@precisa-saude/fhir-ocr-utils 0.38.0 → 0.38.1

package/README.md

@@ -90,9 +90,13 @@ Por isso a correspondência é conservadora:
   `RBC_Urine`, "PH" ancora `pH_Urine`, e "COLOR", "KETONES", "PROTEIN" e
   companhia ancoram mesmo sem valor na linha (texto em colunas). O código do
   sangue não ancora pela mesma linha. A seção acaba na primeira linha com
-  exame de outro painel ou na primeira linha sem valor e sem nome conhecido
-  (o jeito de um cabeçalho); na dúvida ela acaba, e o nome volta ao sentido de
-  fora da seção. Fora dela nada muda.
+  exame de outro painel, no cabeçalho de outro painel ou no fim do texto. Uma
+  linha sem valor e sem nome conhecido pode ser as duas coisas ("COMPREHENSIVE
+  METABOLIC PANEL" ou "MUCUS"), e quem decide é a próxima linha decisiva: se só
+  pode ser da urina (código da urina, a palavra urina, ou /HPF e /LPF), a seção
+  segue; se é exame de outro painel, ou se o texto acaba sem nada decisivo, a
+  seção acaba ali, e o nome volta ao sentido de fora da seção. Fora dela nada
+  muda.
 - **Contexto genético é descartado** — símbolos de gene colidem com nomes de
   biomarcador (o gene `APOB` vs. a lipoproteína `ApoB`). Linhas com acesso
   RefSeq (`NM_000384.2`), notação HGVS (`p.Trp448*`, `c.1234A>G`), `rs` do dbSNP

package/dist/cli.js

@@ -135,6 +135,34 @@ var QUALITATIVE_VALUE_TERMS = /* @__PURE__ */ new Set([
   "undetectable",
   "yellow"
 ]);
+var GENETIC_CONTEXT_PATTERNS = [
+  /\b[nx][mrpc]_\d{6,}/,
+  // RefSeq: NM_000384.2, NP_, NR_, XM_
+  /\bens[gtp]\d{6,}/,
+  // Ensembl: ENSG00000084674
+  /\bp\.[a-z]{3}\d/,
+  // HGVS proteína: p.Trp448*
+  /\bc\.\d+[acgt]?[>_+-]/,
+  // HGVS codificante: c.1234A>G, c.76_78del
+  /\brs\d{4,}\b/,
+  // dbSNP
+  /\bgenes?\b/,
+  /\bvariante?s?\b/,
+  /\bexons?\b/,
+  /\bzygosity\b/,
+  /\bzigosidade\b/,
+  /\balleles?\b/,
+  /\balelos?\b/,
+  /\bmutations?\b/,
+  /\bmutac(ao|oes)\b/,
+  /\bpathogenic/,
+  /\bpatogenic/,
+  /\bheterozyg/,
+  /\bhomozyg/,
+  /\bheterozigot/,
+  /\bhomozigot/,
+  /\bsequence change\b/
+];
 
 // src/body-region.ts
 var WHOLE_BODY_COMPOSITION_CODES = /* @__PURE__ */ new Set([
@@ -283,19 +311,37 @@ function isSectionName(matched) {
   return URINALYSIS_SECTION_NAMES.has(matched) || URINALYSIS_SECTION_NAMES.has(matched.replace(/s$/, ""));
 }
 var MENTIONS_URINE = /(?<![\p{L}\p{N}])urin[ae](?![\p{L}\p{N}])/u;
+var SEDIMENT_FIELD_UNIT = /\/(?:hpf|lpf)(?![\p{L}\p{N}])/u;
+function hasUrineCue(line) {
+  return MENTIONS_URINE.test(line) || SEDIMENT_FIELD_UNIT.test(line);
+}
+function kindOf(line) {
+  const text = line.text.trim();
+  if (!text) return "blank";
+  if (URINALYSIS_HEADER.test(text) && !/\d/.test(text)) return "header";
+  if (line.foreign) return "foreign";
+  if (line.urine) return "urine";
+  return line.known || line.hasValue ? "neutral" : "unknown";
+}
 function urinalysisLineIndexes(lines) {
+  const kinds = lines.map(kindOf);
+  const continuesAsUrine = (from) => {
+    const next = kinds.findIndex(
+      (kind, index) => index > from && (kind === "urine" || kind === "foreign" || kind === "header")
+    );
+    return next !== -1 && kinds[next] === "urine";
+  };
   const inside = /* @__PURE__ */ new Set();
   let open = false;
-  lines.forEach((line, index) => {
-    const text = line.text.trim();
-    if (URINALYSIS_HEADER.test(text) && !/\d/.test(text)) {
+  kinds.forEach((kind, index) => {
+    if (kind === "header") {
       open = true;
       return;
     }
-    if (!open || !text) {
+    if (!open || kind === "blank") {
       return;
     }
-    if (line.foreign || !line.known && !line.hasValue && !MENTIONS_URINE.test(text)) {
+    if (kind === "foreign" || kind === "unknown" && !continuesAsUrine(index)) {
       open = false;
       return;
     }
@@ -339,7 +385,7 @@ function applyUrinalysisSection(normalizedText, candidates, deps) {
     const known = deps.patterns.some(
       ({ name, regex }) => !matchesIn(text, name, regex).next().done
     );
-    return { foreign: false, hasValue: deps.hasValue(text), known, text };
+    return { foreign: false, hasValue: deps.hasValue(text), known, text, urine: hasUrineCue(text) };
   });
   for (const candidate of resolved) {
     let index = lineStarts.length - 1;
@@ -347,7 +393,9 @@ function applyUrinalysisSection(normalizedText, candidates, deps) {
     const line = lines[index];
     line.known = true;
     const matched = normalizedText.slice(candidate.start, candidate.end);
-    if (!isSectionName(matched) && !candidate.entries.some((e) => deps.urineCodes.has(e.code))) {
+    if (candidate.entries.some((e) => deps.urineCodes.has(e.code))) {
+      line.urine = true;
+    } else if (!isSectionName(matched)) {
       line.foreign = true;
     }
   }
@@ -420,40 +468,13 @@ function foldCommas(text) {
     (comma, space, offset) => startsWithCatalogName(text.slice(offset + 1 + space.length)) ? comma : " "
   );
 }
-var GENETIC_CONTEXT_PATTERNS = [
-  /\b[nx][mrpc]_\d{6,}/,
-  // RefSeq: NM_000384.2, NP_, NR_, XM_
-  /\bens[gtp]\d{6,}/,
-  // Ensembl: ENSG00000084674
-  /\bp\.[a-z]{3}\d/,
-  // HGVS proteína: p.Trp448*
-  /\bc\.\d+[acgt]?[>_+-]/,
-  // HGVS codificante: c.1234A>G, c.76_78del
-  /\brs\d{4,}\b/,
-  // dbSNP
-  /\bgenes?\b/,
-  /\bvariante?s?\b/,
-  /\bexons?\b/,
-  /\bzygosity\b/,
-  /\bzigosidade\b/,
-  /\balleles?\b/,
-  /\balelos?\b/,
-  /\bmutations?\b/,
-  /\bmutac(ao|oes)\b/,
-  /\bpathogenic/,
-  /\bpatogenic/,
-  /\bheterozyg/,
-  /\bhomozyg/,
-  /\bheterozigot/,
-  /\bhomozigot/,
-  /\bsequence change\b/
-];
 var DIGIT_PATTERN = /\d/;
 var cachedUnitTokens = null;
 function getUnitTokens() {
   if (!cachedUnitTokens) {
+    const isExamName = (t) => URINALYSIS_SECTION_NAMES.has(t) || getNamePatterns().has(t);
     cachedUnitTokens = new Set(
-      Object.keys(UNIT_TO_UCUM).map((unit) => normalize(unit).trim()).filter(Boolean)
+      Object.keys(UNIT_TO_UCUM).map((unit) => normalize(unit).trim()).filter((unit) => unit && !isExamName(unit))
     );
   }
   return cachedUnitTokens;
@@ -1218,7 +1239,7 @@ async function main() {
     strict: false
   });
   if (values.version) {
-    process.stdout.write(`${"0.38.0"}
+    process.stdout.write(`${"0.38.1"}
 `);
     return;
   }

package/dist/index.cjs

@@ -119,6 +119,34 @@ var QUALITATIVE_VALUE_TERMS = /* @__PURE__ */ new Set([
   "undetectable",
   "yellow"
 ]);
+var GENETIC_CONTEXT_PATTERNS = [
+  /\b[nx][mrpc]_\d{6,}/,
+  // RefSeq: NM_000384.2, NP_, NR_, XM_
+  /\bens[gtp]\d{6,}/,
+  // Ensembl: ENSG00000084674
+  /\bp\.[a-z]{3}\d/,
+  // HGVS proteína: p.Trp448*
+  /\bc\.\d+[acgt]?[>_+-]/,
+  // HGVS codificante: c.1234A>G, c.76_78del
+  /\brs\d{4,}\b/,
+  // dbSNP
+  /\bgenes?\b/,
+  /\bvariante?s?\b/,
+  /\bexons?\b/,
+  /\bzygosity\b/,
+  /\bzigosidade\b/,
+  /\balleles?\b/,
+  /\balelos?\b/,
+  /\bmutations?\b/,
+  /\bmutac(ao|oes)\b/,
+  /\bpathogenic/,
+  /\bpatogenic/,
+  /\bheterozyg/,
+  /\bhomozyg/,
+  /\bheterozigot/,
+  /\bhomozigot/,
+  /\bsequence change\b/
+];
 
 // src/body-region.ts
 var WHOLE_BODY_COMPOSITION_CODES = /* @__PURE__ */ new Set([
@@ -267,19 +295,37 @@ function isSectionName(matched) {
   return URINALYSIS_SECTION_NAMES.has(matched) || URINALYSIS_SECTION_NAMES.has(matched.replace(/s$/, ""));
 }
 var MENTIONS_URINE = /(?<![\p{L}\p{N}])urin[ae](?![\p{L}\p{N}])/u;
+var SEDIMENT_FIELD_UNIT = /\/(?:hpf|lpf)(?![\p{L}\p{N}])/u;
+function hasUrineCue(line) {
+  return MENTIONS_URINE.test(line) || SEDIMENT_FIELD_UNIT.test(line);
+}
+function kindOf(line) {
+  const text = line.text.trim();
+  if (!text) return "blank";
+  if (URINALYSIS_HEADER.test(text) && !/\d/.test(text)) return "header";
+  if (line.foreign) return "foreign";
+  if (line.urine) return "urine";
+  return line.known || line.hasValue ? "neutral" : "unknown";
+}
 function urinalysisLineIndexes(lines) {
+  const kinds = lines.map(kindOf);
+  const continuesAsUrine = (from) => {
+    const next = kinds.findIndex(
+      (kind, index) => index > from && (kind === "urine" || kind === "foreign" || kind === "header")
+    );
+    return next !== -1 && kinds[next] === "urine";
+  };
   const inside = /* @__PURE__ */ new Set();
   let open = false;
-  lines.forEach((line, index) => {
-    const text = line.text.trim();
-    if (URINALYSIS_HEADER.test(text) && !/\d/.test(text)) {
+  kinds.forEach((kind, index) => {
+    if (kind === "header") {
       open = true;
       return;
     }
-    if (!open || !text) {
+    if (!open || kind === "blank") {
       return;
     }
-    if (line.foreign || !line.known && !line.hasValue && !MENTIONS_URINE.test(text)) {
+    if (kind === "foreign" || kind === "unknown" && !continuesAsUrine(index)) {
       open = false;
       return;
     }
@@ -323,7 +369,7 @@ function applyUrinalysisSection(normalizedText, candidates, deps) {
     const known = deps.patterns.some(
       ({ name, regex }) => !matchesIn(text, name, regex).next().done
     );
-    return { foreign: false, hasValue: deps.hasValue(text), known, text };
+    return { foreign: false, hasValue: deps.hasValue(text), known, text, urine: hasUrineCue(text) };
   });
   for (const candidate of resolved) {
     let index = lineStarts.length - 1;
@@ -331,7 +377,9 @@ function applyUrinalysisSection(normalizedText, candidates, deps) {
     const line = lines[index];
     line.known = true;
     const matched = normalizedText.slice(candidate.start, candidate.end);
-    if (!isSectionName(matched) && !candidate.entries.some((e) => deps.urineCodes.has(e.code))) {
+    if (candidate.entries.some((e) => deps.urineCodes.has(e.code))) {
+      line.urine = true;
+    } else if (!isSectionName(matched)) {
       line.foreign = true;
     }
   }
@@ -404,40 +452,13 @@ function foldCommas(text) {
     (comma, space, offset) => startsWithCatalogName(text.slice(offset + 1 + space.length)) ? comma : " "
   );
 }
-var GENETIC_CONTEXT_PATTERNS = [
-  /\b[nx][mrpc]_\d{6,}/,
-  // RefSeq: NM_000384.2, NP_, NR_, XM_
-  /\bens[gtp]\d{6,}/,
-  // Ensembl: ENSG00000084674
-  /\bp\.[a-z]{3}\d/,
-  // HGVS proteína: p.Trp448*
-  /\bc\.\d+[acgt]?[>_+-]/,
-  // HGVS codificante: c.1234A>G, c.76_78del
-  /\brs\d{4,}\b/,
-  // dbSNP
-  /\bgenes?\b/,
-  /\bvariante?s?\b/,
-  /\bexons?\b/,
-  /\bzygosity\b/,
-  /\bzigosidade\b/,
-  /\balleles?\b/,
-  /\balelos?\b/,
-  /\bmutations?\b/,
-  /\bmutac(ao|oes)\b/,
-  /\bpathogenic/,
-  /\bpatogenic/,
-  /\bheterozyg/,
-  /\bhomozyg/,
-  /\bheterozigot/,
-  /\bhomozigot/,
-  /\bsequence change\b/
-];
 var DIGIT_PATTERN = /\d/;
 var cachedUnitTokens = null;
 function getUnitTokens() {
   if (!cachedUnitTokens) {
+    const isExamName = (t) => URINALYSIS_SECTION_NAMES.has(t) || getNamePatterns().has(t);
     cachedUnitTokens = new Set(
-      Object.keys(_fhir.UNIT_TO_UCUM).map((unit) => normalize(unit).trim()).filter(Boolean)
+      Object.keys(_fhir.UNIT_TO_UCUM).map((unit) => normalize(unit).trim()).filter((unit) => unit && !isExamName(unit))
     );
   }
   return cachedUnitTokens;