@precisa-saude/fhir-ocr-utils 0.4.0 → 0.5.0

package/dist/cli.js

@@ -0,0 +1,224 @@
+#!/usr/bin/env node
+
+// src/cli/index.ts
+import { parseArgs } from "util";
+import { exitWithError } from "@precisa-saude/fhir/cli-utils";
+
+// src/cli/commands/codes.ts
+import { getInput, outputJson, outputText } from "@precisa-saude/fhir/cli-utils";
+
+// src/anchor.ts
+import {
+  generateFilteredLLMReference,
+  getAllSearchPatterns
+} from "@precisa-saude/fhir";
+function normalize(text) {
+  return text.normalize("NFD").replace(/[\u0300-\u036f]/g, "").toLowerCase().replace(/\s+/g, " ");
+}
+var UNAMBIGUOUS_SHORT_NAMES = /* @__PURE__ */ new Set([
+  "hdl",
+  "ldl",
+  "lh",
+  "tsh",
+  "crp",
+  "pcr",
+  "ggt",
+  "alt",
+  "ast",
+  "bun",
+  "wbc",
+  "rbc",
+  "mcv",
+  "mch",
+  "rdw",
+  "mpv",
+  "psa",
+  "fsh",
+  "hba1c",
+  "egfr",
+  "acr",
+  "esr",
+  "vhs",
+  "bmc",
+  "bmd",
+  "vat",
+  "dxa",
+  "dmo",
+  "cmo",
+  "ffm",
+  "lbm",
+  "mlg",
+  "tav"
+]);
+var cachedPatterns = null;
+var cachedNormalized = null;
+function getPatterns() {
+  if (!cachedPatterns) {
+    cachedPatterns = getAllSearchPatterns();
+  }
+  return cachedPatterns;
+}
+function getNormalizedPatterns() {
+  if (!cachedNormalized) {
+    const patterns = getPatterns();
+    const map = /* @__PURE__ */ new Map();
+    for (const pattern of patterns) {
+      for (const name of pattern.names) {
+        const normalized = normalize(name);
+        const existing = map.get(normalized) || [];
+        existing.push({
+          code: pattern.code,
+          ...pattern.loinc && { loinc: pattern.loinc },
+          original: name
+        });
+        map.set(normalized, existing);
+      }
+    }
+    cachedNormalized = map;
+  }
+  return cachedNormalized;
+}
+function findBiomarkersInText(ocrText) {
+  const startTime = Date.now();
+  const normalizedText = normalize(ocrText);
+  const matchedCodes = /* @__PURE__ */ new Set();
+  const matches = [];
+  const normalizedPatterns = getNormalizedPatterns();
+  for (const [normalizedName, entries] of normalizedPatterns) {
+    if (normalizedName.length < 3 && !UNAMBIGUOUS_SHORT_NAMES.has(normalizedName)) {
+      continue;
+    }
+    let position = -1;
+    if (normalizedName.length <= 4) {
+      const regex = new RegExp(`\\b${normalizedName}\\b`);
+      const match = regex.exec(normalizedText);
+      if (match) {
+        position = match.index;
+      }
+    } else {
+      position = normalizedText.indexOf(normalizedName);
+    }
+    if (position !== -1) {
+      for (const entry of entries) {
+        if (!matchedCodes.has(entry.code)) {
+          matchedCodes.add(entry.code);
+          matches.push({
+            code: entry.code,
+            confidence: 1,
+            loinc: entry.loinc,
+            matchedName: entry.original,
+            position
+          });
+        }
+      }
+    }
+  }
+  const scanTimeMs = Date.now() - startTime;
+  const matchedCodesArray = Array.from(matchedCodes);
+  return {
+    filteredReference: generateFilteredLLMReference(matchedCodesArray),
+    matches,
+    stats: {
+      matchedCount: matches.length,
+      scanTimeMs,
+      totalPatterns: getPatterns().length
+    }
+  };
+}
+function getMatchedCodes(result) {
+  return result.matches.map((m) => m.code);
+}
+
+// src/cli/commands/codes.ts
+async function codes(args, json) {
+  const text = await getInput(args[0]);
+  const result = findBiomarkersInText(text);
+  const matched = getMatchedCodes(result);
+  if (json) {
+    outputJson(matched);
+    return;
+  }
+  if (matched.length === 0) {
+    outputText("Nenhum biomarcador encontrado no texto.");
+    return;
+  }
+  outputText(matched.join("\n"));
+  outputText(`
+Total: ${matched.length} c\xF3digos encontrados`);
+}
+
+// src/cli/commands/find.ts
+import { formatTable, getInput as getInput2, outputJson as outputJson2, outputText as outputText2 } from "@precisa-saude/fhir/cli-utils";
+async function find(args, json) {
+  const text = await getInput2(args[0]);
+  const result = findBiomarkersInText(text);
+  if (json) {
+    outputJson2(result);
+    return;
+  }
+  if (result.matches.length === 0) {
+    outputText2("Nenhum biomarcador encontrado no texto.");
+    return;
+  }
+  const rows = result.matches.map((m) => [
+    m.code,
+    m.loinc ?? "\u2014",
+    m.matchedName,
+    m.confidence.toFixed(2),
+    String(m.position)
+  ]);
+  outputText2(formatTable(["C\xF3digo", "LOINC", "Match", "Confian\xE7a", "Posi\xE7\xE3o"], rows));
+  outputText2(
+    `
+Encontrados: ${result.stats.matchedCount} de ${result.stats.totalPatterns} padr\xF5es (${result.stats.scanTimeMs}ms)`
+  );
+}
+
+// src/cli/index.ts
+var HELP = `fhir-ocr \u2014 CLI do @precisa-saude/fhir-ocr-utils
+
+Uso: fhir-ocr <comando> [op\xE7\xF5es]
+
+Comandos:
+  find [arquivo]          Encontrar biomarcadores em texto OCR
+  codes [arquivo]         Extrair c\xF3digos de biomarcadores encontrados no texto
+
+Flags globais:
+  --json                  Sa\xEDda em formato JSON
+  --help, -h              Mostrar ajuda
+  --version, -v           Mostrar vers\xE3o
+
+L\xEA de stdin quando nenhum arquivo \xE9 fornecido.
+`;
+var COMMANDS = {
+  codes,
+  find
+};
+async function main() {
+  const { positionals, values } = parseArgs({
+    allowPositionals: true,
+    options: {
+      help: { default: false, short: "h", type: "boolean" },
+      json: { default: false, type: "boolean" },
+      version: { default: false, short: "v", type: "boolean" }
+    },
+    strict: false
+  });
+  if (values.version) {
+    process.stdout.write(`${"0.5.0"}
+`);
+    return;
+  }
+  const [command, ...rest] = positionals;
+  if (values.help || !command) {
+    process.stdout.write(HELP);
+    return;
+  }
+  const handler = COMMANDS[command];
+  if (!handler) {
+    exitWithError(`Comando desconhecido: ${command}
+Use --help para ver os comandos dispon\xEDveis.`);
+  }
+  await handler(rest, Boolean(values.json));
+}
+main().catch((err) => exitWithError(err.message));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@precisa-saude/fhir-ocr-utils",
-  "version": "0.4.0",
+  "version": "0.5.0",
   "description": "Utilitários de ancoragem OCR para extração de biomarcadores de PDFs de resultados laboratoriais",
   "keywords": [
     "fhir",
@@ -34,6 +34,9 @@
   "main": "./dist/index.cjs",
   "module": "./dist/index.js",
   "types": "./dist/index.d.ts",
+  "bin": {
+    "fhir-ocr": "./dist/cli.js"
+  },
   "files": [
     "dist"
   ],