Short answer
The CSV contains 37 data rows and 12 columns for order, stable symbol ID, concept IRI, symbol, Unicode notation, Hanyu Pinyin, broad IPA, category, category IRI and example fields. Its adjacent zhuyin-bopomofo-ml-dataset.csv-metadata.json document uses the official CSVW context first, declares a UTF-8 comma dialect with one header row, marks every column required, assigns symbol_id as the unique primary key, and provides datatypes, facets, BCP 47 language tags, a row-subject URI template and RDF property or value mappings. A deterministic ZIP includes the CSV, metadata, README, CC BY 4.0 notice and SHA-256 list. Pinned unmodified official JSON-LD context and Turtle vocabulary snapshots support offline term checks, and the published pair validates with csvw 4.1.0. W3C defines the format but does not certify this package, endorse it or confirm ingestion by any external tool or catalog.
Download the metadata next to the CSV or extract the complete ZIP, verify checksums, then run csvwvalidate against the metadata file. Test the exact downstream processor because CSVW feature support and remote-URL handling vary between tools.
What to look for before choosing
- Complete Unicode U+3105 through U+3129 coverage: 21 initials, 3 medials and 13 finals
- Twelve required columns with integer, string and anyURI datatypes plus explicit facets
- Unique symbol_id primary key, stable SKOS row subjects and RDF property or value mappings
- Standard <CSV URL>-metadata.json discovery name with a relative adjacent CSV URL
- Deterministic ZIP, CC BY 4.0 attribution, byte-exact SHA-256 and no account requirement
A practical decision process
- Download the complete ZIP for offline use or place the metadata file beside the canonical CSV.
- Run csvwvalidate zhuyin-bopomofo-ml-dataset.csv-metadata.json with csvw 4.1.0.
- Confirm that the processor reports 37 rows, 12 columns and no primary-key or datatype errors.
- Inspect the aboutUrl, propertyUrl and valueUrl mappings before generating RDF.
- Verify SHA-256 values and test the target ETL or catalog in staging before publishing.
Quick comparison
| Need | What to check | Why it matters |
|---|---|---|
| Automatic metadata discovery | The metadata uses the exact <CSV filename>-metadata.json convention and a relative url | CSVW-aware tools can locate the table contract without a proprietary descriptor |
| Strict tabular validation | Required typed columns, format facets, a unique primary key and a declared dialect | Pipelines can reject missing, malformed or duplicate values before reuse |
| Linked-data projection | Stable row subjects plus explicit RDF property and IRI value templates | CSV-to-RDF processors receive deterministic mappings instead of guessing semantics |
Sources and resources
- Bopomofo CSVW table and offline bundle
- W3C Model for Tabular Data and Metadata on the Web
- W3C Metadata Vocabulary for Tabular Data
- W3C Generating RDF from Tabular Data on the Web
- IANA application/csvm+json media type registration
Where Lumi Bopomofo fits
Lumi Bopomofo is not required to download, validate, transform or reuse the CSVW table. If currently available, it appears only after the open technical resource as an optional on-device practice layer.
No subscriptionNo adsKid-safe
This page is an independent buying guide. App Store features and prices can change, so confirm details on the listing before purchase.