pynysiis 1.0.6

NYSIIS phonetic encoding algorithm.

The pynysiis package provides a Python implementation of the New York State Identification and Intelligence System (NYSIIS) phonetic encoding algorithm. NYSIIS encodes names based on pronunciation, which is helpful in name-matching and searching applications.

Requirements

Python 2.7 and later.

Setup

You can install this package by using the pip tool and installing:

$ pip install pynysiis

Or:

$ easy_install pynysiis

Basic Usage

from nysiis import NYSIIS

encoder = NYSIIS()
name = "Watkins"
encoded_name = encoder.encode(name)
print(encoded_name)  # Output: WATCAN

Name Comparison

from nysiis import NYSIIS

encoder = NYSIIS()

# Compare similar names
name1 = "John Smith"
name2 = "John Smyth"

encoded_name1 = encoder.encode(name1)
encoded_name2 = encoder.encode(name2)

if encoded_name1 == encoded_name2:
    print("Names match phonetically")
else:
    print("Names are phonetically different")

# Output: Names match phonetically

Multi-Language Support

The NYSIIS encoder handles names from various languages:

from nysiis import NYSIIS

encoder = NYSIIS()

# Sample names from different languages
names = [
    # English names
    "Watkins",
    "Robert Johnson",

    # Yoruba name
    "Olanrewaju Akinyele",

    # Igbo name
    "Obinwanne Obiora",

    # Hausa name
    "Abdussalamu Abubakar",

    # Hindi name
    "Virat Kohli",

    # Urdu name
    "Usman Shah"
]

# Process each name
for name in names:
    encoded_name = encoder.encode(name)
    print(f"{name:<20} -> {encoded_name}")

# Output:
# Watkins              -> WATCAN
# Robert Johnson       -> RABART
# Olanrewaju Akinyele -> OLANRA
# Obinwanne Obiora    -> OBAWAN
# Abdussalamu Abubakar-> ABDASA
# Virat Kohli         -> VARATC
# Usman Shah          -> USNANS

Common Use Cases

Database Search Optimisation

def find_similar_names(search_name, database_names):
    encoder = NYSIIS()
    search_code = encoder.encode(search_name)

    matches = [
        name for name in database_names
        if encoder.encode(name) == search_code
    ]
    return matches

Name Deduplication

def find_duplicates(names):
    encoder = NYSIIS()
    encoded_names = {}

    for name in names:
        code = encoder.encode(name)
        encoded_names.setdefault(code, []).append(name)

    return {
        code: names
        for code, names in encoded_names.items()
        if len(names) > 1
    }

Fuzzy Name Matching

def match_names(name1, name2, encoder=None):
    if encoder is None:
        encoder = NYSIIS()

    return encoder.encode(name1) == encoder.encode(name2)

Best Practices

Reuse the Encoder Instance

# Good - create once, use many times
encoder = NYSIIS()
for name in large_name_list:
    encoded = encoder.encode(name)

# Less efficient - creating new instance repeatedly
for name in large_name_list:
    encoded = NYSIIS().encode(name)

Handle Empty Inputs

def process_name(name):
    if not name or not name.strip():
        return None

    encoder = NYSIIS()
    return encoder.encode(name)

Case Sensitivity

# The encoder handles case automatically
encoder = NYSIIS()
print(encoder.encode("smith"))  # Same as "SMITH"
print(encoder.encode("SMITH"))  # Same result

Reference

@inproceedings{Rajkovic2007,
  author    = {Petar Rajkovic and Dragan Jankovic},
  title     = {Adaptation and Application of Daitch-Mokotoff Soundex Algorithm on Serbian Names},
  booktitle = {XVII Conference on Applied Mathematics},
  editors   = {D. Herceg and H. Zarin},
  pages     = {193--204},
  year      = {2007},
  publisher = {Department of Mathematics and Informatics, Novi Sad},
  url       = {https://jmp.sh/hukNujCG}
}

Additional References

• Commission Implementing Regulation (EU) 2016/480

• Commission Implementing Regulation (EU) 2023/2381

License

This project is licensed under the MIT License.

Copyright

Copyright © 2024 Finbarrs Oketunji. All Rights Reserved.