q):
+    """
+    Update theme of app.
+    """
+
+    # Copying argument values to client
+    copy_expando(q.args, q.client)
+
+    if q.client.theme_dark:
+        logging.info('Updating theme to dark mode')
+
+        # Update theme from light to dark mode
+        q.page['meta'].theme = 'h2o-dark'
+        q.page['header'].icon_color = 'black'
+    else:
+        logging.info('Updating theme to light mode')
+
+        # Update theme from dark to light mode
+        q.page['meta'].theme = 'light'
+        q.page['header'].icon_color = '#FEC924'
+
+    await q.page.save()
+
+
+@on('next')
+async def show_next_text(q: Q):
+    """
+    Show next NER data.
+    """
+
+    logging.info('Showing the next NER data')
+
+    # Save annotation
+    copy_expando(q.args, q.client)
+    q.client.ner_data[q.client.ner_index] = q.client.ner_annotator
+
+    # Move to next text
+    q.client.ner_index += 1
+    q.client.disable_previous = False
+
+    # Disable 'Next' if last text
+    if q.client.ner_index == len(q.client.ner_data) - 1:
+        q.client.disable_next = True
+
+    # Display data
+    q.page['ner_annotator'] = cards.ner_annotator(
+        ner_tags=q.client.ner_tags,
+        ner_items=q.client.ner_data[q.client.ner_index],
+        disable_next=q.client.disable_next,
+        disable_previous=q.client.disable_previous
+    )
+
+    await q.page.save()
+
+
+@on('previous')
+async def show_previous_text(q: Q):
+    """
+    Show previous NER data.
+    """
+
+    logging.info('Showing the previous NER data')
+
+    # Save annotation
+    copy_expando(q.args, q.client)
+    q.client.ner_data[q.client.ner_index] = q.client.ner_annotator
+
+    # Move to previous text
+    q.client.ner_index -= 1
+    q.client.disable_next = False
+
+    # Disable 'Previous' if first text
+    if q.client.ner_index == 0:
+        q.client.disable_previous = True
+
+    # Display data
+    q.page['ner_annotator'] = cards.ner_annotator(
+        ner_tags=q.client.ner_tags,
+        ner_items=q.client.ner_data[q.client.ner_index],
+        disable_next=q.client.disable_next,
+        disable_previous=q.client.disable_previous
+    )
+
+    await q.page.save()
+
+
+@on('add')
+async def add_entity(q: Q):
+    """
+    Add a new entity to NER tags.
+    """
+
+    logging.info('Adding a new entity')
+
+    # Save annotation
+    copy_expando(q.args, q.client)
+    q.client.ner_data[q.client.ner_index] = q.client.ner_annotator
+
+    # Add new entity
+    if len(q.client.new_entity_name) > 0:
+        q.client.ner_tags.append({
+            'name': q.client.new_entity_name.lower(),
+            'label': q.client.new_entity_name,
+            'color': '#{:02x}{:02x}{:02x}'.format(randint(0, 255), randint(0, 255), randint(0, 255))
+        })
+
+    # Refresh data with new entity
+    q.page['ner_entities'] = cards.ner_entities(ner_tags=q.client.ner_tags)
+    q.page['ner_annotator'] = cards.ner_annotator(
+        ner_tags=q.client.ner_tags,
+        ner_items=q.client.ner_data[q.client.ner_index],
+        disable_next=q.client.disable_next,
+        disable_previous=q.client.disable_previous
+    )
+
+    await q.page.save()
+
+
+@on('delete')
+async def delete_entity(q: Q):
+    """
+    Delete an entity from NER tags.
+    """
+
+    logging.info('Deleting an entity')
+
+    # Save annotation
+    copy_expando(q.args, q.client)
+    q.client.ner_data[q.client.ner_index] = q.client.ner_annotator
+
+    # Delete entity and it's tags
+    if len(q.client.ner_tags) > 1:
+        q.client.ner_tags = [tag for tag in q.client.ner_tags if tag['name'] != q.client.delete_entity_name]
+        for text in q.client.ner_data:
+            for item in text:
+                if 'tag' in item.keys():
+                    if item['tag'] == q.client.delete_entity_name:
+                        item.pop('tag')
+    else:
+        logging.info('No entities deleted since annotator requires at least one entity available')
+
+    # Refresh data with remaining entities
+    q.page['ner_entities'] = cards.ner_entities(ner_tags=q.client.ner_tags)
+    q.page['ner_annotator'] = cards.ner_annotator(
+        ner_tags=q.client.ner_tags,
+        ner_items=q.client.ner_data[q.client.ner_index],
+        disable_next=q.client.disable_next,
+        disable_previous=q.client.disable_previous
+    )
+
+    await q.page.save()
+
+
+def clear_cards(q: Q, card_names: list):
+    """
+    Clear cards from the page.
+    """
+
+    logging.info('Clearing cards')
+
+    # Delete cards from the page
+    for card_name in card_names:
+        del q.page[card_name]
+
+
+async def show_error(q: Q, error: str):
+    """
+    Displays errors.
+    """
+
+    logging.error(error)
+
+    # Clear all cards
+    clear_cards(q, q.app.cards)
+
+    # Format and display the error
+    q.page['error'] = cards.crash_report(q)
+
+    await q.page.save()
+
+
+@on('reload')
+async def reload_client(q: Q):
+    """
+    Reset the client (browser tab).
+    This function is called when the user clicks "Reload" on the crash report.
+    """
+
+    logging.info('Reloading client')
+
+    # Clear all cards
+    clear_cards(q, q.app.cards)
+
+    # Reload the client
+    await initialize_client(q)
+
+
+async def handle_fallback(q: Q):
+    """
+    Handle fallback cases.
+    This function should never get called unless there is a bug in our code or query handling logic.
+    """
+
+    logging.info('Adding fallback page')
+
+    q.page['fallback'] = cards.fallback
+
+    await q.page.save()

cards.py

@@ -0,0 +1,177 @@
+import sys
+import traceback
+
+from h2o_wave import Q, expando_to_dict, ui
+
+# App name
+app_name = 'NER Annotation'
+
+# Link to repo. Report bugs/features here :)
+repo_url = 'https://github.com/vopani/waveton'
+issue_url = f'{repo_url}/issues/new?assignees=vopani&labels=bug&template=error-report.md&title=%5BERROR%5D'
+
+# A meta card to hold the app's title, layouts, dialogs, theme and other meta information
+meta = ui.meta_card(
+    box='',
+    title='WaveTon',
+    layouts=[
+        ui.layout(
+            breakpoint='xs',
+            zones=[
+                ui.zone(name='header'),
+                ui.zone(
+                    name='main',
+                    size='calc(100vh - 150px)',
+                    direction='row',
+                    zones=[
+                        ui.zone(name='ner_entities', size='20%'),
+                        ui.zone(name='ner_annotator', size='80%')
+                    ]
+                ),
+                ui.zone(name='footer')
+            ]
+        )
+    ],
+    theme='h2o-dark'
+)
+
+# The header shown on all the app's pages
+header = ui.header_card(
+    box='header',
+    title='NER Annotation',
+    subtitle='Annotate entities for Named-Entity Recognition tasks',
+    icon='Handwriting',
+    icon_color='black',
+    items=[
+        ui.toggle(name='theme_dark', label='Dark Mode', value=True, trigger=True)
+    ]
+)
+
+# The footer shown on all the app's pages
+footer = ui.footer_card(
+    box='footer',
+    caption=f'Learn more about <a href="{repo_url}" target="_blank"> WaveTon: 💯 Wave Applications</a>'
+)
+
+# A fallback card for handling bugs
+fallback = ui.form_card(
+    box='fallback',
+    items=[ui.text('Uh-oh, something went wrong!')]
+)
+
+
+def ner_entities(ner_tags: list[dict]) -> ui.FormCard:
+    """
+    Card for NER entities.
+    """
+
+    card = ui.form_card(
+        box='ner_entities',
+        items=[
+            ui.textbox(name='new_entity_name', label='Type a new entity to be added'),
+            ui.buttons(
+                items=[
+                    ui.button(name='add', label='Add', primary=True)
+                ],
+                justify='center'
+            ),
+            ui.separator(),
+            ui.dropdown(
+                name='delete_entity_name',
+                label='Select an entity to delete',
+                choices=[ui.choice(name=tag['name'], label=tag['label']) for tag in ner_tags]
+            ),
+            ui.buttons(
+                items=[
+                    ui.button(name='delete', label='Delete', primary=True)
+                ],
+                justify='center'
+            )
+        ]
+    )
+
+    return card
+
+
+def ner_annotator(
+    ner_tags: list[dict],
+    ner_items: list[dict],
+    disable_next: bool = False,
+    disable_previous: bool = False
+) -> ui.FormCard:
+    """
+    Card for NER annotator.
+    """
+
+    card = ui.form_card(
+        box='ner_annotator',
+        items=[
+            ui.text_annotator(
+                name='ner_annotator',
+                title='Click and/or drag text to annotate',
+                tags=[ui.text_annotator_tag(**tag) for tag in ner_tags],
+                items=[ui.text_annotator_item(**item) for item in ner_items]
+            ),
+            ui.buttons(
+                items=[
+                    ui.button(name='next', label='Next', primary=True, disabled=disable_next),
+                    ui.button(name='previous', label='Previous', disabled=disable_previous)
+                ]
+            )
+        ]
+    )
+
+    return card
+
+
+def crash_report(q: Q) -> ui.FormCard:
+    """
+    Card for capturing the stack trace and current application state, for error reporting.
+    This function is called by the main serve() loop on uncaught exceptions.
+    """
+
+    def code_block(content): return '\n'.join(['```', *content, '```'])
+
+    type_, value_, traceback_ = sys.exc_info()
+    stack_trace = traceback.format_exception(type_, value_, traceback_)
+
+    dump = [
+        '### Stack Trace',
+        code_block(stack_trace),
+    ]
+
+    states = [
+        ('q.app', q.app),
+        ('q.user', q.user),
+        ('q.client', q.client),
+        ('q.events', q.events),
+        ('q.args', q.args)
+    ]
+    for name, source in states:
+        dump.append(f'### {name}')
+        dump.append(code_block([f'{k}: {v}' for k, v in expando_to_dict(source).items()]))
+
+    return ui.form_card(
+        box='main',
+        items=[
+            ui.stats(
+                items=[
+                    ui.stat(
+                        label='',
+                        value='Oops!',
+                        caption='Something went wrong',
+                        icon='Error'
+                    )
+                ],
+            ),
+            ui.separator(),
+            ui.text_l(content='Apologies for the inconvenience!'),
+            ui.buttons(items=[ui.button(name='reload', label='Reload', primary=True)]),
+            ui.expander(name='report', label='Error Details', items=[
+                ui.text(
+                    f'To report this issue, <a href="{issue_url}" target="_blank">please open an issue</a> with the details below:'),
+                ui.text_l(content=f'Report Issue in App: **{app_name}**'),
+                ui.text(content='\n'.join(dump)),
+            ])
+        ]
+    )

constants.py

@@ -0,0 +1,125 @@
+NER_TAGS = [
+    {"name": "organization", "label": "Organization", "color": "#F1CBCB"},
+    {"name": "metric", "label": "Metric", "color": "#CAEACA"}
+]
+
+NER_DATA = [
+    [
+        {"text": "At "},
+        {"text": "Santander", "tag": "organization"},
+        {"text": " our mission is to help people and businesses prosper. "},
+        {"text": "We are always looking for ways to help our customers understand their financial health "},
+        {"text": "and identify which products and services might help them achieve their monetary goals. "},
+        {"text": "Our data science team is continually challenging our machine learning algorithms, working with "},
+        {"text": "the global data science community to make sure we can more accurately identify new ways "},
+        {"text": "to solve our most common challenge, binary classification problems such as: "},
+        {"text": "is a customer satisfied? Will a customer buy this product? Can a customer pay this loan? "},
+        {"text": "In this challenge, we invite Kagglers to help us identify which customers will make "},
+        {"text": "a specific transaction in the future, irrespective of the amount of money transacted. "},
+        {"text": "The data provided for this competition has the same structure as the real data we have available "},
+        {"text": "to solve this problem."}
+    ],
+    [
+        {"text": "Many people struggle to get loans due to insufficient or non-existent credit histories. "},
+        {"text": "And, unfortunately, this population is often taken advantage of by untrustworthy lenders. "},
+        {"text": "Home Credit", "tag": "organization"},
+        {"text": " strives to broaden financial inclusion for the unbanked population by providing "},
+        {"text": "a positive and safe borrowing experience. "},
+        {"text": "In order to make sure this underserved population has a positive loan experience, "},
+        {"text": "Home Credit", "tag": "organization"},
+        {"text": " makes use of a variety of alternative data--including telco & transactional information"},
+        {"text": "--to predict their clients repayment abilities. While "},
+        {"text": "Home Credit", "tag": "organization"},
+        {"text": " is currently using various statistical and machine learning methods to make "},
+        {"text": "predictions, they're challenging Kagglers to help them unlock "},
+        {"text": "the full potential of their data. "},
+        {"text": "Doing so will ensure that clients capable of repayment are not rejected "},
+        {"text": "and that loans are given with a principal, maturity, and repayment calendar that will empower "},
+        {"text": "their clients to be successful."}
+    ],
+    [
+        {"text": "Imagine standing at the check-out counter at the grocery store with a long line behind you "},
+        {"text": "and the cashier not-so-quietly announces that your card has been declined. "},
+        {"text": "In this moment, you probably aren’t thinking about the data science that determined your fate. "},
+        {"text": "Embarrassed, and certain you have the funds to cover everything needed for an epic "},
+        {"text": "nacho party for 50 of your closest friends, you try your card again. "},
+        {"text": "Same result. As you step aside and allow the cashier to tend to the next customer, "},
+        {"text": "you receive a text message from your bank. "},
+        {"text": "'Press 1 if you really tried to spend $500 on cheddar cheese.' "},
+        {"text": "While perhaps cumbersome (and often embarrassing) in the moment, "},
+        {"text": "this fraud prevention system is actually saving consumers millions of dollars per year. "},
+        {"text": "Researchers from the "},
+        {"text": "IEEE Computational Intelligence Society (IEEE-CIS)", "tag": "organization"},
+        {"text": " want to improve this figure, while also improving the customer experience. With higher "},
+        {"text": "accuracy", "tag": "metric"},
+        {"text": " fraud detection, you can get on with your chips without the hassle. "},
+        {"text": "IEEE-CIS", "tag": "organization"},
+        {"text": " works across a variety of AI and machine learning areas, including deep neural networks, "},
+        {"text": "fuzzy systems, evolutionary computation, and swarm intelligence. "},
+        {"text": "Today they’re partnering with the world’s leading payment service company, "},
+        {"text": "Vesta Corporation", "tag": "organization"},
+        {"text": ", seeking the best solutions for fraud prevention industry, "},
+        {"text": "and now you are invited to join the challenge. "},
+        {"text": "In this competition, you’ll benchmark machine learning models on a challenging large-scale dataset. "},
+        {"text": "The data comes from "},
+        {"text": "Vesta", "tag": "organization"},
+        {"text": "'s real-world e-commerce transactions "},
+        {"text": "and contains a wide range of features from device type to product features. "},
+        {"text": "You also have the opportunity to create new features to improve your results. "},
+        {"text": "If successful, you’ll improve the efficacy of fraudulent transaction alerts for millions of people "},
+        {"text": "around the world, helping hundreds of thousands of businesses reduce their "},
+        {"text": "fraud loss", "tag": "metric"},
+        {"text": " and increase their "},
+        {"text": "revenue", "tag": "metric"},
+        {"text": ". And of course, you will save party people just like you the hassle of "},
+        {"text": "false positives", "tag": "metric"},
+        {"text": "."}
+    ],
+    [
+        {"text": "How much camping gear will one store sell each month in a year? "},
+        {"text": "To the uninitiated, calculating sales at this level may seem as difficult as predicting the weather. "},
+        {"text": "Both types of forecasting rely on science and historical data. "},
+        {"text": "While a wrong weather forecast may result in you carrying around an umbrella on a sunny day, "},
+        {"text": "inaccurate business forecasts could result in actual or opportunity losses. "},
+        {"text": "In this competition, in addition to traditional forecasting methods you’re also challenged to use "},
+        {"text": "machine learning to improve forecast "},
+        {"text": "accuracy", "tag": "metric"},
+        {"text": ". The Makridakis Open Forecasting Center (MOFC) at the "},
+        {"text": "University of Nicosia", "tag": "organization"},
+        {"text": " conducts cutting-edge forecasting research and provides business forecast training. "},
+        {"text": "It helps companies achieve accurate predictions, estimate the levels of uncertainty, "},
+        {"text": "avoiding costly mistakes, and apply best forecasting practices. "},
+        {"text": "The MOFC is well known for its Makridakis Competitions, the first of which ran in the 1980s. "},
+        {"text": "In this competition, the fifth iteration, you will use hierarchical sales data from Walmart, "},
+        {"text": "the world’s largest company by "},
+        {"text": "revenue", "tag": "metric"},
+        {"text": ", to forecast daily sales for the next 28 days. "},
+        {"text": "The data, covers stores in three US States (California, Texas, and Wisconsin) "},
+        {"text": "and includes item level, department, product categories, and store details. "},
+        {"text": "In addition, it has explanatory variables such as "},
+        {"text": "price, promotions, day of the week, and special events. "},
+        {"text": "Together, this robust dataset can be used to improve forecasting "},
+        {"text": "accuracy", "tag": "metric"},
+        {"text": ". If successful, your work will continue to advance the theory and practice of forecasting. "},
+        {"text": "The methods used can be applied in various business areas, such as setting up appropriate "},
+        {"text": "inventory or service levels. Through its business support and training, "},
+        {"text": "the MOFC will help distribute the tools and knowledge so others can achieve more accurate "},
+        {"text": "and better calibrated forecasts, reduce waste and be able to appreciate uncertainty and its risk "},
+        {"text": "implications."}
+    ],
+    [
+        {"text": "Nothing ruins the thrill of buying a brand new car more quickly than seeing your new insurance bill. "},
+        {"text": "The sting’s even more painful when you know you’re a good driver. "},
+        {"text": "It doesn’t seem fair that you have to pay so much if you’ve been cautious on the road for years. "},
+        {"text": "Porto Seguro, one of Brazil’s largest auto and homeowner insurance companies, completely agrees. "},
+        {"text": "Inaccuracies in car insurance company’s claim predictions raise the cost of insurance for "},
+        {"text": "good drivers and reduce the price for bad ones. "},
+        {"text": "In this competition, you’re challenged to build a model that predicts the probability that "},
+        {"text": "a driver will initiate an auto insurance claim in the next year. While "},
+        {"text": "Porto Seguro", "tag": "organization"},
+        {"text": " has used machine learning for the past 20 years, "},
+        {"text": "they’re looking to Kaggle’s machine learning community to explore new, more powerful methods. "},
+        {"text": "A more accurate prediction will allow them to further tailor their prices, and hopefully "},
+        {"text": "make auto insurance coverage more accessible to more drivers."}
+    ]
+]

requirements.txt

@@ -0,0 +1 @@
+h2o_wave==0.23.1