Improve unsupported PDF character errors

backend/src/routes/functions.ts

@@ -25,6 +25,7 @@ import { s3 } from "../utils/s3";
 import { secrets } from "../utils/secrets";
 import { storeExtractedTextForFile } from "../utils/documentText";
 import { generateLiquidityForecast } from "../utils/liquidityForecast";
+import { getUnsupportedPdfCharacterMessage } from "../utils/pdfCharacterError";
 dayjs.extend(customParseFormat)
 dayjs.extend(isoWeek)
 dayjs.extend(isBetween)
@@ -134,7 +135,15 @@ export default async function functionRoutes(server: FastifyInstance) {
             return pdf // Fastify wandelt automatisch in JSON
         } catch (err) {
             console.log(err)
-            reply.code(500).send({ error: "Failed to create PDF" })
+            const unsupportedCharacterMessage = getUnsupportedPdfCharacterMessage(err, body.data)
+            if (unsupportedCharacterMessage) {
+                return reply.code(422).send({
+                    error: unsupportedCharacterMessage,
+                    code: "UNSUPPORTED_PDF_CHARACTER",
+                })
+            }
+
+            return reply.code(500).send({ error: "Failed to create PDF" })
         }
     })
 

backend/src/utils/pdfCharacterError.ts

@@ -0,0 +1,79 @@
+const PDF_CHARACTER_ERROR_PATTERN = /WinAnsi cannot encode "[\s\S]+?" \(0x([0-9a-f]+)\)/i
+
+const DOCUMENT_FIELD_LABELS: Record<string, string> = {
+    title: "Dokumenttitel",
+    description: "Dokumentbeschreibung",
+    startText: "Einleitungstext",
+    endText: "Schlusstext",
+    text: "Bezeichnung",
+    descriptionText: "Beschreibung",
+    unit: "Einheit",
+}
+
+type CharacterLocation = {
+    label: string
+    characterPosition: number
+}
+
+const labelForPath = (path: Array<string | number>, root: any): string => {
+    if (path[0] === "rows" && typeof path[1] === "number") {
+        const row = root?.rows?.[path[1]]
+        const position = row?.pos ? `Position ${row.pos}` : `Dokumentzeile ${path[1] + 1}`
+        const field = DOCUMENT_FIELD_LABELS[String(path[path.length - 1])] || String(path[path.length - 1])
+        return `${position}, ${field}`
+    }
+
+    const field = String(path[path.length - 1])
+    return DOCUMENT_FIELD_LABELS[field] || field
+}
+
+const findCharacterLocations = (
+    value: unknown,
+    unsupportedCharacter: string,
+    root: any,
+    path: Array<string | number> = [],
+): CharacterLocation[] => {
+    if (typeof value === "string") {
+        const characters = Array.from(value)
+        return characters.flatMap((character, index) =>
+            character === unsupportedCharacter
+                ? [{ label: labelForPath(path, root), characterPosition: index + 1 }]
+                : []
+        )
+    }
+
+    if (Array.isArray(value)) {
+        return value.flatMap((entry, index) =>
+            findCharacterLocations(entry, unsupportedCharacter, root, [...path, index])
+        )
+    }
+
+    if (value && typeof value === "object") {
+        return Object.entries(value).flatMap(([key, entry]) =>
+            findCharacterLocations(entry, unsupportedCharacter, root, [...path, key])
+        )
+    }
+
+    return []
+}
+
+export const getUnsupportedPdfCharacterMessage = (error: unknown, documentData: any): string | null => {
+    const errorMessage = error instanceof Error ? error.message : String(error)
+    const match = errorMessage.match(PDF_CHARACTER_ERROR_PATTERN)
+    if (!match) return null
+
+    const numericCodePoint = Number.parseInt(match[1], 16)
+    if (!Number.isSafeInteger(numericCodePoint) || numericCodePoint > 0x10FFFF) return null
+
+    const unsupportedCharacter = String.fromCodePoint(numericCodePoint)
+    const [location] = findCharacterLocations(documentData, unsupportedCharacter, documentData)
+    const codePoint = numericCodePoint
+        .toString(16)
+        .toUpperCase()
+        .padStart(4, "0")
+    const locationText = location
+        ? ` Fundstelle: ${location.label}, Zeichen ${location.characterPosition}.`
+        : ""
+
+    return `Das Zeichen „${unsupportedCharacter}“ (Unicode U+${codePoint}) wird in PDF-Dokumenten nicht unterstützt.${locationText} Bitte ersetze oder entferne das Zeichen.`
+}

backend/tests/pdfCharacterError.test.ts

@@ -0,0 +1,27 @@
+import assert from "node:assert/strict"
+import test from "node:test"
+import { getUnsupportedPdfCharacterMessage } from "../src/utils/pdfCharacterError"
+
+test("reports unsupported characters with the document position and character offset", () => {
+    // pdf-lib displays supplementary Unicode characters as a private-use glyph,
+    // while the hexadecimal value still contains the original code point.
+    const error = new Error('WinAnsi cannot encode "" (0x1f600)')
+    const documentData = {
+        rows: [
+            { pos: "1", text: "Montage" },
+            { pos: "2.3", descriptionText: "Gerät 😀 montieren" },
+        ],
+    }
+
+    assert.equal(
+        getUnsupportedPdfCharacterMessage(error, documentData),
+        "Das Zeichen „😀“ (Unicode U+1F600) wird in PDF-Dokumenten nicht unterstützt. Fundstelle: Position 2.3, Beschreibung, Zeichen 7. Bitte ersetze oder entferne das Zeichen."
+    )
+})
+
+test("returns null for unrelated PDF errors", () => {
+    assert.equal(
+        getUnsupportedPdfCharacterMessage(new Error("Background PDF is invalid"), {}),
+        null
+    )
+})

frontend/plugins/api.ts

@@ -45,6 +45,11 @@ export default defineNuxtPlugin(() => {
             let title = "Fehler"
             let description = "Ein unerwarteter Fehler ist aufgetreten."
 
+            if (status === 422 && response._data?.code === "UNSUPPORTED_PDF_CHARACTER") {
+                title = "Nicht unterstütztes Zeichen"
+                description = response._data.error
+            }
+
             switch (status) {
                 case 400:
                     title = "Anfrage fehlerhaft"