initial commit

packages/extractor/extractor.ts

@@ -0,0 +1,176 @@
+// Use the legacy build — the main build relies on browser-only APIs (DOMMatrix
+// etc.) that aren't available in a Bun/Node server context.
+// @ts-ignore — legacy path has no separate .d.ts; types are compatible
+import * as pdfjsLib from "pdfjs-dist/legacy/build/pdf.mjs";
+
+// Point at the legacy worker. import.meta.resolve returns a file:// URL that
+// Bun can load as a Worker, which pdfjs requires even in server contexts.
+pdfjsLib.GlobalWorkerOptions.workerSrc = import.meta.resolve(
+  "pdfjs-dist/legacy/build/pdf.worker.mjs",
+);
+
+// ─── Types ────────────────────────────────────────────────────────────────────
+
+/** Raw extracted field values keyed by PDF field name. */
+export type FieldValues = Record<string, string>;
+
+export interface FrageVorlage {
+  id?: string;
+  text: string;
+}
+
+export interface TemplateLeafObject {
+  fieldName: string;
+  label?: string;
+  antwortFormat?: "einzelfrage" | "mehrfachfrage_ein_antwortfeld";
+  fragen?: FrageVorlage[];
+}
+
+export type TemplateLeaf = string | TemplateLeafObject;
+
+export interface FieldIndexEntry {
+  path: string;
+  fieldName: string;
+  label?: string;
+  antwortFormat?: "einzelfrage" | "mehrfachfrage_ein_antwortfeld";
+  fragen?: FrageVorlage[];
+  isRich?: boolean;
+}
+
+/**
+ * Flattened view of a template: PDF field name → metadata about the target path.
+ */
+export type FieldIndex = Record<string, FieldIndexEntry>;
+
+// ─── Template helpers ─────────────────────────────────────────────────────────
+
+function isTemplateLeafObject(value: unknown): value is TemplateLeafObject {
+  return typeof value === "object" && value !== null && typeof (value as Record<string, unknown>).fieldName === "string";
+}
+
+/**
+ * Recursively flattens the nested template produced by templatebuilder.
+ */
+export function flattenTemplate(
+  node: Record<string, unknown>,
+  prefix = "",
+): FieldIndex {
+  const index: FieldIndex = {};
+
+  for (const [key, value] of Object.entries(node)) {
+    const path = prefix ? `${prefix}.${key}` : key;
+
+    if (typeof value === "string") {
+      index[value] = {
+        path,
+        fieldName: value,
+        antwortFormat: "einzelfrage",
+        isRich: false,
+      };
+    } else if (isTemplateLeafObject(value)) {
+      index[value.fieldName] = {
+        path,
+        fieldName: value.fieldName,
+        label: value.label,
+        antwortFormat: value.antwortFormat ?? "einzelfrage",
+        fragen: value.fragen,
+        isRich: true,
+      };
+    } else if (typeof value === "object" && value !== null) {
+      Object.assign(index, flattenTemplate(value as Record<string, unknown>, path));
+    }
+  }
+
+  return index;
+}
+
+// ─── PDF extraction ───────────────────────────────────────────────────────────
+
+export async function extractFields(
+  data: ArrayBuffer,
+  options: { maxPages?: number } = {},
+): Promise<FieldValues> {
+  const pdf = await pdfjsLib.getDocument({ data }).promise;
+  if (options.maxPages && pdf.numPages > options.maxPages) {
+    throw new Error(`PDF has ${pdf.numPages} pages; the configured limit is ${options.maxPages}.`);
+  }
+  const values: FieldValues = {};
+
+  for (let p = 1; p <= pdf.numPages; p++) {
+    const page = await pdf.getPage(p);
+    const annotations = await page.getAnnotations();
+
+    for (const ann of annotations) {
+      if (ann.subtype !== "Widget" || !ann.fieldName) continue;
+
+      const raw = ann.fieldValue;
+      if (raw === undefined || raw === null) continue;
+
+      values[ann.fieldName] = Array.isArray(raw) ? raw.join(", ") : String(raw);
+    }
+  }
+
+  return values;
+}
+
+// ─── Output builder ───────────────────────────────────────────────────────────
+
+export function buildOutput(
+  values: FieldValues,
+  index: FieldIndex,
+): Record<string, unknown> {
+  const output: Record<string, unknown> = {};
+
+  for (const entry of Object.values(index)) {
+    const value = values[entry.fieldName];
+    if (value === undefined) continue;
+
+    if (!entry.isRich) {
+      setNested(output, entry.path, value);
+      continue;
+    }
+
+    const antwortFormat = entry.antwortFormat ?? "einzelfrage";
+    const fragen = (entry.fragen ?? []).map((frage, i) => ({
+      id: frage.id ?? `${entry.path.split(".").at(-1) ?? "frage"}_${i + 1}`,
+      text: frage.text,
+      ...(antwortFormat === "einzelfrage" ? { antwort: value } : {}),
+    }));
+
+    setNested(output, entry.path, {
+      fieldName: entry.fieldName,
+      label: entry.label,
+      antwortFormat,
+      zuordnungsmodus:
+        antwortFormat === "mehrfachfrage_ein_antwortfeld"
+          ? "gemeinsame_antwort"
+          : "einzelfrage",
+      fragen,
+      antwort: value,
+    });
+  }
+
+  return output;
+}
+
+export function getNested(obj: Record<string, unknown>, path: string): unknown {
+  let cur: unknown = obj;
+  for (const key of path.split(".")) {
+    if (typeof cur !== "object" || cur === null) return undefined;
+    cur = (cur as Record<string, unknown>)[key];
+  }
+  return cur;
+}
+
+// ─── Internal helpers ─────────────────────────────────────────────────────────
+
+function setNested(obj: Record<string, unknown>, path: string, value: unknown): void {
+  const parts = path.split(".");
+  let cur = obj;
+  for (let i = 0; i < parts.length - 1; i++) {
+    const key = parts[i]!;
+    if (typeof cur[key] !== "object" || cur[key] === null) cur[key] = {};
+    cur = cur[key] as Record<string, unknown>;
+  }
+  cur[parts[parts.length - 1]!] = value;
+}