Update UVW 2026 dataset

README.md

@@ -5,40 +5,123 @@ license: cc-by-sa-4.0
 task_categories:
 - text-generation
 - fill-mask
+- text-classification
+- feature-extraction
+- sentence-similarity
 tags:
 - wikipedia
 - vietnamese
 - nlp
 - underthesea
+- wikidata
+- pretraining
+- language-modeling
+pretty_name: UVW 2026 - Vietnamese Wikipedia Dataset
 size_categories:
 - 1M<n<10M
+source_datasets:
+- original
+dataset_info:
+  features:
+  - name: id
+    dtype: string
+  - name: title
+    dtype: string
+  - name: content
+    dtype: string
+  - name: num_chars
+    dtype: int32
+  - name: num_sentences
+    dtype: int32
+  - name: quality_score
+    dtype: int32
+  - name: wikidata_id
+    dtype: string
+  - name: main_category
+    dtype: string
+  splits:
+  - name: train
+    num_examples: 894579
+  - name: validation
+    num_examples: 111822
+  - name: test
+    num_examples: 111823
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: train.parquet
+  - split: validation
+    path: validation.parquet
+  - split: test
+    path: test.parquet
 ---
 
 # UVW 2026: Underthesea Vietnamese Wikipedia Dataset
 
+<div align="center">
+
+[![License: CC BY-SA 4.0](https://img.shields.io/badge/License-CC%20BY--SA%204.0-lightgrey.svg)](https://creativecommons.org/licenses/by-sa/4.0/)
+[![Language: Vietnamese](https://img.shields.io/badge/Language-Vietnamese-blue.svg)](https://vi.wikipedia.org)
+[![Wikidata Enriched](https://img.shields.io/badge/Wikidata-Enriched-green.svg)](https://www.wikidata.org)
+
+</div>
+
 ## Dataset Description
 
-UVW 2026 (Underthesea Vietnamese Wikipedia) is a cleaned and processed dataset of Vietnamese Wikipedia articles, designed for Vietnamese NLP tasks such as language modeling, text generation, and pretraining.
+**UVW 2026** (Underthesea Vietnamese Wikipedia) is a high-quality, cleaned dataset of Vietnamese Wikipedia articles enriched with Wikidata metadata. Designed for Vietnamese NLP research including language modeling, text generation, text classification, named entity recognition, and model pretraining.
+
+### Key Features
+
+- **Clean text**: Wikipedia markup, templates, references, and formatting removed
+- **Wikidata integration**: Articles linked to Wikidata entities with semantic categories
+- **Quality scoring**: Each article scored 1-10 based on content quality metrics
+- **Unicode normalized**: NFC normalization applied for consistent text processing
+- **Ready to use**: Pre-split into train/validation/test sets
 
 ### Dataset Summary
 
-- **Language:** Vietnamese
-- **Source:** Vietnamese Wikipedia (vi.wikipedia.org)
-- **License:** CC BY-SA 4.0
-- **Year:** 2026
+| Property | Value |
+|----------|-------|
+| **Language** | Vietnamese (vi) |
+| **Source** | Vietnamese Wikipedia + Wikidata |
+| **License** | CC BY-SA 4.0 |
+| **Generated** | 2026-01-31 |
+| **Total Articles** | 1,118,224 |
+| **Wikidata Coverage** | 99.4% |
+| **Category Coverage** | 97.0% |
+| **Unique Categories** | 11,549 |
+| **Avg. Characters** | 1,190 |
+| **Avg. Sentences** | 10 |
+
+## Quick Start
 
-### Statistics
+```python
+from datasets import load_dataset
+
+# Load the dataset
+dataset = load_dataset("undertheseanlp/UVW-2026")
 
-| Metric | Value |
-|--------|-------|
-| Total Articles | 1,118,224 |
-| Total Characters | 1,331,021,085 |
-| Total Sentences | 11,772,330 |
-| Avg. Characters/Article | 1,190 |
+# Access splits
+train = dataset["train"]
+validation = dataset["validation"]
+test = dataset["test"]
+
+# View an example
+print(train[0])
+```
 
 ## Dataset Structure
 
-### Data Instances
+### Data Splits
+
+| Split | Examples | Description |
+|-------|----------|-------------|
+| `train` | 894,579 | Training set (80%) |
+| `validation` | 111,822 | Validation set (10%) |
+| `test` | 111,823 | Test set (10%) |
+
+### Schema
 
 ```json
 {
@@ -47,115 +130,183 @@ UVW 2026 (Underthesea Vietnamese Wikipedia) is a cleaned and processed dataset o
   "content": "Việt Nam, tên chính thức là Cộng hòa Xã hội chủ nghĩa Việt Nam...",
   "num_chars": 45000,
   "num_sentences": 500,
-  "quality": 8
+  "quality_score": 9,
+  "wikidata_id": "Q881",
+  "main_category": "quốc gia có chủ quyền"
 }
 ```
 
-### Data Fields
-
-- `id` (string): Article identifier (title with underscores)
-- `title` (string): Article title
-- `content` (string): Cleaned article content
-- `num_chars` (int): Number of characters in content
-- `num_sentences` (int): Estimated number of sentences
-- `quality` (int): Quality score from 1-10 based on article metrics
+### Field Descriptions
 
-### Quality Score
+| Field | Type | Description |
+|-------|------|-------------|
+| `id` | string | Unique article identifier (URL-safe title) |
+| `title` | string | Human-readable article title |
+| `content` | string | Cleaned article text content |
+| `num_chars` | int32 | Character count of content |
+| `num_sentences` | int32 | Estimated sentence count |
+| `quality_score` | int32 | Quality score from 1 (lowest) to 10 (highest) |
+| `wikidata_id` | string | Wikidata Q-identifier (e.g., "Q881" for Vietnam) |
+| `main_category` | string | Primary category from Wikidata P31 (instance of) |
 
-Quality score (1-10) is computed based on Wikipedia quality research:
+## Usage Examples
 
-- **Length score (40%)**: Article comprehensiveness based on character count
-- **Sentence score (30%)**: Content depth based on number of sentences
-- **Density score (30%)**: Readability based on average sentence length
+### Filter High-Quality Articles
 
-| Score | Count | Percentage |
-|-------|-------|------------|
-| 2 | 124,937 | 11.2% |
-| 3 | 496,169 | 44.4% |
-| 4 | 224,161 | 20.0% |
-| 5 | 111,906 | 10.0% |
-| 6 | 121,956 | 10.9% |
-| 7 | 25,178 | 2.3% |
-| 8 | 10,905 | 1.0% |
-| 9 | 2,958 | 0.3% |
-| 10 | 54 | 0.0% |
-
-References:
-- [Wikipedia Language-Agnostic Quality](https://meta.wikimedia.org/wiki/Research:Prioritization_of_Wikipedia_Articles/Language-Agnostic_Quality)
-- [Automatic Quality Assessment of Wikipedia Articles](https://dl.acm.org/doi/10.1145/3625286)
+```python
+# Get articles with quality score >= 7
+high_quality = dataset["train"].filter(lambda x: x["quality_score"] >= 7)
+print(f"High-quality articles: {len(high_quality):,}")
+```
 
-### Data Splits
+### Filter by Category
 
-| Split | Articles |
-|-------|----------|
-| train | 1,118,224 |
+```python
+# Get articles about people
+people = dataset["train"].filter(lambda x: x["main_category"] == "người")
+print(f"Articles about people: {len(people):,}")
+
+# Get articles about locations
+locations = dataset["train"].filter(
+    lambda x: "khu định cư" in (x["main_category"] or "")
+)
+```
 
-## Usage
+### Filter by Wikidata
 
 ```python
-from datasets import load_dataset
+# Get articles with Wikidata links
+with_wikidata = dataset["train"].filter(lambda x: x["wikidata_id"] != "")
 
-# Load from HuggingFace Hub
-dataset = load_dataset("undertheseanlp/UVW-2026")
+# Lookup specific entity
+vietnam = dataset["train"].filter(lambda x: x["wikidata_id"] == "Q881")
+```
 
-# Access data
-train_data = dataset["train"]
+### Use for Language Modeling
 
-# Example usage
-for example in train_data:
-    print(example["title"])
-    print(example["content"][:200])
-    break
-```
+```python
+from transformers import AutoTokenizer
 
-## Dataset Creation
+tokenizer = AutoTokenizer.from_pretrained("vinai/phobert-base")
 
-### Source Data
+def tokenize(examples):
+    return tokenizer(examples["content"], truncation=True, max_length=512)
 
-The dataset is created from the Vietnamese Wikipedia dump, with the following processing steps:
+tokenized = dataset["train"].map(tokenize, batched=True)
+```
 
-1. Download Vietnamese Wikipedia XML dump
-2. Parse XML and extract article content
-3. Remove Wikipedia markup (templates, categories, references, etc.)
-4. Unicode normalization (NFC)
-5. Filter out:
-   - Special pages (Wikipedia:, User:, Template:, etc.)
-   - Redirect pages
-   - Disambiguation pages
-   - Articles with less than 100 characters
+## Quality Score
 
-### Processing Pipeline
+Articles are scored 1-10 based on multiple factors:
 
-```bash
-# 1. Download Wikipedia dump
-python scripts/download_wikipedia.py
+| Component | Weight | Criteria |
+|-----------|--------|----------|
+| **Length** | 40% | Character count (200 - 100,000 optimal) |
+| **Sentences** | 30% | Sentence count (3 - 1,000 optimal) |
+| **Density** | 30% | Avg sentence length (80-150 chars optimal) |
+| **Wikidata bonus** | +0.5 | Has wikidata_id |
+| **Category bonus** | +0.5 | Has main_category |
+| **Markup penalty** | -1 to -3 | Remaining Wikipedia markup |
 
-# 2. Extract and clean articles
-python scripts/extract_articles.py
+### Quality Distribution
 
-# 3. Create train/dev/test splits
-python scripts/create_splits.py
+| Score | Count | Percentage |
+|-------|------:|----------:|
+| 1 | 134 | 0.0% |
+| 2 | 376 | 0.0% |
+| 3 | 28,267 | 2.5% |
+| 4 | 607,081 | 54.3% |
+| 5 | 208,304 | 18.6% |
+| 6 | 134,385 | 12.0% |
+| 7 | 70,345 | 6.3% |
+| 8 | 57,054 | 5.1% |
+| 9 | 9,649 | 0.9% |
+| 10 | 2,629 | 0.2% |
+
+## Top Categories
+
+| Category (Vietnamese) | Count | Percentage |
+|----------------------|------:|----------:|
+| đơn vị phân loại | 618,281 | 55.3% |
+| người | 78,191 | 7.0% |
+| xã của Pháp | 35,635 | 3.2% |
+| khu định cư | 20,276 | 1.8% |
+| village of Turkey | 18,540 | 1.7% |
+| tiểu hành tinh | 17,891 | 1.6% |
+| mahalle | 16,419 | 1.5% |
+| xã của Việt Nam | 7,088 | 0.6% |
+| đô thị của Ý | 6,700 | 0.6% |
+| trang định hướng Wikimedia | 6,202 | 0.6% |
+
+## Data Processing
+
+### Pipeline Steps
+
+1. **Download**: Fetch Vietnamese Wikipedia XML dump from Wikimedia
+2. **Extract**: Parse XML and extract article content
+3. **Clean**: Remove Wikipedia markup (templates, refs, links, tables, categories)
+4. **Normalize**: Apply Unicode NFC normalization
+5. **Score**: Calculate quality metrics for each article
+6. **Enrich**: Add Wikidata IDs and semantic categories via Wikidata API
+7. **Filter**: Remove special pages, redirects, disambiguation, and short articles (<100 chars)
+8. **Split**: Create train/validation/test splits (80/10/10) with seed=42
+
+### Removed Content
+
+- Wikipedia templates (`{{...}}`)
+- References and citations (`<ref>...</ref>`)
+- HTML tags and comments
+- Category links (`[[Thể loại:...]]`)
+- File/image links (`[[Tập tin:...]]`, `[[File:...]]`)
+- Interwiki links
+- Tables (`{| ... |}`)
+- Infoboxes and navigation templates
+
+### Reproduction
 
-# 4. Prepare for HuggingFace
-python scripts/prepare_huggingface.py
+```bash
+git clone https://github.com/undertheseanlp/UVW-2026
+cd UVW-2026
+uv sync --extra huggingface
+
+# Run full pipeline
+uv run python scripts/build_dataset.py
+
+# Or run individual steps
+uv run python scripts/download_wikipedia.py
+uv run python scripts/extract_articles.py
+uv run python scripts/wikipedia_quality_score.py
+uv run python scripts/add_wikidata.py
+uv run python scripts/create_splits.py
+uv run python scripts/prepare_huggingface.py --push
 ```
 
 ## Citation
 
 ```bibtex
-@misc{uvw2026,
-  title={UVW 2026: Underthesea Vietnamese Wikipedia Dataset},
-  author={Underthesea NLP},
-  year={2026},
-  url={https://github.com/undertheseanlp/underthesea}
+@dataset{uvw2026,
+  title = {UVW 2026: Underthesea Vietnamese Wikipedia Dataset},
+  author = {Underthesea NLP},
+  year = {2026},
+  publisher = {Hugging Face},
+  url = {https://huggingface.co/datasets/undertheseanlp/UVW-2026},
+  note = {Vietnamese Wikipedia articles enriched with Wikidata metadata}
 }
 ```
 
 ## Related Resources
 
 - [Underthesea](https://github.com/undertheseanlp/underthesea) - Vietnamese NLP Toolkit
+- [PhoBERT](https://github.com/VinAIResearch/PhoBERT) - Pre-trained language models for Vietnamese
 - [Vietnamese Wikipedia](https://vi.wikipedia.org)
+- [Wikidata](https://www.wikidata.org)
 
 ## License
 
-This dataset is released under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/), following the Wikipedia license.
+This dataset is released under the [Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA 4.0)](https://creativecommons.org/licenses/by-sa/4.0/), consistent with the Wikipedia content license.
+
+---
+
+<div align="center">
+Made with ❤️ by <a href="https://github.com/undertheseanlp">Underthesea NLP</a>
+</div>

test.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60fcbe70fd7c110fae09b34ee9306960f8db3738c5a7ce679753060bd7c4e323
+size 79550587

train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:524374d40fc7b25501c9d3c7420d9d6f41973d24476418523d3f0536a8c955f2
+size 608316204

validation.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3da59890851b2e68698b4fd8e67aef5439dbad8132e1a264d98eaab9936a83f
+size 78047554