textinsight-nlp 0.1.0__tar.gz

textinsight_nlp-0.1.0/.gitignore

@@ -0,0 +1,34 @@
+# Python
+__pycache__/
+*.py[cod]
+*.egg-info/
+build/
+dist/
+
+# Environments
+.venv/
+venv/
+.env
+
+# Tools
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.coverage
+coverage.xml
+htmlcov/
+
+# Editors / OS
+.vscode/
+.idea/
+Thumbs.db
+.DS_Store
+
+# Large data and model files never belong in git
+/data/
+*.onnx
+*.joblib
+*.bin
+
+# Office lock files
+~$*

textinsight_nlp-0.1.0/CHANGELOG.md

@@ -0,0 +1,24 @@
+# Changelog
+
+All notable changes to this project are documented here.
+The format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and the project uses [Semantic Versioning](https://semver.org/).
+
+## [0.1.0] - Unreleased
+
+### Added
+- Project skeleton: src-layout package, `pyproject.toml`, ruff, mypy, pytest, pre-commit, GitHub Actions CI.
+- Core foundation: configuration, exception hierarchy, input validation, typed result classes,
+  backend base class, backend registry with `"auto"` selection, and a model manager with
+  on-disk caching, SHA-256 verification and offline mode.
+- Configurable `clean()`: HTML, URLs, emails, mentions, hashtags, emojis, numbers, punctuation.
+- `tokenize()` and `sentences()` for informal English and Hinglish, including Devanagari and the danda.
+- Hinglish spelling normalization: `normalize_hinglish()` and `cluster_variants()` for mining new variants.
+- `normalize()` and `remove_stopwords()` for English and Hinglish.
+- `detect_language()` with a from-scratch character n-gram classifier.
+- `sentiment()` with three backends: `lexicon` (English and Hinglish, with post-word negation),
+  bundled `classical` (NB-LR with Hinglish spelling features, trained on SentiMix), and optional
+  `hybrid` (blended with HingBERT, via the `transformers` extra). Short, plain text falls back to the lexicon when the model
+  says neutral but the lexicon is clear.
+- Evaluation tools: SentiMix loader, metrics, and the `textinsight benchmark` command.
+- Benchmarks on the held-out SentiMix test split, and a model card.

textinsight_nlp-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Harsh Vardhan Ajmera
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

textinsight_nlp-0.1.0/PKG-INFO

@@ -0,0 +1,163 @@
+Metadata-Version: 2.5
+Name: textinsight-nlp
+Version: 0.1.0
+Summary: Lightweight, offline-first NLP for English and Hinglish text, with pluggable backends.
+Project-URL: Homepage, https://github.com/harsh01103/textinsight
+Project-URL: Issues, https://github.com/harsh01103/textinsight/issues
+Author: Harsh Vardhan Ajmera
+License-Expression: MIT
+License-File: LICENSE
+Keywords: code-mixed,hinglish,keywords,nlp,sentiment,summarization,text-analysis
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Text Processing :: Linguistic
+Classifier: Typing :: Typed
+Requires-Python: >=3.10
+Provides-Extra: dev
+Requires-Dist: build>=1.2; extra == 'dev'
+Requires-Dist: mypy>=1.10; extra == 'dev'
+Requires-Dist: pre-commit>=3.7; extra == 'dev'
+Requires-Dist: pytest-cov>=5; extra == 'dev'
+Requires-Dist: pytest>=8; extra == 'dev'
+Requires-Dist: ruff>=0.6; extra == 'dev'
+Requires-Dist: twine>=5; extra == 'dev'
+Provides-Extra: training
+Requires-Dist: scikit-learn>=1.5; extra == 'training'
+Requires-Dist: vadersentiment>=3.3.2; extra == 'training'
+Provides-Extra: transformers
+Requires-Dist: torch>=2.2; extra == 'transformers'
+Requires-Dist: transformers>=4.40; extra == 'transformers'
+Description-Content-Type: text/markdown
+
+# TextInsight
+
+**Lightweight, offline-first NLP for English and Hinglish text, with pluggable backends.**
+
+> Status: alpha. Preprocessing, language detection and sentiment work; other tasks arrive in upcoming releases.
+> See `TextInsight_Architecture_and_Roadmap.docx` for the full plan.
+
+## Why TextInsight?
+
+- **Built for Hinglish** – handles code-mixed text such as *"yeh phone mast hai but battery bekaar hai"*: spelling variants, Hindi word order for negation, and a model trained on real Hinglish tweets.
+- **Lightweight by default** – small base install; heavy models are optional extras.
+- **No surprises** – importing the library never downloads anything.
+- **Pluggable backends** – swap lexicon, classical ML, ONNX, transformer or LLM implementations with one argument.
+- **Honest benchmarks** – every backend ships with measured results.
+
+## Installation
+
+```bash
+pip install textinsight-nlp
+```
+
+The package installs as `textinsight-nlp` and is imported as `textinsight`.
+
+## Quick start
+
+```python
+import textinsight as ti
+
+ti.clean("Visit https://example.com now!", lowercase=True)  # 'visit now!'
+ti.tokenize("Don't miss it :) #sale")  # ["Don't", 'miss', 'it', ':)', '#sale']
+ti.sentences("Dr. Rao aaye. yeh accha hai!")  # ['Dr. Rao aaye.', 'yeh accha hai!']
+ti.normalize_hinglish("bohot acha hai")  # 'bahut accha hai'
+ti.detect_language("yeh phone mast hai but battery bekaar").language  # 'hi-en'
+ti.sentiment("bahut accha product hai")  # SentimentResult(label='positive', ...)
+ti.sentiment("yeh accha nahi hai", backend="lexicon")  # handles Hindi negation
+
+ti.list_backends()  # registered backends
+ti.config.offline = True  # never download models
+```
+
+Language labels are `en`, `hi` (mainly Devanagari), `hi-en` (code-mixed or romanized Hindi)
+and `und` (undetermined).
+
+## Benchmarks
+
+Sentiment on the held-out test split of [SentiMix](https://doi.org/10.5281/zenodo.3974927)
+(3,000 real Hinglish tweets):
+
+| Method | Accuracy | Macro F1 | Weighted F1 |
+|---|---|---|---|
+| Majority class | 0.367 | 0.179 | 0.197 |
+| VADER | 0.492 | 0.470 | 0.475 |
+| TextInsight `lexicon` | 0.561 | 0.540 | 0.546 |
+| **TextInsight `classical`** (default, 529 KB, no dependencies) | **0.702** | **0.707** | **0.703** |
+| **TextInsight `hybrid`** (optional transformer blend) | **0.723** | **0.728** | **0.725** |
+
+Cross-validated macro F1 on the other 17,000 tweets is lower and more conservative: 0.639 for
+`classical` and 0.666 for `hybrid`. Word-level language detection on the test tweets: 72.7%.
+
+Short, plain sentences such as reviews use a lexicon fallback, because a model trained on tweets
+reads them as neutral too often. On 60 hand-written review-style Hinglish sentences kept back for
+this check, macro F1 rose from 0.581 to 0.819 for `classical` and from 0.499 to 0.834 for `hybrid`.
+That set is small and not independent, so treat it as a behaviour check.
+
+The test-split scores above are for tweets only. See the
+[model card](docs/model_cards/sentiment_classical.md) for method, limitations and bias.
+
+To reproduce them, install the `training` extra, download the dataset and run:
+
+```bash
+textinsight benchmark sentiment --data-dir path/to/Hinglish --split test
+```
+
+## Most accurate option
+
+```bash
+pip install "textinsight-nlp[transformers]"
+```
+
+```python
+ti.sentiment("zabardast movie thi, maza aa gaya", backend="hybrid")
+```
+
+First use downloads the 438 MB HingBERT encoder, so `hybrid` runs only when you ask for it.
+
+## Credits
+
+The bundled sentiment model derives from SentiMix (CC-BY-4.0): Patwa et al., "SemEval-2020
+Task 9: Overview of Sentiment Analysis of Code-Mixed Tweets", SemEval-2020. The `hybrid` backend
+uses HingBERT (CC-BY-4.0): Joshi, "L3Cube-HingCorpus and HingBERT", 2022.
+
+## Planned features
+
+| Release | Features |
+|---|---|
+| 0.1.0 | Preprocessing, language ID, English + Hinglish sentiment, benchmarks |
+| 0.2.0 | Keywords (TF-IDF, RAKE, TextRank), extractive summarization |
+| 0.3.0 | Aspect sentiment, text stats, PII redaction, `analyze()`, CLI, pandas |
+| 1.0.0 | Optional ONNX / transformer / LLM backends, docs site, live demo |
+
+## Development
+
+```bash
+python -m venv .venv
+.venv\Scripts\activate          # Windows  (source .venv/bin/activate on macOS/Linux)
+pip install -e ".[dev]"
+pre-commit install
+
+pytest                          # tests + coverage
+ruff check . && ruff format --check .
+mypy
+```
+
+## Configuration
+
+| Setting | Environment variable | Default |
+|---|---|---|
+| `config.cache_dir` | `TEXTINSIGHT_CACHE_DIR` | OS user cache folder + `/textinsight` |
+| `config.offline` | `TEXTINSIGHT_OFFLINE` | `False` |
+| `config.device` | `TEXTINSIGHT_DEVICE` | `"cpu"` |
+| `config.max_input_chars` | `TEXTINSIGHT_MAX_INPUT_CHARS` | `1_000_000` |
+
+## License
+
+MIT

textinsight_nlp-0.1.0/README.md

@@ -0,0 +1,125 @@
+# TextInsight
+
+**Lightweight, offline-first NLP for English and Hinglish text, with pluggable backends.**
+
+> Status: alpha. Preprocessing, language detection and sentiment work; other tasks arrive in upcoming releases.
+> See `TextInsight_Architecture_and_Roadmap.docx` for the full plan.
+
+## Why TextInsight?
+
+- **Built for Hinglish** – handles code-mixed text such as *"yeh phone mast hai but battery bekaar hai"*: spelling variants, Hindi word order for negation, and a model trained on real Hinglish tweets.
+- **Lightweight by default** – small base install; heavy models are optional extras.
+- **No surprises** – importing the library never downloads anything.
+- **Pluggable backends** – swap lexicon, classical ML, ONNX, transformer or LLM implementations with one argument.
+- **Honest benchmarks** – every backend ships with measured results.
+
+## Installation
+
+```bash
+pip install textinsight-nlp
+```
+
+The package installs as `textinsight-nlp` and is imported as `textinsight`.
+
+## Quick start
+
+```python
+import textinsight as ti
+
+ti.clean("Visit https://example.com now!", lowercase=True)  # 'visit now!'
+ti.tokenize("Don't miss it :) #sale")  # ["Don't", 'miss', 'it', ':)', '#sale']
+ti.sentences("Dr. Rao aaye. yeh accha hai!")  # ['Dr. Rao aaye.', 'yeh accha hai!']
+ti.normalize_hinglish("bohot acha hai")  # 'bahut accha hai'
+ti.detect_language("yeh phone mast hai but battery bekaar").language  # 'hi-en'
+ti.sentiment("bahut accha product hai")  # SentimentResult(label='positive', ...)
+ti.sentiment("yeh accha nahi hai", backend="lexicon")  # handles Hindi negation
+
+ti.list_backends()  # registered backends
+ti.config.offline = True  # never download models
+```
+
+Language labels are `en`, `hi` (mainly Devanagari), `hi-en` (code-mixed or romanized Hindi)
+and `und` (undetermined).
+
+## Benchmarks
+
+Sentiment on the held-out test split of [SentiMix](https://doi.org/10.5281/zenodo.3974927)
+(3,000 real Hinglish tweets):
+
+| Method | Accuracy | Macro F1 | Weighted F1 |
+|---|---|---|---|
+| Majority class | 0.367 | 0.179 | 0.197 |
+| VADER | 0.492 | 0.470 | 0.475 |
+| TextInsight `lexicon` | 0.561 | 0.540 | 0.546 |
+| **TextInsight `classical`** (default, 529 KB, no dependencies) | **0.702** | **0.707** | **0.703** |
+| **TextInsight `hybrid`** (optional transformer blend) | **0.723** | **0.728** | **0.725** |
+
+Cross-validated macro F1 on the other 17,000 tweets is lower and more conservative: 0.639 for
+`classical` and 0.666 for `hybrid`. Word-level language detection on the test tweets: 72.7%.
+
+Short, plain sentences such as reviews use a lexicon fallback, because a model trained on tweets
+reads them as neutral too often. On 60 hand-written review-style Hinglish sentences kept back for
+this check, macro F1 rose from 0.581 to 0.819 for `classical` and from 0.499 to 0.834 for `hybrid`.
+That set is small and not independent, so treat it as a behaviour check.
+
+The test-split scores above are for tweets only. See the
+[model card](docs/model_cards/sentiment_classical.md) for method, limitations and bias.
+
+To reproduce them, install the `training` extra, download the dataset and run:
+
+```bash
+textinsight benchmark sentiment --data-dir path/to/Hinglish --split test
+```
+
+## Most accurate option
+
+```bash
+pip install "textinsight-nlp[transformers]"
+```
+
+```python
+ti.sentiment("zabardast movie thi, maza aa gaya", backend="hybrid")
+```
+
+First use downloads the 438 MB HingBERT encoder, so `hybrid` runs only when you ask for it.
+
+## Credits
+
+The bundled sentiment model derives from SentiMix (CC-BY-4.0): Patwa et al., "SemEval-2020
+Task 9: Overview of Sentiment Analysis of Code-Mixed Tweets", SemEval-2020. The `hybrid` backend
+uses HingBERT (CC-BY-4.0): Joshi, "L3Cube-HingCorpus and HingBERT", 2022.
+
+## Planned features
+
+| Release | Features |
+|---|---|
+| 0.1.0 | Preprocessing, language ID, English + Hinglish sentiment, benchmarks |
+| 0.2.0 | Keywords (TF-IDF, RAKE, TextRank), extractive summarization |
+| 0.3.0 | Aspect sentiment, text stats, PII redaction, `analyze()`, CLI, pandas |
+| 1.0.0 | Optional ONNX / transformer / LLM backends, docs site, live demo |
+
+## Development
+
+```bash
+python -m venv .venv
+.venv\Scripts\activate          # Windows  (source .venv/bin/activate on macOS/Linux)
+pip install -e ".[dev]"
+pre-commit install
+
+pytest                          # tests + coverage
+ruff check . && ruff format --check .
+mypy
+```
+
+## Configuration
+
+| Setting | Environment variable | Default |
+|---|---|---|
+| `config.cache_dir` | `TEXTINSIGHT_CACHE_DIR` | OS user cache folder + `/textinsight` |
+| `config.offline` | `TEXTINSIGHT_OFFLINE` | `False` |
+| `config.device` | `TEXTINSIGHT_DEVICE` | `"cpu"` |
+| `config.max_input_chars` | `TEXTINSIGHT_MAX_INPUT_CHARS` | `1_000_000` |
+
+## License
+
+MIT

textinsight_nlp-0.1.0/pyproject.toml

@@ -0,0 +1,109 @@
+[build-system]
+requires = ["hatchling>=1.27"]
+build-backend = "hatchling.build"
+
+[project]
+name = "textinsight-nlp"
+dynamic = ["version"]
+description = "Lightweight, offline-first NLP for English and Hinglish text, with pluggable backends."
+readme = "README.md"
+license = "MIT"
+license-files = ["LICENSE"]
+requires-python = ">=3.10"
+authors = [{ name = "Harsh Vardhan Ajmera" }]
+keywords = ["nlp", "sentiment", "hinglish", "code-mixed", "keywords", "summarization", "text-analysis"]
+classifiers = [
+    "Development Status :: 3 - Alpha",
+    "Intended Audience :: Developers",
+    "Intended Audience :: Science/Research",
+    "Operating System :: OS Independent",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Topic :: Text Processing :: Linguistic",
+    "Typing :: Typed",
+]
+# Runtime dependencies are added in the phase that first uses them
+# (numpy / scikit-learn / regex arrive with preprocessing and sentiment).
+dependencies = []
+
+[project.optional-dependencies]
+dev = [
+    "pytest>=8",
+    "pytest-cov>=5",
+    "ruff>=0.6",
+    "mypy>=1.10",
+    "pre-commit>=3.7",
+    "build>=1.2",
+    "twine>=5",
+]
+transformers = [
+    "torch>=2.2",
+    "transformers>=4.40",
+]
+training = [
+    "scikit-learn>=1.5",
+    "vaderSentiment>=3.3.2",
+]
+
+[project.scripts]
+textinsight = "textinsight.cli:main"
+
+[project.urls]
+Homepage = "https://github.com/harsh01103/textinsight"
+Issues = "https://github.com/harsh01103/textinsight/issues"
+
+[tool.hatch.version]
+path = "src/textinsight/_version.py"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/textinsight"]
+
+[tool.hatch.build.targets.sdist]
+include = ["src/textinsight", "tests", "README.md", "CHANGELOG.md", "LICENSE"]
+
+# ---------------------------------------------------------------- pytest
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+addopts = "-ra --strict-markers --cov=textinsight --cov-report=term-missing"
+markers = [
+    "slow: tests that load real models or hit the network (deselect with -m 'not slow')",
+]
+
+[tool.coverage.run]
+branch = true
+source = ["textinsight"]
+
+[tool.coverage.report]
+fail_under = 90
+exclude_also = ["if TYPE_CHECKING:", "raise NotImplementedError", "\\.\\.\\."]
+
+# ---------------------------------------------------------------- ruff
+[tool.ruff]
+line-length = 100
+target-version = "py310"
+src = ["src", "tests"]
+
+[tool.ruff.lint]
+select = ["E", "F", "W", "I", "B", "UP", "SIM", "RUF", "PT", "N"]
+ignore = ["SIM905"]  # long word lists read better as one split string
+
+[tool.ruff.lint.per-file-ignores]
+# Tests use non-ASCII text on purpose (Devanagari, full-width letters, curly quotes).
+"tests/*" = ["RUF001"]
+# Curly quotes are matched deliberately when tokenizing.
+"src/textinsight/preprocess/*" = ["RUF001"]
+
+# ---------------------------------------------------------------- mypy
+[tool.mypy]
+python_version = "3.10"
+mypy_path = "src"
+packages = ["textinsight"]
+strict = true
+warn_unreachable = true
+
+[[tool.mypy.overrides]]
+module = ["torch", "transformers"]
+ignore_missing_imports = true

textinsight_nlp-0.1.0/src/textinsight/__init__.py

@@ -0,0 +1,60 @@
+"""TextInsight: lightweight, offline-first NLP for English and Hinglish text.
+
+Importing this package is cheap: it never downloads models or opens network
+connections. Models are fetched lazily the first time a backend needs them.
+"""
+
+import logging
+
+from ._version import __version__
+from .core import (
+    BackendInfo,
+    BackendNotAvailableError,
+    Config,
+    ConfigError,
+    InvalidInputError,
+    Keyword,
+    KeywordResult,
+    LanguageResult,
+    ModelDownloadError,
+    SentimentResult,
+    SummaryResult,
+    TextInsightError,
+    UnknownBackendError,
+    UnsupportedLanguageError,
+    config,
+    list_backends,
+)
+from .langid import detect_language
+from .preprocess import clean, normalize, normalize_hinglish, remove_stopwords, sentences, tokenize
+from .sentiment import sentiment
+
+logging.getLogger(__name__).addHandler(logging.NullHandler())
+
+__all__ = [
+    "BackendInfo",
+    "BackendNotAvailableError",
+    "Config",
+    "ConfigError",
+    "InvalidInputError",
+    "Keyword",
+    "KeywordResult",
+    "LanguageResult",
+    "ModelDownloadError",
+    "SentimentResult",
+    "SummaryResult",
+    "TextInsightError",
+    "UnknownBackendError",
+    "UnsupportedLanguageError",
+    "__version__",
+    "clean",
+    "config",
+    "detect_language",
+    "list_backends",
+    "normalize",
+    "normalize_hinglish",
+    "remove_stopwords",
+    "sentences",
+    "sentiment",
+    "tokenize",
+]

textinsight_nlp-0.1.0/src/textinsight/_data/__init__.py

@@ -0,0 +1 @@
+"""Small word lists bundled with the package."""

textinsight_nlp-0.1.0/src/textinsight/_data/english.py

@@ -0,0 +1,87 @@
+"""Common English words and stopwords."""
+
+from __future__ import annotations
+
+# Negations are left out on purpose: dropping "not" flips the meaning of a sentence.
+STOPWORDS_EN = frozenset(
+    """
+    a about above after again all also am an and any are as at be because been before being
+    below between both but by can could did do does doing down during each few for from
+    further had has have having he her here hers herself him himself his how i if in into is
+    it its itself just me more most my myself of off on once only or other our ours ourselves
+    out over own same she should so some such than that the their theirs them themselves then
+    there these they this those through to too under until up very was we were what when
+    where which while who whom why will with would you your yours yourself yourselves
+    """.split()
+)
+
+ENGLISH_COMMON = frozenset(
+    """
+    a about above absolutely across actually add after again against ago agree air all allow
+    almost alone along already also although always am amazing among an and angry animal
+    annoying another answer any anyone anything app apple apps are area arm around arrive art
+    as ask at attention available average awesome away awful baby back bad bag bank battery
+    be beautiful because become bed been before began begin behind being believe below best
+    better between big bill black blue boat body book boring both box boy brand break bring
+    broke broken brother brought build building bus business busy but buy by call came camera
+    can cannot car card care carry case cause change charge charger cheap check child
+    children choose city class clean clear close clothes coffee cold college color come
+    comfortable company complete computer confused consider contact continue cool cost could
+    country couple course cover create crazy customer cut dark data date daughter day days
+    dead deal dear decide delivered delivery design did die different difficult dinner
+    disappointed disappointing do doctor does doing dog dollar done door down dream drink
+    drive drop during each early earth easy eat economy education effect either else email
+    end enjoy enough entire even evening ever every everyone everything exactly example
+    excellent except excited expect expensive experience explain eye face fact fail failed
+    fair fake fall family fan fantastic far fast father fault favorite feature features feel
+    feeling few field fight figure fill film final finally find fine finish first fit five
+    fix flight floor follow food foot for force forget form forward found four free fresh
+    friend friends from front full fun funny future game garden gave get getting girl give
+    given glad go goes going gone good got government great green ground group grow guess guy
+    guys had hair half hand happen happened happy hard has hate have having he head health
+    hear heard heart heavy hello help her here high him himself his history hit hold holiday
+    home hope horrible hospital hot hotel hour hours house how however huge human hundred
+    hungry hurt husband i idea if ignore important improve in include including increase
+    indeed information inside instead interest interesting internet into is issue issues it
+    item its itself job join just keep kept key kid kids kind kitchen knew know known lack
+    laptop large last late later laugh law lay lead learn least leave left less let letter
+    level life light like likely line list listen little live living local long look looking
+    lose lost lot lots love lovely low lunch machine made main major make making man manager
+    many market marriage matter may maybe me mean meant meet meeting member men message met
+    might mind minute minutes miss mobile model moment money month months more morning most
+    mother move movie movies much music must my myself name nation near nearly need needed
+    never new news next nice night nine no nobody none nor not note nothing notice now number
+    of off offer office often oh ok okay old on once one online only open opinion option or
+    order other others our out outside over own owner page paid pain paper parents park part
+    particular party pass past pay payment people perfect perhaps person personal phone pick
+    picture piece place plan play player please point police policy poor popular possible
+    power prefer present president pretty price probably problem problems process product
+    products program promise proper provide public pull purchase purchased put quality
+    question quick quickly quite rain raise ran rather reach read ready real really reason
+    receive received recent recommend red refund refused relationship remember remove repair
+    replace replacement reply report rest restaurant result return review reviews rich ride
+    right road room rude rule run sad safe said sale same saturday save saw say says scared
+    school screen sea season seat second see seem seen sell send sense sent series serious
+    service set seven several shall share she ship shipping shirt shop short should show
+    shows side sign simple since sing sister sit site situation six size sleep slow small
+    smart smell smile so social software some someone something sometimes son song soon sorry
+    sound space speak special speed spend spent sport staff stand standard star start started
+    state station stay still stop store story street strong student study stuff stupid style
+    subject such suddenly summer sunday support suppose sure surprised sweet system table
+    take taken talk taste tea teach teacher team tell ten terrible test than thank thanks
+    that the their them themselves then there therefore these they thing things think third
+    this those though thought thousand three through throughout thus ticket till time tired
+    to today together told tomorrow tonight too took top total touch towards town traffic
+    train travel tried trip trouble true trust truth try trying turn twenty two type under
+    understand unfortunately unit university until up update upon us use used useful useless
+    user using usually value very video view village visit voice wait waiting walk wall want
+    wanted war warm warranty was wash waste watch water way we wear weather website week
+    weekend weeks weight welcome well went were west what whatever when where whether which
+    while white who whole why wife will win window wish with within without woman women won
+    wonder wonderful word work worked working world worried worse worst worth would write
+    writing wrong yeah year years yellow yes yesterday yet you young your yourself
+    don't doesn't didn't won't can't isn't aren't wasn't weren't haven't hasn't hadn't
+    wouldn't shouldn't couldn't i'm i've i'll i'd you're you've you'll they're they've
+    we're we've it's that's there's what's let's he's she's
+    """.split()
+)