README.md

metadata

language:
  - af
  - am
  - ar
  - bn
  - bg
  - ca
  - hr
  - cs
  - da
  - nl
  - en
  - et
  - tl
  - fi
  - fr
  - de
  - el
  - gu
  - he
  - hi
  - hu
  - id
  - it
  - ja
  - kn
  - ko
  - lv
  - lt
  - ms
  - ml
  - mr
  - 'no'
  - fa
  - pl
  - pt
  - ro
  - ru
license: other
license_name: other
license_details: >-
  This dataset is proprietary to Keeper Security and intended for internal use
  only.
tags: []
annotations_creators:
  - machine-generated
pretty_name: Html Input Text Only Dataset
size_categories:
  - 100K<n<1M
task_categories:
  - text-classification
task_ids: []
configs:
  - config_name: default
    data_files:
      - split: train
        path: html-input-text-only-1755812747.parquet
dataset_info:
  features:
    - name: text
      dtype: string
    - name: label
      dtype: string
    - name: lang
      dtype: string
  config_name: default
  splits:
    - name: train
      num_bytes: 21009409
      num_examples: 449994
  download_size: 21009409
  dataset_size: 21009409

📖 Html Input Text Only Dataset

This dataset contains HTML input text with all punctuation removed, making it ideal for training models on web form elements and URLs.

📊 Dataset Overview

A preprocessed version of keeper-security/html_input_dataset with punctuation removed from HTML sections to reduce tokenization errors during model training. Contains web form elements and URLs for security analysis tasks.

This dataset is designed to facilitate the training of models that require clean HTML input without the noise introduced by punctuation.

🔗 Dependencies

The dataset uses data from the following datasets:

• keeper-security/html_input_dataset

---

💡 Usage Examples

Load with Hugging Face Datasets

from datasets import load_dataset
ds = load_dataset("keeper-security/html-input-text-only", data_files={"train": "html-input-text-only-1755812747.parquet"})
print(ds["train"][0])

Convert to Pandas

import pandas as pd
df_train = ds["train"].to_pandas()
df_train.head()

---

📁 Repository Structure

html-input-text-only/
├── html-input-text-only-1755812747.parquet # Main processed dataset
├── README.md                               # Dataset documentation
└── CHANGELOG.md                            # Version history and changes

---

📊 Data Overview

• Shape: (449,994 rows × 3 columns)
• File Size: ~20 MB

---

📌 Statistics

• 🎯 Unique text: 302,616
• 🎯 Unique label: 45
• 🎯 Unique lang: 38

Columns

Column Name	Type	Description
text	string	The main text content URL and HTML combined
label	string	Classification label
lang	string	Language of the text

Ontology

lang	Description
Gujarati	Gujarati language
Japanese	Japanese language
Lithuanian	Lithuanian language
Indonesian	Indonesian language
Bengali	Bengali language
Afrikaans	Afrikaans language
Estonian	Estonian language
Dutch	Dutch language
Italian	Italian language
Hebrew	Hebrew language
Croatian	Croatian language
Finnish	Finnish language
Filipino	Filipino language
English	English language
Bulgarian	Bulgarian language
Korean	Korean language
Czech	Czech language
Danish	Danish language
Hindi	Hindi language
Latvian	Latvian language
German	German language
unknown	Unknown or unclassified language
Amharic	Amharic language
Hungarian	Hungarian language
Arabic	Arabic language
Greek	Greek language
Catalan	Catalan language
Kannada	Kannada language
French	French language
Marathi	Marathi language
Malay	Malay language
Malayalam	Malayalam language
Persian	Persian language
Norwegian Bokmål	Norwegian language
Polish	Polish language
Romanian	Romanian language
Portuguese	Portuguese language
Russian	Russian language

label	Description
ACCOUNT_CREATION_PASSWORD	Account creation password
ADDRESS_CITY	City name in an address
ADDRESS_COUNTRY	Country name in an address
ADDRESS_LINE1	First line of a street address
ADDRESS_LINE2	Second line of a street address (e.g., apartment, suite)
ADDRESS_STATE	State, province, or region in an address
ADDRESS_ZIP	Postal or ZIP code in an address
ALTERNATIVE_FAMILY_NAME	Alternative or secondary family/last name
ALTERNATIVE_FULL_NAME	Alternative or secondary full name
ALTERNATIVE_GIVEN_NAME	Alternative or secondary given/first name
AMBIGUOUS	Ambiguous or unclear input that cannot be classified
BIRTH_DATE_DAY	Day component of a birth date
BIRTH_DATE_MONTH	Month component of a birth date
BIRTH_DATE_YEAR	Year component of a birth date
COMPANY_NAME	Name of a company or organization
CONFIRMATION_PASSWORD	Password confirmation field
CREDIT_CARD_EXP_DATE_MONTH_AND_YEAR	Credit card expiration date (month and year)
CREDIT_CARD_EXP_DATE_YEAR	Credit card expiration year
CREDIT_CARD_EXP_MONTH	Credit card expiration month
CREDIT_CARD_NUMBER	Credit card number
CREDIT_CARD_STANDALONE_VERIFICATION_CODE	Credit card verification code (standalone field)
CREDIT_CARD_TYPE	Type or brand of credit card (e.g., Visa, MasterCard)
CREDIT_CARD_VERIFICATION_CODE	Credit card verification code (CVV, CVC, etc.)
EMAIL_ADDRESS	Email address
IBAN_VALUE	International Bank Account Number (IBAN)
MALICIOUS_LABEL	Label for potentially malicious or harmful input
MERCHANT_EMAIL_SIGNUP	Email address used for merchant signup
MERCHANT_PROMO_CODE	Promotional or discount code for merchants
NAME_FIRST	First or given name
NAME_FULL	Full name
NAME_LAST	Last or family name
NAME_MIDDLE	Middle name
NAME_MIDDLE_INITIAL	Middle initial
NAME_PREFIX	Name prefix (e.g., Mr., Dr., Ms.)
NAME_SUFFIX	Name suffix (e.g., Jr., Sr., III)
NATIONAL_IDENTITY_NUMBER	National identity or government-issued number
NEW_PASSWORD	New password field
PASSWORD	Password field
PHONE_NUMBER	Phone number
PIN_CODE	Personal identification number (PIN)
PROBABLY_NEW_PASSWORD	Field likely to be a new password
SEARCH	Search query or search field
TWO_FACTOR_CODE	Two-factor authentication code
UNKNOWN	Unknown or unclassified input
USERNAME	Username or user ID

---

---

🔧 Reproduction

1. Clone the repository:

git clone https://github.com/keeper-security/ai-factory.git
cd ai-factory
git checkout 245a96b976f46aefbd71fc1716c16fd69d503b73

2. Navigate to the dataset directory and run the pipeline:

cd datasets/html-input-text-only
python run_pipeline.py

---