ner — Named-entity annotations
One entity mention per line (relational pattern: several lines share a document id).
- Dominio
- NLP / data labelling
- Versión
- 1
- Padre
- — (raíz)
- Clave de registros
entities- Separador de lista
,- Archivo
- forks/ner/contract.json
Campos por posición
| # | Tipo | Descripción | Parte |
|---|---|---|---|
doc | str | núcleo | |
start | int | núcleo | |
end | int | núcleo | |
text | str | núcleo | |
type | enum | (valores: PERSON, ORG, LOC, DATE, PRODUCT, CONCEPT, MISC) | núcleo |
conf | float | (max: 1) | núcleo |
Ejemplo válido
ner|n=12|d=20260603|l=en|model=ner-base-2026|schema=ontonotes-lite
d1|0|17|Whisper|PRODUCT|0.98
d1|25|31|OpenAI|ORG|0.99
d1|60|64|2023|DATE|0.97
d2|4|10|Django|PRODUCT|0.96
d2|30|40|PostgreSQL|PRODUCT|0.95
d3|0|4|Lima|LOC|0.99
d3|6|10|Perú|LOC|0.99
d3|45|48|UPC|ORG|0.93
d4|12|30|Item Response Theory|CONCEPT|0.91
d4|35|38|CAT|CONCEPT|0.88
d5|0|12|Adrián Palma|PERSON|0.97
d5|17|31|Erick Palomino|PERSON|0.97
Con el ZIP opcional extraído, valida con mini --forks forks validate forks/ner/fixtures/valid.mini.
Notas de la familia
.mini-ner — Anotaciones de entidades nombradas#
Prefijo: ner Versión: 1 Padre: — Dominio: PLN / etiquetado de datos
Una mención de entidad por línea (patrón relacional: varias líneas comparten un id de documento).
Estructura del registro#
doc:str | start:int[0..] | end:int[0..] | text:str | type:enum{PERSON|ORG|LOC|DATE|PRODUCT|CONCEPT|MISC} | conf:float[0..1]
| # | campo | tipo | notas |
|---|---|---|---|
| 1 | doc |
str |
núcleo |
| 2 | start |
int[0..] |
núcleo |
| 3 | end |
int[0..] |
núcleo |
| 4 | text |
str |
núcleo |
| 5 | type |
enum{PERSON|ORG|LOC|DATE|PRODUCT|CONCEPT|MISC} |
núcleo |
| 6 | conf |
float[0..1] |
núcleo |
Claves de cabecera#
| clave | tipo | requerida | descripción |
|---|---|---|---|
d |
str |
no | fecha de generación AAAAMMDD |
l |
str |
no | código de idioma |
model |
str |
no | |
schema |
str |
no | |
n |
int |
sí | número de registros |
v |
int |
no | versión del contrato |
Ejemplo#
ner|n=12|d=20260603|l=en|model=ner-base-2026|schema=ontonotes-lite
d1|0|17|Whisper|PRODUCT|0.98
d1|25|31|OpenAI|ORG|0.99
…
Fixtures#
fixtures/valid.mini ↔ fixtures/canonical.json (ida y vuelta), fixtures/escaping.mini ↔ fixtures/escaping.json, y los casos negativos bad_count.mini, bad_arity.mini, bad_type.mini.
Bloque de prompt#
Genera la especificación transferible con mini prompt ner.
ner — Named-entity annotations
One entity mention per line (relational pattern: several lines share a document id).
- Domain
- NLP / data labelling
- Version
- 1
- Parent
- — (root)
- Records key
entities- List separator
,- File
- forks/ner/contract.json
Fields by position
| # | Type | Description | Part |
|---|---|---|---|
doc | str | core | |
start | int | core | |
end | int | core | |
text | str | core | |
type | enum | (values: PERSON, ORG, LOC, DATE, PRODUCT, CONCEPT, MISC) | core |
conf | float | (max: 1) | core |
Valid example
ner|n=12|d=20260603|l=en|model=ner-base-2026|schema=ontonotes-lite
d1|0|17|Whisper|PRODUCT|0.98
d1|25|31|OpenAI|ORG|0.99
d1|60|64|2023|DATE|0.97
d2|4|10|Django|PRODUCT|0.96
d2|30|40|PostgreSQL|PRODUCT|0.95
d3|0|4|Lima|LOC|0.99
d3|6|10|Perú|LOC|0.99
d3|45|48|UPC|ORG|0.93
d4|12|30|Item Response Theory|CONCEPT|0.91
d4|35|38|CAT|CONCEPT|0.88
d5|0|12|Adrián Palma|PERSON|0.97
d5|17|31|Erick Palomino|PERSON|0.97
With the optional ZIP extracted, validate with mini --forks forks validate forks/ner/fixtures/valid.mini.
Family notes
.mini-ner — Named-entity annotations#
Prefix: ner Version: 1 Parent: — Domain: NLP / data labelling
One entity mention per line (relational pattern: several lines share a document id).
Record layout#
doc:str | start:int[0..] | end:int[0..] | text:str | type:enum{PERSON|ORG|LOC|DATE|PRODUCT|CONCEPT|MISC} | conf:float[0..1]
| # | field | type | notes |
|---|---|---|---|
| 1 | doc |
str |
core |
| 2 | start |
int[0..] |
core |
| 3 | end |
int[0..] |
core |
| 4 | text |
str |
core |
| 5 | type |
enum{PERSON|ORG|LOC|DATE|PRODUCT|CONCEPT|MISC} |
core |
| 6 | conf |
float[0..1] |
core |
Header keys#
| key | type | required | description |
|---|---|---|---|
d |
str |
no | generation date YYYYMMDD |
l |
str |
no | language code |
model |
str |
no | |
schema |
str |
no | |
n |
int |
yes | number of records |
v |
int |
no | contract version |
Example#
ner|n=12|d=20260603|l=en|model=ner-base-2026|schema=ontonotes-lite
d1|0|17|Whisper|PRODUCT|0.98
d1|25|31|OpenAI|ORG|0.99
…
Fixtures#
fixtures/valid.mini ↔ fixtures/canonical.json (round-trip), fixtures/escaping.mini ↔ fixtures/escaping.json, and the negative cases bad_count.mini, bad_arity.mini, bad_type.mini.
Prompt block#
Generate the transferable specification with mini prompt ner.