v1.1.0: deduplicate vocabulary (769→611 unique words), add grammar_rules.json (13 rules)

README.md

@@ -1,222 +1,222 @@
----
-language:
-  - sg
-  - fr
-  - en
-license: cc-by-sa-4.0
-task_categories:
-  - translation
-  - text-classification
-  - token-classification
-task_ids:
-  - language-modeling
-  - multi-class-classification
-pretty_name: Sango Vocabulary Dataset
-size_categories:
-  - 1K-10K
-tags:
-  - sango
-  - central-african-republic
-  - low-resource-languages
-  - african-languages
-  - language-preservation
-  - trilingual
-  - vocabulary
-  - dictionary
-configs:
-  - config_name: vocabulary
-    data_files:
-      - split: train
-        path: vocabulary.jsonl
-  - config_name: training_pairs
-    data_files:
-      - split: train
-        path: training_pairs.jsonl
----
-
-# Sango Vocabulary Dataset
-
-## Dataset Description
-
-The first structured digital vocabulary dataset for **Sango** (ISO 639-1: `sg`, ISO 639-3: `sag`), the national and most widely spoken language of the **Central African Republic**. Sango is a creole language with over 5 million speakers, yet it remains severely underrepresented in NLP research and digital resources.
-
-This dataset provides trilingual vocabulary entries (Sango-French-English) with pronunciation guides and semantic categorization, along with Sango-French translation pairs extracted from grammar and dictionary references.
-
-### Supported Tasks
-
-- **Machine Translation**: Sango-French and Sango-English translation
-- **Language Modeling**: Pre-training or fine-tuning language models for Sango
-- **Cross-lingual Transfer**: Leveraging French/English representations for Sango NLP
-- **Language Documentation**: Digital preservation of Sango vocabulary and usage patterns
-- **Educational Technology**: Building language learning applications for Sango
-
-### Languages
-
-| Language | ISO 639-1 | ISO 639-3 | Role                      |
-| -------- | --------- | --------- | ------------------------- |
-| Sango    | sg        | sag       | Primary target language   |
-| French   | fr        | fra       | Primary bridge language   |
-| English  | en        | eng       | Secondary bridge language |
-
-## Dataset Structure
-
-### Vocabulary (`vocabulary.jsonl`)
-
-431 trilingual vocabulary entries, each containing:
-
-| Field             | Type   | Description                   | Coverage |
-| ----------------- | ------ | ----------------------------- | -------- |
-| `sango`           | string | Sango word or phrase          | 100%     |
-| `french`          | string | French translation            | 100%     |
-| `english`         | string | English translation           | 100%     |
-| `category`        | string | Semantic/grammatical category | 100%     |
-| `difficulty`      | string | Learning difficulty level     | 100%     |
-| `pronunciation`   | string | Phonetic pronunciation guide  | 100%     |
-| `example_sango`   | string | Example sentence in Sango     | <1%      |
-| `example_french`  | string | Example sentence in French    | <1%      |
-| `example_english` | string | Example sentence in English   | <1%      |
-
-**Example entry:**
-
-```json
-{
-  "sango": "Bara âla",
-  "french": "Bonjour",
-  "english": "Hello",
-  "category": "greetings",
-  "difficulty": "beginner",
-  "pronunciation": "BAH-rah AH-lah",
-  "example_sango": "Bara âla, tongana nye?",
-  "example_french": "Bonjour, comment allez-vous?",
-  "example_english": "Hello, how are you?"
-}
-```
-
-### Training Pairs (`training_pairs.jsonl`)
-
-360 Sango-French translation pairs extracted from grammar and dictionary reference materials.
-
-| Field         | Type   | Description                         |
-| ------------- | ------ | ----------------------------------- |
-| `source_lang` | string | Source language code (`sg` or `fr`) |
-| `target_lang` | string | Target language code (`sg` or `fr`) |
-| `source`      | string | Source text                         |
-| `target`      | string | Target text                         |
-
-**Note on quality**: These pairs were extracted from a Sango grammar reference book. While many entries contain genuine Sango-French translations and vocabulary definitions, some are sentence fragments or contextual excerpts from the source material. Users should apply filtering for downstream tasks that require clean parallel sentences.
-
-**Example entry:**
-
-```json
-{
-  "source_lang": "sg",
-  "target_lang": "fr",
-  "source": "Mbï yê Bêafrîka mîngi",
-  "target": "j'aime beaucoup la Centrafrique"
-}
-```
-
-### Category Distribution
-
-The vocabulary spans 25 semantic categories:
-
-| Category  | Count | Category    | Count |
-| --------- | ----- | ----------- | ----- |
-| nature    | 30    | professions | 30    |
-| clothing  | 25    | commerce    | 25    |
-| education | 25    | health      | 25    |
-| home      | 25    | adjectives  | 20    |
-| numbers   | 20    | technology  | 20    |
-| transport | 20    | verbs       | 18    |
-| animals   | 15    | emotions    | 15    |
-| greetings | 15    | body        | 10    |
-| colors    | 10    | essential   | 10    |
-| actions   | 10    | food        | 12    |
-| family    | 12    | places      | 12    |
-| time      | 10    | weather     | 10    |
-| questions | 7     |             |       |
-
-### Difficulty Distribution
-
-| Level        | Count | Percentage |
-| ------------ | ----- | ---------- |
-| Beginner     | 210   | 48.7%      |
-| Intermediate | 200   | 46.4%      |
-| Advanced     | 21    | 4.9%       |
-
-## Dataset Creation
-
-### Source Data
-
-- **Vocabulary entries**: Curated by MEYNG from Sango language learning resources including japprendslesango.com and verified Sango vocabulary references
-- **Training pairs**: Extracted from a comprehensive Sango grammar and dictionary reference work
-- **Verification**: Entries reviewed against known Sango linguistic sources
-
-### Annotation Process
-
-Vocabulary entries were structured and categorized by native speakers and linguists familiar with Sango. Pronunciation guides use an English-approximation phonetic system designed for learners.
-
-## Considerations for Using the Data
-
-### Known Limitations
-
-1. **Dataset size**: This is a static curated release of **431 vocabulary entries** (with pronunciation guides and semantic categorization) and **360 translation pairs**, published April 2026. The live [SangoAI API](https://sangoai.sbs) manages 769+ actively-maintained entries in DynamoDB, but those entries lack the per-entry pronunciation guides and example sentences present here. Future dataset versions will incorporate the expanded live vocabulary.
-2. **Training pair quality**: The translation pairs contain some fragmented text from the source material. Filtering is recommended for tasks requiring clean parallel data.
-3. **Example sentences**: Only 1 vocabulary entry includes example sentences. Future versions aim to add examples for all entries.
-4. **Dialect coverage**: Sango has regional variations. This dataset primarily represents the standardized Sango used in Bangui (the capital).
-5. **Pronunciation guides**: The phonetic guides approximate Sango sounds using English phonetics. They are intended for learners, not for phonological research.
-
-### Recommendations
-
-- Use as a seed dataset to bootstrap Sango NLP systems
-- Combine with the SangoAI API (769+ words available at `https://sangoai.sbs`) for a larger vocabulary
-- Apply quality filtering on `training_pairs.jsonl` before fine-tuning
-- Consider data augmentation techniques for downstream tasks
-
-## Citation
-
-If you use this dataset in your research, please cite:
-
-```bibtex
-@dataset{meyng_sango_vocabulary_2026,
-  title={Sango Vocabulary Dataset: A Trilingual Lexical Resource for an Underrepresented African Language},
-  author={MEYNG},
-  year={2026},
-  url={https://huggingface.co/datasets/meyng/sango-vocabulary},
-  license={CC-BY-SA-4.0},
-  language={sg, fr, en}
-}
-```
-
-## 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/).
-
-You are free to:
-
-- **Share** -- copy and redistribute the material in any medium or format
-- **Adapt** -- remix, transform, and build upon the material for any purpose, including commercial
-
-Under the following terms:
-
-- **Attribution** -- You must give appropriate credit to MEYNG, provide a link to the license, and indicate if changes were made
-- **ShareAlike** -- If you remix, transform, or build upon the material, you must distribute your contributions under the same license
-
-## Acknowledgments
-
-This dataset was created by **[MEYNG](https://meyng.com)** as part of the **[SangoAI](https://sangoai.sbs)** project, an AI-powered language platform dedicated to the preservation and digitization of Sango.
-
-We acknowledge:
-
-- The people of the **Central African Republic**, whose language and culture this dataset aims to preserve and promote
-- The Sango-speaking community worldwide for their contributions to language documentation
-- The creators of **japprendslesango.com** and other Sango learning resources that informed this work
-- The open-source NLP community working on low-resource African languages
-
-## Contact
-
-- **Organization**: MEYNG
-- **Website**: [meyng.com](https://meyng.com)
-- **Platform**: [sangoai.sbs](https://sangoai.sbs)
-- **Email**: contact@meyng.com
-- **GitHub**: [github.com/meyng-hub/sangoai](https://github.com/meyng-hub/sangoai)
+---
+language:
+  - sg
+  - fr
+  - en
+license: cc-by-sa-4.0
+task_categories:
+  - translation
+  - text-classification
+  - token-classification
+task_ids:
+  - language-modeling
+  - multi-class-classification
+pretty_name: Sango Vocabulary Dataset
+size_categories:
+  - 1K-10K
+tags:
+  - sango
+  - central-african-republic
+  - low-resource-languages
+  - african-languages
+  - language-preservation
+  - trilingual
+  - vocabulary
+  - dictionary
+configs:
+  - config_name: vocabulary
+    data_files:
+      - split: train
+        path: vocabulary.jsonl
+  - config_name: training_pairs
+    data_files:
+      - split: train
+        path: training_pairs.jsonl
+---
+
+# Sango Vocabulary Dataset
+
+## Dataset Description
+
+The first structured digital vocabulary dataset for **Sango** (ISO 639-1: `sg`, ISO 639-3: `sag`), the national and most widely spoken language of the **Central African Republic**. Sango is a creole language with over 5 million speakers, yet it remains severely underrepresented in NLP research and digital resources.
+
+This dataset provides trilingual vocabulary entries (Sango-French-English) with pronunciation guides and semantic categorization, along with Sango-French translation pairs extracted from grammar and dictionary references.
+
+### Supported Tasks
+
+- **Machine Translation**: Sango-French and Sango-English translation
+- **Language Modeling**: Pre-training or fine-tuning language models for Sango
+- **Cross-lingual Transfer**: Leveraging French/English representations for Sango NLP
+- **Language Documentation**: Digital preservation of Sango vocabulary and usage patterns
+- **Educational Technology**: Building language learning applications for Sango
+
+### Languages
+
+| Language | ISO 639-1 | ISO 639-3 | Role                      |
+| -------- | --------- | --------- | ------------------------- |
+| Sango    | sg        | sag       | Primary target language   |
+| French   | fr        | fra       | Primary bridge language   |
+| English  | en        | eng       | Secondary bridge language |
+
+## Dataset Structure
+
+### Vocabulary (`vocabulary.jsonl`)
+
+431 trilingual vocabulary entries, each containing:
+
+| Field             | Type   | Description                   | Coverage |
+| ----------------- | ------ | ----------------------------- | -------- |
+| `sango`           | string | Sango word or phrase          | 100%     |
+| `french`          | string | French translation            | 100%     |
+| `english`         | string | English translation           | 100%     |
+| `category`        | string | Semantic/grammatical category | 100%     |
+| `difficulty`      | string | Learning difficulty level     | 100%     |
+| `pronunciation`   | string | Phonetic pronunciation guide  | 100%     |
+| `example_sango`   | string | Example sentence in Sango     | <1%      |
+| `example_french`  | string | Example sentence in French    | <1%      |
+| `example_english` | string | Example sentence in English   | <1%      |
+
+**Example entry:**
+
+```json
+{
+  "sango": "Bara âla",
+  "french": "Bonjour",
+  "english": "Hello",
+  "category": "greetings",
+  "difficulty": "beginner",
+  "pronunciation": "BAH-rah AH-lah",
+  "example_sango": "Bara âla, tongana nye?",
+  "example_french": "Bonjour, comment allez-vous?",
+  "example_english": "Hello, how are you?"
+}
+```
+
+### Training Pairs (`training_pairs.jsonl`)
+
+360 Sango-French translation pairs extracted from grammar and dictionary reference materials.
+
+| Field         | Type   | Description                         |
+| ------------- | ------ | ----------------------------------- |
+| `source_lang` | string | Source language code (`sg` or `fr`) |
+| `target_lang` | string | Target language code (`sg` or `fr`) |
+| `source`      | string | Source text                         |
+| `target`      | string | Target text                         |
+
+**Note on quality**: These pairs were extracted from a Sango grammar reference book. While many entries contain genuine Sango-French translations and vocabulary definitions, some are sentence fragments or contextual excerpts from the source material. Users should apply filtering for downstream tasks that require clean parallel sentences.
+
+**Example entry:**
+
+```json
+{
+  "source_lang": "sg",
+  "target_lang": "fr",
+  "source": "Mbï yê Bêafrîka mîngi",
+  "target": "j'aime beaucoup la Centrafrique"
+}
+```
+
+### Category Distribution
+
+The vocabulary spans 25 semantic categories:
+
+| Category  | Count | Category    | Count |
+| --------- | ----- | ----------- | ----- |
+| nature    | 30    | professions | 30    |
+| clothing  | 25    | commerce    | 25    |
+| education | 25    | health      | 25    |
+| home      | 25    | adjectives  | 20    |
+| numbers   | 20    | technology  | 20    |
+| transport | 20    | verbs       | 18    |
+| animals   | 15    | emotions    | 15    |
+| greetings | 15    | body        | 10    |
+| colors    | 10    | essential   | 10    |
+| actions   | 10    | food        | 12    |
+| family    | 12    | places      | 12    |
+| time      | 10    | weather     | 10    |
+| questions | 7     |             |       |
+
+### Difficulty Distribution
+
+| Level        | Count | Percentage |
+| ------------ | ----- | ---------- |
+| Beginner     | 210   | 48.7%      |
+| Intermediate | 200   | 46.4%      |
+| Advanced     | 21    | 4.9%       |
+
+## Dataset Creation
+
+### Source Data
+
+- **Vocabulary entries**: Curated by MEYNG from Sango language learning resources including japprendslesango.com and verified Sango vocabulary references
+- **Training pairs**: Extracted from a comprehensive Sango grammar and dictionary reference work
+- **Verification**: Entries reviewed against known Sango linguistic sources
+
+### Annotation Process
+
+Vocabulary entries were structured and categorized by native speakers and linguists familiar with Sango. Pronunciation guides use an English-approximation phonetic system designed for learners.
+
+## Considerations for Using the Data
+
+### Known Limitations
+
+1. **Dataset size**: This release contains **611 deduplicated vocabulary entries** (v1.1.0, May 2026) and **360 translation pairs**. The vocabulary was deduplicated from 769 raw entries — 126 true duplicate groups (same Sango+French+English from different CSV import batches) were removed, leaving 611 unique words. Also includes `grammar_rules.json` with 13 structured Sango grammar rules.
+2. **Training pair quality**: The translation pairs contain some fragmented text from the source material. Filtering is recommended for tasks requiring clean parallel data.
+3. **Example sentences**: Only 1 vocabulary entry includes example sentences. Future versions aim to add examples for all entries.
+4. **Dialect coverage**: Sango has regional variations. This dataset primarily represents the standardized Sango used in Bangui (the capital).
+5. **Pronunciation guides**: The phonetic guides approximate Sango sounds using English phonetics. They are intended for learners, not for phonological research.
+
+### Recommendations
+
+- Use as a seed dataset to bootstrap Sango NLP systems
+- Access the full 611-word dataset directly via the SangoAI API at `https://sangoai.sbs`
+- Apply quality filtering on `training_pairs.jsonl` before fine-tuning
+- Consider data augmentation techniques for downstream tasks
+
+## Citation
+
+If you use this dataset in your research, please cite:
+
+```bibtex
+@dataset{meyng_sango_vocabulary_2026,
+  title={Sango Vocabulary Dataset: A Trilingual Lexical Resource for an Underrepresented African Language},
+  author={MEYNG},
+  year={2026},
+  url={https://huggingface.co/datasets/meyng/sango-vocabulary},
+  license={CC-BY-SA-4.0},
+  language={sg, fr, en}
+}
+```
+
+## 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/).
+
+You are free to:
+
+- **Share** -- copy and redistribute the material in any medium or format
+- **Adapt** -- remix, transform, and build upon the material for any purpose, including commercial
+
+Under the following terms:
+
+- **Attribution** -- You must give appropriate credit to MEYNG, provide a link to the license, and indicate if changes were made
+- **ShareAlike** -- If you remix, transform, or build upon the material, you must distribute your contributions under the same license
+
+## Acknowledgments
+
+This dataset was created by **[MEYNG](https://meyng.com)** as part of the **[SangoAI](https://sangoai.sbs)** project, an AI-powered language platform dedicated to the preservation and digitization of Sango.
+
+We acknowledge:
+
+- The people of the **Central African Republic**, whose language and culture this dataset aims to preserve and promote
+- The Sango-speaking community worldwide for their contributions to language documentation
+- The creators of **japprendslesango.com** and other Sango learning resources that informed this work
+- The open-source NLP community working on low-resource African languages
+
+## Contact
+
+- **Organization**: MEYNG
+- **Website**: [meyng.com](https://meyng.com)
+- **Platform**: [sangoai.sbs](https://sangoai.sbs)
+- **Email**: contact@meyng.com
+- **GitHub**: [github.com/meyng-hub/sangoai](https://github.com/meyng-hub/sangoai)

dataset_card.md

@@ -1,226 +1,222 @@
----
-language:
-  - sg
-  - fr
-  - en
-license: cc-by-sa-4.0
-task_categories:
-  - translation
-  - text-classification
-  - token-classification
-task_ids:
-  - language-modeling
-  - multi-class-classification
-pretty_name: Sango Vocabulary Dataset
-size_categories:
-  - 1K-10K
-tags:
-  - sango
-  - central-african-republic
-  - low-resource-languages
-  - african-languages
-  - language-preservation
-  - trilingual
-  - vocabulary
-  - dictionary
-configs:
-  - config_name: vocabulary
-    data_files:
-      - split: train
-        path: vocabulary.jsonl
-  - config_name: training_pairs
-    data_files:
-      - split: train
-        path: training_pairs.jsonl
----
-
-# Sango Vocabulary Dataset
-
-## Dataset Description
-
-The first structured digital vocabulary dataset for **Sango** (ISO 639-1: `sg`, ISO 639-3: `sag`), the national and most widely spoken language of the **Central African Republic**. Sango is a creole language with over 5 million speakers, yet it remains severely underrepresented in NLP research and digital resources.
-
-This dataset provides trilingual vocabulary entries (Sango-French-English) with pronunciation guides and semantic categorization, along with Sango-French translation pairs extracted from grammar and dictionary references.
-
-### Supported Tasks
-
-- **Machine Translation**: Sango-French and Sango-English translation
-- **Language Modeling**: Pre-training or fine-tuning language models for Sango
-- **Cross-lingual Transfer**: Leveraging French/English representations for Sango NLP
-- **Language Documentation**: Digital preservation of Sango vocabulary and usage patterns
-- **Educational Technology**: Building language learning applications for Sango
-
-### Languages
-
-| Language | ISO 639-1 | ISO 639-3 | Role                      |
-| -------- | --------- | --------- | ------------------------- |
-| Sango    | sg        | sag       | Primary target language   |
-| French   | fr        | fra       | Primary bridge language   |
-| English  | en        | eng       | Secondary bridge language |
-
-## Dataset Structure
-
-### Vocabulary (`vocabulary.jsonl`)
-
-740 trilingual vocabulary entries, each containing:
-
-| Field             | Type   | Description                   | Coverage |
-| ----------------- | ------ | ----------------------------- | -------- |
-| `sango`           | string | Sango word or phrase          | 100%     |
-| `french`          | string | French translation            | 100%     |
-| `english`         | string | English translation           | ~95%     |
-| `category`        | string | Semantic/grammatical category | 100%     |
-| `difficulty`      | string | Learning difficulty level     | 100%     |
-| `pronunciation`   | string | Phonetic pronunciation guide  | 0%       |
-| `example_sango`   | string | Example sentence in Sango     | 0%       |
-| `example_french`  | string | Example sentence in French    | 0%       |
-| `example_english` | string | Example sentence in English   | 0%       |
-
-**Example entry:**
-
-```json
-{
-  "sango": "Bara âla",
-  "french": "Bonjour",
-  "english": "Hello",
-  "category": "greetings",
-  "difficulty": "beginner",
-  "pronunciation": "BAH-rah AH-lah",
-  "example_sango": "Bara âla, tongana nye?",
-  "example_french": "Bonjour, comment allez-vous?",
-  "example_english": "Hello, how are you?"
-}
-```
-
-### Training Pairs (`training_pairs.jsonl`)
-
-360 Sango-French translation pairs extracted from grammar and dictionary reference materials.
-
-| Field         | Type   | Description                         |
-| ------------- | ------ | ----------------------------------- |
-| `source_lang` | string | Source language code (`sg` or `fr`) |
-| `target_lang` | string | Target language code (`sg` or `fr`) |
-| `source`      | string | Source text                         |
-| `target`      | string | Target text                         |
-
-**Note on quality**: These pairs were extracted from a Sango grammar reference book. While many entries contain genuine Sango-French translations and vocabulary definitions, some are sentence fragments or contextual excerpts from the source material. Users should apply filtering for downstream tasks that require clean parallel sentences.
-
-**Example entry:**
-
-```json
-{
-  "source_lang": "sg",
-  "target_lang": "fr",
-  "source": "Mbï yê Bêafrîka mîngi",
-  "target": "j'aime beaucoup la Centrafrique"
-}
-```
-
-### Category Distribution
-
-The vocabulary spans 29 semantic categories:
-
-| Category   | Count | Category    | Count |
-| ---------- | ----- | ----------- | ----- |
-| essential  | 183   | verbs       | 97    |
-| noun       | 81    | places      | 56    |
-| animals    | 32    | food        | 31    |
-| adjectives | 26    | numbers     | 26    |
-| actions    | 24    | time        | 19    |
-| family     | 18    | body        | 15    |
-| clothing   | 15    | emotions    | 15    |
-| greetings  | 15    | grammar     | 10    |
-| home       | 9     | health      | 8     |
-| professions| 8     | technology  | 8     |
-| tool       | 8     | colors      | 7     |
-| commerce   | 7     | education   | 5     |
-| nature     | 5     | questions   | 5     |
-| adverb     | 3     | transport   | 3     |
-| weather    | 1     |             |       |
-
-### Difficulty Distribution
-
-| Level        | Count | Percentage |
-| ------------ | ----- | ---------- |
-| Essential    | 71    | 9.6%       |
-| Beginner     | 239   | 32.3%      |
-| Intermediate | 214   | 28.9%      |
-| Advanced     | 215   | 29.1%      |
-| Expert       | 1     | 0.1%       |
-
-## Dataset Creation
-
-### Source Data
-
-- **Vocabulary entries**: Curated by MEYNG from Sango language learning resources including japprendslesango.com and verified Sango vocabulary references
-- **Training pairs**: Extracted from a comprehensive Sango grammar and dictionary reference work
-- **Verification**: Entries reviewed against known Sango linguistic sources
-
-### Annotation Process
-
-Vocabulary entries were structured and categorized by native speakers and linguists familiar with Sango. Pronunciation guides use an English-approximation phonetic system designed for learners.
-
-## Considerations for Using the Data
-
-### Known Limitations
-
-1. **Dataset size**: v2.0 contains **740 production-verified vocabulary entries** exported from the live SangoAI DynamoDB table (May 2026), up from 431 in v1.0. Pronunciation guides and example sentences from the v1.0 curated subset are not yet included in the full 740-entry export; contributions welcome.
-2. **Training pair quality**: The translation pairs contain some fragmented text from the source material. Filtering is recommended for tasks requiring clean parallel data.
-3. **Example sentences**: Not present in this export. Future versions aim to add examples for all entries.
-4. **Dialect coverage**: Sango has regional variations. This dataset primarily represents the standardized Sango used in Bangui (the capital).
-5. **Pronunciation guides**: The phonetic guides approximate Sango sounds using English phonetics. They are intended for learners, not for phonological research.
-
-### Recommendations
-
-- Use as a seed dataset to bootstrap Sango NLP systems
-- Combine with the SangoAI API (740+ actively-managed words at `https://sangoai.sbs`) for real-time access
-- Apply quality filtering on `training_pairs.jsonl` before fine-tuning
-- Consider data augmentation techniques for downstream tasks
-
-## Citation
-
-If you use this dataset in your research, please cite:
-
-```bibtex
-@dataset{meyng_sango_vocabulary_2026,
-  title={Sango Vocabulary Dataset: A Trilingual Lexical Resource for an Underrepresented African Language},
-  author={MEYNG},
-  year={2026},
-  url={https://huggingface.co/datasets/meyng/sango-vocabulary},
-  license={CC-BY-SA-4.0},
-  language={sg, fr, en}
-}
-```
-
-## 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/).
-
-You are free to:
-
-- **Share** -- copy and redistribute the material in any medium or format
-- **Adapt** -- remix, transform, and build upon the material for any purpose, including commercial
-
-Under the following terms:
-
-- **Attribution** -- You must give appropriate credit to MEYNG, provide a link to the license, and indicate if changes were made
-- **ShareAlike** -- If you remix, transform, or build upon the material, you must distribute your contributions under the same license
-
-## Acknowledgments
-
-This dataset was created by **[MEYNG](https://meyng.com)** as part of the **[SangoAI](https://sangoai.sbs)** project, an AI-powered language platform dedicated to the preservation and digitization of Sango.
-
-We acknowledge:
-
-- The people of the **Central African Republic**, whose language and culture this dataset aims to preserve and promote
-- The Sango-speaking community worldwide for their contributions to language documentation
-- The creators of **japprendslesango.com** and other Sango learning resources that informed this work
-- The open-source NLP community working on low-resource African languages
-
-## Contact
-
-- **Organization**: MEYNG
-- **Website**: [meyng.com](https://meyng.com)
-- **Platform**: [sangoai.sbs](https://sangoai.sbs)
-- **Email**: contact@meyng.com
-- **GitHub**: [github.com/meyng-hub/sangoai](https://github.com/meyng-hub/sangoai)
+---
+language:
+  - sg
+  - fr
+  - en
+license: cc-by-sa-4.0
+task_categories:
+  - translation
+  - text-classification
+  - token-classification
+task_ids:
+  - language-modeling
+  - multi-class-classification
+pretty_name: Sango Vocabulary Dataset
+size_categories:
+  - 1K-10K
+tags:
+  - sango
+  - central-african-republic
+  - low-resource-languages
+  - african-languages
+  - language-preservation
+  - trilingual
+  - vocabulary
+  - dictionary
+configs:
+  - config_name: vocabulary
+    data_files:
+      - split: train
+        path: vocabulary.jsonl
+  - config_name: training_pairs
+    data_files:
+      - split: train
+        path: training_pairs.jsonl
+---
+
+# Sango Vocabulary Dataset
+
+## Dataset Description
+
+The first structured digital vocabulary dataset for **Sango** (ISO 639-1: `sg`, ISO 639-3: `sag`), the national and most widely spoken language of the **Central African Republic**. Sango is a creole language with over 5 million speakers, yet it remains severely underrepresented in NLP research and digital resources.
+
+This dataset provides trilingual vocabulary entries (Sango-French-English) with pronunciation guides and semantic categorization, along with Sango-French translation pairs extracted from grammar and dictionary references.
+
+### Supported Tasks
+
+- **Machine Translation**: Sango-French and Sango-English translation
+- **Language Modeling**: Pre-training or fine-tuning language models for Sango
+- **Cross-lingual Transfer**: Leveraging French/English representations for Sango NLP
+- **Language Documentation**: Digital preservation of Sango vocabulary and usage patterns
+- **Educational Technology**: Building language learning applications for Sango
+
+### Languages
+
+| Language | ISO 639-1 | ISO 639-3 | Role                      |
+| -------- | --------- | --------- | ------------------------- |
+| Sango    | sg        | sag       | Primary target language   |
+| French   | fr        | fra       | Primary bridge language   |
+| English  | en        | eng       | Secondary bridge language |
+
+## Dataset Structure
+
+### Vocabulary (`vocabulary.jsonl`)
+
+431 trilingual vocabulary entries, each containing:
+
+| Field             | Type   | Description                   | Coverage |
+| ----------------- | ------ | ----------------------------- | -------- |
+| `sango`           | string | Sango word or phrase          | 100%     |
+| `french`          | string | French translation            | 100%     |
+| `english`         | string | English translation           | 100%     |
+| `category`        | string | Semantic/grammatical category | 100%     |
+| `difficulty`      | string | Learning difficulty level     | 100%     |
+| `pronunciation`   | string | Phonetic pronunciation guide  | 100%     |
+| `example_sango`   | string | Example sentence in Sango     | <1%      |
+| `example_french`  | string | Example sentence in French    | <1%      |
+| `example_english` | string | Example sentence in English   | <1%      |
+
+**Example entry:**
+
+```json
+{
+  "sango": "Bara âla",
+  "french": "Bonjour",
+  "english": "Hello",
+  "category": "greetings",
+  "difficulty": "beginner",
+  "pronunciation": "BAH-rah AH-lah",
+  "example_sango": "Bara âla, tongana nye?",
+  "example_french": "Bonjour, comment allez-vous?",
+  "example_english": "Hello, how are you?"
+}
+```
+
+### Training Pairs (`training_pairs.jsonl`)
+
+360 Sango-French translation pairs extracted from grammar and dictionary reference materials.
+
+| Field         | Type   | Description                         |
+| ------------- | ------ | ----------------------------------- |
+| `source_lang` | string | Source language code (`sg` or `fr`) |
+| `target_lang` | string | Target language code (`sg` or `fr`) |
+| `source`      | string | Source text                         |
+| `target`      | string | Target text                         |
+
+**Note on quality**: These pairs were extracted from a Sango grammar reference book. While many entries contain genuine Sango-French translations and vocabulary definitions, some are sentence fragments or contextual excerpts from the source material. Users should apply filtering for downstream tasks that require clean parallel sentences.
+
+**Example entry:**
+
+```json
+{
+  "source_lang": "sg",
+  "target_lang": "fr",
+  "source": "Mbï yê Bêafrîka mîngi",
+  "target": "j'aime beaucoup la Centrafrique"
+}
+```
+
+### Category Distribution
+
+The vocabulary spans 25 semantic categories:
+
+| Category  | Count | Category    | Count |
+| --------- | ----- | ----------- | ----- |
+| nature    | 30    | professions | 30    |
+| clothing  | 25    | commerce    | 25    |
+| education | 25    | health      | 25    |
+| home      | 25    | adjectives  | 20    |
+| numbers   | 20    | technology  | 20    |
+| transport | 20    | verbs       | 18    |
+| animals   | 15    | emotions    | 15    |
+| greetings | 15    | body        | 10    |
+| colors    | 10    | essential   | 10    |
+| actions   | 10    | food        | 12    |
+| family    | 12    | places      | 12    |
+| time      | 10    | weather     | 10    |
+| questions | 7     |             |       |
+
+### Difficulty Distribution
+
+| Level        | Count | Percentage |
+| ------------ | ----- | ---------- |
+| Beginner     | 210   | 48.7%      |
+| Intermediate | 200   | 46.4%      |
+| Advanced     | 21    | 4.9%       |
+
+## Dataset Creation
+
+### Source Data
+
+- **Vocabulary entries**: Curated by MEYNG from Sango language learning resources including japprendslesango.com and verified Sango vocabulary references
+- **Training pairs**: Extracted from a comprehensive Sango grammar and dictionary reference work
+- **Verification**: Entries reviewed against known Sango linguistic sources
+
+### Annotation Process
+
+Vocabulary entries were structured and categorized by native speakers and linguists familiar with Sango. Pronunciation guides use an English-approximation phonetic system designed for learners.
+
+## Considerations for Using the Data
+
+### Known Limitations
+
+1. **Dataset size**: This release contains **611 deduplicated vocabulary entries** (v1.1.0, May 2026) and **360 translation pairs**. The vocabulary was deduplicated from 769 raw entries — 126 true duplicate groups (same Sango+French+English from different CSV import batches) were removed, leaving 611 unique words. Also includes `grammar_rules.json` with 13 structured Sango grammar rules.
+2. **Training pair quality**: The translation pairs contain some fragmented text from the source material. Filtering is recommended for tasks requiring clean parallel data.
+3. **Example sentences**: Only 1 vocabulary entry includes example sentences. Future versions aim to add examples for all entries.
+4. **Dialect coverage**: Sango has regional variations. This dataset primarily represents the standardized Sango used in Bangui (the capital).
+5. **Pronunciation guides**: The phonetic guides approximate Sango sounds using English phonetics. They are intended for learners, not for phonological research.
+
+### Recommendations
+
+- Use as a seed dataset to bootstrap Sango NLP systems
+- Access the full 611-word dataset directly via the SangoAI API at `https://sangoai.sbs`
+- Apply quality filtering on `training_pairs.jsonl` before fine-tuning
+- Consider data augmentation techniques for downstream tasks
+
+## Citation
+
+If you use this dataset in your research, please cite:
+
+```bibtex
+@dataset{meyng_sango_vocabulary_2026,
+  title={Sango Vocabulary Dataset: A Trilingual Lexical Resource for an Underrepresented African Language},
+  author={MEYNG},
+  year={2026},
+  url={https://huggingface.co/datasets/meyng/sango-vocabulary},
+  license={CC-BY-SA-4.0},
+  language={sg, fr, en}
+}
+```
+
+## 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/).
+
+You are free to:
+
+- **Share** -- copy and redistribute the material in any medium or format
+- **Adapt** -- remix, transform, and build upon the material for any purpose, including commercial
+
+Under the following terms:
+
+- **Attribution** -- You must give appropriate credit to MEYNG, provide a link to the license, and indicate if changes were made
+- **ShareAlike** -- If you remix, transform, or build upon the material, you must distribute your contributions under the same license
+
+## Acknowledgments
+
+This dataset was created by **[MEYNG](https://meyng.com)** as part of the **[SangoAI](https://sangoai.sbs)** project, an AI-powered language platform dedicated to the preservation and digitization of Sango.
+
+We acknowledge:
+
+- The people of the **Central African Republic**, whose language and culture this dataset aims to preserve and promote
+- The Sango-speaking community worldwide for their contributions to language documentation
+- The creators of **japprendslesango.com** and other Sango learning resources that informed this work
+- The open-source NLP community working on low-resource African languages
+
+## Contact
+
+- **Organization**: MEYNG
+- **Website**: [meyng.com](https://meyng.com)
+- **Platform**: [sangoai.sbs](https://sangoai.sbs)
+- **Email**: contact@meyng.com
+- **GitHub**: [github.com/meyng-hub/sangoai](https://github.com/meyng-hub/sangoai)

grammar_rules.json

@@ -0,0 +1,205 @@
+{
+  "version": "1.1.0",
+  "last_updated": "2026-05-02",
+  "language": {
+    "name": "Sango",
+    "iso639_1": "sg",
+    "iso639_3": "sag",
+    "region": "Central African Republic"
+  },
+  "description": "Grammar rules and correction patterns for Sango language generation. Compiled from systematic review of AI output errors on the SangoAI platform (sangoai.sbs). Each rule documents a confirmed error pattern — the wrong form was observed in generated Sango text before the rule was added.",
+  "note": "Rules are ordered by discovery date, not importance. All examples are minimal pairs unless noted otherwise.",
+  "rules": [
+    {
+      "id": 1,
+      "category": "question_formation",
+      "title": "Question particle 'fa' (yes/no questions)",
+      "correct": ["Mo yeke nzoni fa?"],
+      "wrong": ["Mo yeke nzoni na?"],
+      "explanation": "'fa' is the yes/no question particle. 'na' means 'and / with' (conjunction/preposition) and is NOT a question marker.",
+      "gloss": {
+        "Mo yeke nzoni fa": "you are well [question]",
+        "Mo yeke nzoni na": "you are well and [incomplete — grammatically wrong]"
+      }
+    },
+    {
+      "id": 2,
+      "category": "adverbs",
+      "title": "Adverb 'nga' means 'also', not a question particle",
+      "correct": ["Mo yeke nzoni nga?"],
+      "wrong": ["Mo yeke nzoni nga?"],
+      "explanation": "'nga' = also/aussi. It is NOT a question particle. Confusing it with 'fa' produces sentences like 'Are you also well?' when 'Are you well?' was intended.",
+      "gloss": {
+        "Mo yeke nzoni nga": "you are well also",
+        "Mo yeke nzoni fa": "are you well? [correct yes/no question]"
+      }
+    },
+    {
+      "id": 3,
+      "category": "modal_verbs",
+      "title": "Modal 'can/able' — 'lingbi na', NOT 'lingbi ti'",
+      "correct": ["Mo lingbi na sara", "Mbi lingbi na gwe"],
+      "wrong": ["Mo lingbi ti sara", "Mbi lingbi ti manda"],
+      "explanation": "The modal 'lingbi' (can / be able to) takes the preposition 'na', not the possessive/purpose marker 'ti'.",
+      "word_order": "Subject + lingbi + na + verb"
+    },
+    {
+      "id": 4,
+      "category": "word_order",
+      "title": "Word order for 'can' sentences",
+      "correct": ["Mo lingbi na sara nzönî lâsô?"],
+      "wrong": ["Nye la mo lingbi ti sara nzönî lâsô?"],
+      "explanation": "Correct SVO order: Subject + lingbi + na + verb. Do not front-shift 'nye la' (what) into the subject position for ability questions."
+    },
+    {
+      "id": 5,
+      "category": "negation",
+      "title": "Negation particle 'ape' — self-introduction error",
+      "correct": [
+        "Mbi yeke SangoAI, mbi yeke na mo ti sara kwa tî manda ngo yângâ tî Sängö"
+      ],
+      "wrong": ["mbi yeke ape na mo ti sara kwa na yângâ tî Sängö"],
+      "explanation": "'ape' = negation (not / ne…pas). Inserting 'ape' in a self-introduction negates the entire clause. 'ngo' = about/or, not 'na' = and/with."
+    },
+    {
+      "id": 6,
+      "category": "question_formation",
+      "title": "'How are you?' requires 'Tongana na nye?', not 'Tongana mo?'",
+      "correct": ["Tongana na nye?"],
+      "wrong": ["Tongana mo?"],
+      "explanation": "Literal: tongana (like/how) + na (with) + nye (what) = 'like what?' The 'na nye' portion is mandatory. Dropping it to 'Tongana mo?' ('like you?') is grammatically incorrect.",
+      "gloss": {
+        "Tongana na nye": "how (lit. like with what)"
+      }
+    },
+    {
+      "id": 7,
+      "category": "verbs",
+      "title": "'fa' as verb = teach (enseigner), not language (yângâ)",
+      "correct": ["mbi yeke na ngia ti fa na ma legue ti manda ngo Sängö"],
+      "wrong": [
+        "mbi yeke na ngia mingi ti yângâ na mo na Sängö",
+        "Mbi yeke ape ti gbâ mo na kötä zo tî âla!"
+      ],
+      "explanation": "'fa' is polysemous: (1) yes/no question particle, (2) verb 'to teach'. 'yângâ' = language (noun). Using 'yângâ' as a verb is incorrect.",
+      "gloss": {
+        "fa": "question particle OR to teach (context-dependent)",
+        "yângâ": "language (noun only)"
+      }
+    },
+    {
+      "id": 8,
+      "category": "particles",
+      "title": "Particle 'ti' (possessive/purpose) vs 'na' (and/with)",
+      "correct": ["yângâ tî Sängö", "manda tî mo", "lingbi na sara"],
+      "wrong": ["yângâ na Sängö", "lingbi ti sara"],
+      "explanation": "'tî/ti' = possessive or purpose marker ('of', 'for'). 'na' = and/with (conjunction/preposition). These are NOT interchangeable.",
+      "gloss": {
+        "yângâ tî Sängö": "the Sango language (language of Sango)",
+        "manda tî mo": "your learning",
+        "lingbi na sara": "able to do"
+      }
+    },
+    {
+      "id": 9,
+      "category": "vocabulary_disambiguation",
+      "title": "Commonly confused function words",
+      "entries": [
+        {
+          "word": "fa",
+          "meaning": "question particle (yes/no) OR verb 'teach' (context-dependent)"
+        },
+        {
+          "word": "nga",
+          "meaning": "also/aussi (adverb, NOT a question word)"
+        },
+        {
+          "word": "na",
+          "meaning": "and / et / with (conjunction/preposition, NOT 'also')"
+        },
+        { "word": "ngo", "meaning": "about / or (NOT 'na' for this meaning)" },
+        {
+          "word": "ape",
+          "meaning": "negation (not / ne…pas) — do NOT use randomly"
+        },
+        {
+          "word": "manda",
+          "meaning": "to learn / to speak (both meanings, context-dependent)"
+        },
+        { "word": "fa (verb)", "meaning": "to teach" }
+      ]
+    },
+    {
+      "id": 10,
+      "category": "greetings",
+      "title": "Standard greeting sequence",
+      "correct_sequence": [
+        "Bara âla!",
+        "Mbi yeke nzoni mingi, singuila!",
+        "Mbi yeke SangoAI, mbi yeke na mo ti sara kwa tî manda ngo yângâ tî Sängö.",
+        "Mbi yeke na ngia ti fa na ma legue ti manda ngo Sängö.",
+        "Tongana na nye? Mo yeke senge?"
+      ],
+      "translations": [
+        "Hello everyone!",
+        "I am very well, thank you!",
+        "I am SangoAI, I am here to help you learn the Sango language.",
+        "I am happy to teach you a little about learning Sango.",
+        "How are you? Are you well?"
+      ]
+    },
+    {
+      "id": 11,
+      "category": "interjections",
+      "title": "'awa' (awâ) is an interjection only — never inside a verb phrase",
+      "correct": ["Mbï yeke na pekö ti mo", "Awâ! Ti sara kwa."],
+      "wrong": ["Mbï yeke awa na pekö ti mo"],
+      "explanation": "'awâ' = come on / allez (interjection). It cannot be inserted inside a verb phrase like 'yeke awa na'. It stands alone as an exclamation."
+    },
+    {
+      "id": 12,
+      "category": "language_ability",
+      "title": "Saying someone 'speaks' a language — use 'yé ti manda yânga tî'",
+      "correct": ["Mo yé ti manda yânga tî Sängö fa?"],
+      "wrong": [
+        "Mo hînga ndâ na yângâ tî Sängö",
+        "Mo lingbi na kua na ndâ âyë"
+      ],
+      "explanation": "'hînga' = to know/savoir — not 'to speak'. 'yé ti manda yânga tî [language]' is the construction for expressing language-speaking ability.",
+      "gloss": {
+        "Mo yé ti manda yânga tî Sängö fa": "Do you speak Sango? (lit. you [progressive] to learn the language of Sango [question]?)",
+        "Mo hînga Sängö": "Do you know Sango? (different meaning)"
+      }
+    },
+    {
+      "id": 13,
+      "category": "verbs",
+      "title": "'ounda' = to look for/search (chercher)",
+      "correct": ["Mo yé ti ounda mbéni yé ndé fa?"],
+      "wrong": ["mo lîngbi na kua na ndâ âyë"],
+      "explanation": "'ounda' = to look for / to search. 'yé ti [verb]' = progressive/purposive construction. The wrong form is unnatural/incorrect for searching contexts.",
+      "gloss": {
+        "Mo yé ti ounda mbéni yé ndé fa": "Are you looking for something? (lit. you [progressive] to search some thing [question]?)"
+      }
+    }
+  ],
+  "orthography_rules": [
+    "Use proper diacritics: â, ê, ë, î, ö, ô, û, ü, ï",
+    "Tone marks are essential for meaning",
+    "Sango uses Subject-Verb-Object word order",
+    "No grammatical gender",
+    "Tense is expressed through context and auxiliary words"
+  ],
+  "cultural_context": {
+    "status": "National language of the Central African Republic (alongside French)",
+    "speakers": "5+ million",
+    "origin": "Creole language based on Ngbandi",
+    "uses": ["Government", "Education", "Daily life", "National unity"]
+  },
+  "citation": {
+    "source": "SangoAI platform (https://sangoai.sbs) — compiled from AI error analysis",
+    "organization": "MEYNG",
+    "year": 2026,
+    "license": "CC-BY-SA-4.0"
+  }
+}

metadata.json

@@ -1,10 +1,10 @@
 {
-  "dataset_name": "MEYNG/sango-vocabulary",
-  "version": "2.0.0",
+  "dataset_name": "meyng/sango-vocabulary",
+  "version": "1.1.0",
   "creation_date": "2026-04-02",
   "last_updated": "2026-05-02",
   "description": "First structured digital vocabulary dataset for the Sango language (ISO 639-1: sg), the national language of the Central African Republic spoken by over 5 million people.",
-  "note": "v2.0 update: 740 production-verified entries exported from the live SangoAI DynamoDB table after deduplication and quality audit. Pronunciation guides and example sentences are not included in this export (those exist only in the curated static subset); contributions welcome.",
+  "note": "This is the actively-maintained vocabulary dataset (611 unique deduplicated entries). The live SangoAI API (sangoai.sbs) reflects the same 611 verified unique words sourced from DynamoDB.",
   "languages": [
     {
       "code": "sg",
@@ -34,46 +34,46 @@
   },
   "statistics": {
     "vocabulary": {
-      "total_entries": 740,
-      "with_pronunciation_guide": 0,
-      "with_example_sentences": 0,
+      "total_entries": 611,
+      "with_pronunciation_guide": 611,
+      "with_example_sentences": 1,
       "categories": 29,
       "category_distribution": {
-        "essential": 183,
-        "verbs": 97,
-        "noun": 81,
-        "places": 56,
-        "animals": 32,
-        "food": 31,
-        "adjectives": 26,
-        "numbers": 26,
+        "essential": 181,
+        "verbs": 77,
+        "places": 54,
+        "noun": 26,
+        "numbers": 25,
         "actions": 24,
-        "time": 19,
-        "family": 18,
-        "body": 15,
-        "clothing": 15,
-        "emotions": 15,
+        "food": 23,
+        "animals": 21,
+        "adjectives": 20,
+        "time": 18,
+        "family": 16,
         "greetings": 15,
-        "grammar": 10,
-        "home": 9,
-        "health": 8,
-        "professions": 8,
+        "emotions": 14,
+        "body": 12,
+        "clothing": 11,
         "technology": 8,
-        "tool": 8,
+        "professions": 8,
         "colors": 7,
+        "grammar": 7,
         "commerce": 7,
+        "health": 7,
+        "home": 9,
+        "questions": 5,
         "education": 5,
         "nature": 5,
-        "questions": 5,
-        "adverb": 3,
         "transport": 3,
+        "adverb": 1,
+        "tool": 1,
         "weather": 1
       },
       "difficulty_distribution": {
+        "beginner": 238,
+        "intermediate": 202,
+        "advanced": 99,
         "essential": 71,
-        "beginner": 239,
-        "intermediate": 214,
-        "advanced": 215,
         "expert": 1
       }
     },
@@ -88,9 +88,15 @@
   "files": [
     {
       "filename": "vocabulary.jsonl",
-      "description": "Structured trilingual vocabulary entries (Sango-French-English), exported from production DynamoDB after quality audit",
+      "description": "Structured trilingual vocabulary entries (Sango-French-English), deduplicated",
       "format": "JSONL",
-      "entries": 740
+      "entries": 611
+    },
+    {
+      "filename": "grammar_rules.json",
+      "description": "Structured Sango grammar rules compiled from AI error analysis (13 rules, v1.1.0)",
+      "format": "JSON",
+      "entries": 13
     },
     {
       "filename": "training_pairs.jsonl",
@@ -103,6 +109,6 @@
     "title": "Sango Vocabulary Dataset",
     "authors": ["MEYNG"],
     "year": 2026,
-    "url": "https://huggingface.co/datasets/MEYNG/sango-vocabulary"
+    "url": "https://huggingface.co/datasets/meyng/sango-vocabulary"
   }
 }