feat: estado e cidade do filtro e da árvore também a partir de coverage e outras grafias
This commit is contained in:
+104
-19
@@ -1448,36 +1448,106 @@ const UF = { AC: 'Acre', AL: 'Alagoas', AP: 'Amapá', AM: 'Amazonas', BA: 'Bahia
|
||||
RS: 'Rio Grande do Sul', RO: 'Rondônia', RR: 'Roraima', SC: 'Santa Catarina', SP: 'São Paulo', SE: 'Sergipe', TO: 'Tocantins' };
|
||||
const STATE_BY_KEY = new Map(Object.values(UF).map(n => [deburr(n), n]));
|
||||
|
||||
// ── places, from subject AND coverage, in the many ways the catalogue writes them:
|
||||
// "Porto Velho (RO)" · "Recife, PE" · "Rio de Janeiro - RJ" · "Recife, Pernambuco, Brasil"
|
||||
// "Bahia (Estado)" · "RJ" · "Centro (Rio de Janeiro, RJ)" · bare "Recife" (learned below)
|
||||
const UF_CODE = /^[A-Z]{2}$/;
|
||||
const CITY_UF = new Map();
|
||||
const CANON = new Map(); // deburred city → its most common spelling ("Rio de janeiro" → "Rio de Janeiro")
|
||||
const HOOD = new Map(); // deburred bairro → {state, city}, learned from "Bairro (Cidade, UF)"
|
||||
const FEATURE = /^(?:ba[ií]a|morro|ilha|serra|lago|lagoa|floresta|estrada|avenida|rua|pra[cç]a|largo|parque|jardim|campo|cais|porto|rio|igreja|esta[cç][aã]o|forte|palacio|palácio|centro)\b/i; // deburred city → its UF, learned from the explicit forms
|
||||
const stateOf = x => UF_CODE.test(x) ? UF[x] : STATE_BY_KEY.get(deburr(x));
|
||||
function parsePlaceExplicit(raw) {
|
||||
const t = raw.trim().replace(/\.$/, '').replace(/,\s*(?:brasil|brazil)$/i, '').replace(/^\d{4}(?:-\d{4})?\s*:\s*/, '');
|
||||
let m;
|
||||
if ((m = t.match(/^(.+?)\s*\((?:estado|brazil\s*:\s*state)\)$/i)) && stateOf(m[1].trim())) return { state: stateOf(m[1].trim()) };
|
||||
if (UF_CODE.test(t) && UF[t]) return { state: UF[t] };
|
||||
if (STATE_BY_KEY.has(deburr(t))) return { state: STATE_BY_KEY.get(deburr(t)), city: /^(rio de janeiro|sao paulo)$/.test(deburr(t)) ? t : null };
|
||||
// "Bairro (Cidade, UF)"
|
||||
if ((m = t.match(/^(.+?)\s*\(([^(),]+),\s*([A-Za-z]{2})\)$/)) && UF[m[3].toUpperCase()] && !/\s-\s/.test(m[1]))
|
||||
return { state: UF[m[3].toUpperCase()], city: m[2].replace(/^\d{4}(?:-\d{4})?\s*:\s*/, '').trim(), sub: m[1].trim() };
|
||||
if (/\s-\s/.test(t.replace(/\s+-\s+(?=[A-Z]{2}$)/, ''))) return null; // a topic path, handled by the caller
|
||||
// "Cidade (UF)" · "Cidade, UF" · "Cidade - UF" · "Cidade(UF)"
|
||||
if ((m = t.match(/^(.+?)\s*(?:\(|,\s*|\s-\s*)([A-Z]{2})\)?$/)) && UF[m[2]]) return { state: UF[m[2]], city: m[1].trim() };
|
||||
// "Cidade (Estado por extenso)" · "Cidade, Estado por extenso"
|
||||
if ((m = t.match(/^(.+?)\s*(?:\(|,\s*)([^(),]+)\)?$/)) && STATE_BY_KEY.has(deburr(m[2]))) return { state: STATE_BY_KEY.get(deburr(m[2])), city: m[1].trim() };
|
||||
return null;
|
||||
}
|
||||
function parsePlace(raw) {
|
||||
const p = parsePlaceExplicit(raw);
|
||||
if (p) {
|
||||
if (p.city) p.city = CANON.get(deburr(p.city)) || p.city;
|
||||
// "Ceará, CE": the state written twice is no city (Rio de Janeiro and São Paulo are both)
|
||||
if (p.city && deburr(p.city) === deburr(p.state) && !/^(rio de janeiro|sao paulo)$/.test(deburr(p.state))) p.city = null;
|
||||
return p;
|
||||
}
|
||||
const k = deburr(raw.trim().replace(/\.$/, ''));
|
||||
const uf = CITY_UF.get(k);
|
||||
return uf ? { state: UF[uf], city: raw.trim().replace(/\.$/, '') } : null;
|
||||
}
|
||||
const NO_PLACE = /^\[?s\.?\s?l\.?\]?$|^\[?s\.?\s?n\.?\]?$/i;
|
||||
|
||||
// every place-bearing tag of a piece, PT side only
|
||||
function geoTags(meta) {
|
||||
const seen = new Set(), out = [];
|
||||
const add = t => {
|
||||
const k = subjectKey(t);
|
||||
if (!k || seen.has(k) || NO_PLACE.test(String(t).trim())) return;
|
||||
seen.add(k); out.push(String(t).trim());
|
||||
};
|
||||
subjectsOf(meta).forEach(add);
|
||||
const cov = [].concat((meta || {}).coverage || []).map(t => String(t).trim()).filter(t => !EN_WORD.test(t));
|
||||
cov.forEach(add);
|
||||
return out;
|
||||
}
|
||||
|
||||
function subjectTree(meta) {
|
||||
const root = { kids: new Map() };
|
||||
const put = path => path.reduce((n, label) => {
|
||||
if (!n.kids.has(label)) n.kids.set(label, { kids: new Map(), place: false });
|
||||
return n.kids.get(label);
|
||||
}, root);
|
||||
const place = path => path.forEach((_, i) => { put(path.slice(0, i + 1)).place = true; });
|
||||
const others = [];
|
||||
let hasPlace = false;
|
||||
subjectsOf(meta).forEach(tag => {
|
||||
const key = deburr(tag).replace(/\s*\(estado\)$/, '');
|
||||
if (key === 'brasil') { put(['Brasil']).place = true; hasPlace = true; return; }
|
||||
if (STATE_BY_KEY.has(key)) {
|
||||
const st = STATE_BY_KEY.get(key);
|
||||
put(['Brasil']).place = true; put(['Brasil', st]).place = true; hasPlace = true; return;
|
||||
const tags = geoTags(meta);
|
||||
// a bairro / feature ("Centro", "Baía de Guanabara") is not a city: it hangs under the
|
||||
// piece's real city, or under the city it was learned to belong to
|
||||
const cities = new Set();
|
||||
tags.forEach(tag => { const p = parsePlace(tag); if (p && p.city && !FEATURE.test(p.city) && !HOOD.has(deburr(p.city))) cities.add(p.state + '|' + p.city); });
|
||||
tags.forEach(tag => {
|
||||
const key = deburr(tag).replace(/\.$/, '');
|
||||
if (key === 'brasil' || key === 'brazil') { place(['Brasil']); hasPlace = true; return; }
|
||||
let p = parsePlace(tag);
|
||||
if (p && p.city && !p.sub && (FEATURE.test(p.city) || HOOD.has(deburr(p.city)))) {
|
||||
const h = HOOD.get(deburr(p.city));
|
||||
const real = h ? { state: h.state, city: h.city } : [...cities].map(c => c.split('|')).filter(c => c[0] === p.state).map(c => ({ state: c[0], city: c[1] }))[0];
|
||||
p = real ? { state: real.state, city: real.city, sub: p.city } : { state: p.state, city: null, sub: p.city };
|
||||
}
|
||||
const parts = tag.split(/\s+-\s+/);
|
||||
// "Espírito Santo (Estado) - Colônias": a state carrying a topic
|
||||
const sm = parts[0].match(/^(.+?)\s*\(Estado\)$/i);
|
||||
if (sm && STATE_BY_KEY.has(deburr(sm[1]))) {
|
||||
const path = ['Brasil', STATE_BY_KEY.get(deburr(sm[1]))];
|
||||
path.forEach((_, i) => { put(path.slice(0, i + 1)).place = true; });
|
||||
parts.slice(1).forEach((_, i) => put(path.concat(parts.slice(1, i + 2))));
|
||||
if (p) {
|
||||
const path = ['Brasil', p.state].concat(p.city ? [p.city] : []);
|
||||
place(path);
|
||||
if (p.sub) put(path.concat(p.sub)).place = true;
|
||||
hasPlace = true; return;
|
||||
}
|
||||
const m = parts[parts.length - 1].match(/^(.+?)\s*\(([A-Z]{2})\)$/);
|
||||
if (!m || !UF[m[2]]) { others.push(tag); return; }
|
||||
const path = ['Brasil', UF[m[2]], m[1]];
|
||||
path.slice(0, 3).forEach((_, i) => { put(path.slice(0, i + 1)).place = true; });
|
||||
parts.slice(0, -1).forEach((_, i) => { put(path.concat(parts.slice(0, i + 1))); });
|
||||
hasPlace = true;
|
||||
// "Espírito Santo (Estado) - Colônias": a state carrying a topic
|
||||
const parts = tag.split(/\s+-\s+/);
|
||||
const sm = parts[0].match(/^(.+?)\s*\(Estado\)$/i);
|
||||
if (parts.length > 1 && sm && STATE_BY_KEY.has(deburr(sm[1]))) {
|
||||
const path = ['Brasil', STATE_BY_KEY.get(deburr(sm[1]))];
|
||||
place(path); hasPlace = true;
|
||||
parts.slice(1).forEach((_, i) => put(path.concat(parts.slice(1, i + 2))));
|
||||
return;
|
||||
}
|
||||
// "Tópico - Cidade (UF)": the last segment carries the place
|
||||
const lp = parts.length > 1 ? parsePlace(parts[parts.length - 1]) : null;
|
||||
if (lp && lp.city) {
|
||||
const path = ['Brasil', lp.state, lp.city];
|
||||
place(path); hasPlace = true;
|
||||
parts.slice(0, -1).forEach((_, i) => put(path.concat(parts.slice(0, i + 1))));
|
||||
return;
|
||||
}
|
||||
others.push(tag);
|
||||
});
|
||||
// a lone "Brasil" is no tree — it reads better as a chip with the rest
|
||||
if (hasPlace && root.kids.get('Brasil').kids.size === 0) { root.kids.clear(); others.push('Brasil'); }
|
||||
@@ -1532,8 +1602,23 @@ function subjectFacets(meta) {
|
||||
return { st, ct, tp };
|
||||
}
|
||||
|
||||
function learnCities(entries) {
|
||||
const votes = new Map(), spell = new Map();
|
||||
entries.forEach(e => geoTags(e.meta).forEach(t => {
|
||||
const p = parsePlaceExplicit(t);
|
||||
if (!p || !p.city) return;
|
||||
{ const k = deburr(p.city), v = spell.get(k) || new Map(); v.set(p.city, (v.get(p.city) || 0) + 1); spell.set(k, v); }
|
||||
if (p.sub) HOOD.set(deburr(p.sub), { state: p.state, city: p.city });
|
||||
const uf = Object.keys(UF).find(k => UF[k] === p.state), k = deburr(p.city);
|
||||
const v = votes.get(k) || new Map(); v.set(uf, (v.get(uf) || 0) + 1); votes.set(k, v);
|
||||
}));
|
||||
spell.forEach((v, k) => CANON.set(k, [...v].sort((a, b) => b[1] - a[1])[0][0]));
|
||||
votes.forEach((v, k) => { const [uf, n] = [...v].sort((a, b) => b[1] - a[1])[0]; if (n >= 2 && !STATE_BY_KEY.has(k)) CITY_UF.set(k, uf); });
|
||||
}
|
||||
|
||||
function setupFilters(mapData) {
|
||||
const entries = Object.values(mapData.map);
|
||||
learnCities(entries);
|
||||
const instCounts = new Map(), typeCounts = new Map(), authorGroups = new Map();
|
||||
let minYear = Infinity, maxYear = -Infinity;
|
||||
|
||||
|
||||
Reference in New Issue
Block a user