Skip to main content

Quick Start

This guide walks you through the most common reclink operations.

Compare two strings​

from reclink import jaro_winkler, levenshtein, cosine

jaro_winkler("Jon Smith", "John Smyth")   # 0.832
levenshtein("Jon Smith", "John Smyth")     # 3
cosine("Jon Smith", "John Smyth")          # 0.5

Find the best match​

from reclink import match_best, match_batch

# Best single match
match_best("hello", ["hallo", "world", "help"])
# ("hallo", 0.933, 0)   -> (string, score, index)

# All matches above a threshold
match_batch("hello", ["hallo", "world", "help"], threshold=0.5, limit=2)
# [("hallo", 0.933, 0), ("help", 0.733, 2)]

Batch comparison​

from reclink import cdist

# All-pairs similarity matrix (parallelized across CPU cores)
matrix = cdist(["Jon", "Jane"], ["John", "Janet"], scorer="jaro_winkler")
# array([[0.93, 0.0 ],
#        [0.0 , 0.93]])

Phonetic encoding​

from reclink import soundex, double_metaphone

soundex("Smith")          # "S530"
soundex("Smyth")          # "S530"  -- same code!
double_metaphone("John")  # ("JN", "AN")

Preprocessing​

from reclink import clean_name, fold_case, normalize_whitespace

clean_name("  DR. John  Smith Jr.  ")  # "john smith"
fold_case("HELLO World")               # "hello world"
normalize_whitespace("  too   many   spaces  ")  # "too many spaces"

Scoring presets​

from reclink.presets import name_matching

scorer = name_matching()
scorer.similarity("Jon Smith", "John Smyth")  # 0.89
scorer.match_best("Jon Smith", ["John Smith", "Jane Doe"])
# ("John Smith", 0.96, 0)

Record linkage pipeline​

import pandas as pd
from reclink.pipeline import ReclinkPipeline

df = pd.DataFrame({
    "id": ["1", "2", "3"],
    "first_name": ["Jon", "John", "Jane"],
    "last_name": ["Smith", "Smyth", "Doe"],
})

pipeline = (
    ReclinkPipeline.builder()
    .preprocess("first_name", ["fold_case", "strip_punctuation"])
    .preprocess("last_name", ["fold_case"])
    .block_phonetic("last_name", algorithm="soundex")
    .compare_string("first_name", metric="jaro_winkler")
    .compare_string("last_name", metric="jaro_winkler")
    .classify_threshold(0.85)
    .build()
)

matches = pipeline.dedup(df)
print(matches)
#    left_id right_id     score            scores
# 0       1        2  0.921...  [0.832..., 1.0...]

Next steps​