Named Entity Recognition (NER)

The NER module in Shekar provides a fast and quantized Named Entity Recognition pipeline powered by a fine-tuned ALBERT model (default) exported to ONNX format for efficient inference.

It automatically identifies common Persian entities such as persons, locations, organizations, dates, and events. The NER pipeline is designed for speed and easy integration with other preprocessing components like normalization and tokenization.

Example usage:

from shekar import NER
from shekar import Normalizer

input_text = (
    "شاهرخ مسکوب به سالِ ۱۳۰۴ در بابل زاده شد و دوره ابتدایی را در تهران و در مدرسه علمیه پشت "
    "مسجد سپهسالار گذراند. از کلاس پنجم ابتدایی مطالعه رمان و آثار ادبی را شروع کرد. از همان زمان "
    "در دبیرستان ادب اصفهان ادامه تحصیل داد. پس از پایان تحصیلات دبیرستان در سال ۱۳۲۴ از اصفهان به تهران رفت و "
    "در رشته حقوق دانشگاه تهران مشغول به تحصیل شد."
)

normalizer = Normalizer()
normalized_text = normalizer(input_text)

albert_ner = NER()
entities = albert_ner(normalized_text)

for text, label in entities:
    print(f"{text} → {label}")

شاهرخ مسکوب → PER
سال ۱۳۰۴ → DAT
بابل → LOC
دوره ابتدایی → DAT
تهران → LOC
مدرسه علمیه → LOC
مسجد سپهسالار → LOC
دبیرستان ادب اصفهان → LOC
در سال ۱۳۲۴ → DAT
اصفهان → LOC
تهران → LOC
دانشگاه تهران → ORG
فرانسه → LOC

Entity Tags

The following table summarizes the entity types used by the model (aggregating B- and I- tags):

Tag	Description
PER	Person names
LOC	Locations (cities, countries, landmarks)
ORG	Organizations (companies, institutions)
DAT	Dates and temporal expressions
EVE	Events (festivals, historical events)
O	Outside (non-entity text)

Chaining with Pipelines

You can seamlessly chain NER with other components using the | operator:

from shekar import NER
from shekar import Normalizer

normalizer = Normalizer()
albert_ner = NER()

ner_pipeline = normalizer | albert_ner
entities = ner_pipeline(input_text)

for text, label in entities:
    print(f"{text} → {label}")

This chaining enables clean and readable code, letting you build custom NLP flows with preprocessing and tagging in one pass.