Country Explorer

import marimo as mo
import pandas as pd

from novelentitymatcher import Matcher

mo.md(
    """
# Zero-Shot Country Matching Explorer

Type a country name (with typos, aliases, or alternate languages) and see how
the Matcher resolves it to a canonical entity — no training required.
"""
)

entities = [
    {"id": "DE", "name": "Germany", "aliases": ["Deutschland", "Bundesrepublik"]},
    {"id": "FR", "name": "France", "aliases": ["Frankreich", "La France"]},
    {"id": "US", "name": "United States", "aliases": ["USA", "America"]},
    {"id": "JP", "name": "Japan", "aliases": ["Nihon", "Nippon"]},
    {"id": "CN", "name": "China", "aliases": ["Zhongguo"]},
    {"id": "BR", "name": "Brazil", "aliases": ["Brasil"]},
    {"id": "IN", "name": "India", "aliases": ["Bharat"]},
    {"id": "GB", "name": "United Kingdom", "aliases": ["UK", "Britain", "England"]},
    {"id": "AU", "name": "Australia", "aliases": ["Oz"]},
    {"id": "CA", "name": "Canada", "aliases": ["Canuck Land"]},
]

matcher = Matcher(entities=entities, mode="zero-shot")
matcher.fit()

query_input = mo.ui.text(
    value="Deutchland",
    label="Enter a country name (try typos!):",
    full_width=True,
)
top_k_slider = mo.ui.slider(1, 5, value=3, label="Top-K results")
mo.hstack([query_input, top_k_slider], justify="start")

_query = query_input.value
_k = top_k_slider.value

_empty = not _query.strip()

if _empty:
    table_output = mo.md("Type something above to search!")
else:
    _results = matcher.match(_query, top_k=_k)
    if isinstance(_results, list):
        _df = pd.DataFrame(_results)
    else:
        _df = pd.DataFrame([_results])
    table_output = mo.ui.table(_df, label=f"Matches for '{_query}'")

table_output

_query = query_input.value
if _query.strip():
    _result = matcher.match(_query)
    _entry = _result if isinstance(_result, dict) else _result
    _best_id = _entry.get("id", "N/A") if isinstance(_entry, dict) else "N/A"
    _best_score = _entry.get("score", 0) if isinstance(_entry, dict) else 0
    summary = mo.md(
        f"**Best match:** `{_best_id}` with confidence **{_best_score:.2%}**"
    )
else:
    summary = mo.md("")

summary