Add result and job management APIs

packages/extractor/index.ts

@@ -1,9 +1,9 @@
 import { join } from "node:path";
 import { dataPath } from "../../deployPaths";
-import { jsonError, normalizeStem, securityHeaders, withAuth } from "../../security";
+import { isSafeStem, jsonError, normalizeStem, securityHeaders, withAuth } from "../../security";
 
 const UPLOAD_PAGE_PATH = join(import.meta.dir, "upload.html");
-import { mkdir } from "node:fs/promises";
+import { mkdir, readdir } from "node:fs/promises";
 import { flattenTemplate, extractFields, buildOutput, getNested } from "./extractor";
 
 const TEMPLATE_PATH =
@@ -59,6 +59,60 @@ function extractScalarText(value: unknown): string | undefined {
   return undefined;
 }
 
+function extractorDownloadUrl(stem: string, file: string): string {
+  return `/downloads/extractor/${encodeURIComponent(stem)}/${encodeURIComponent(file)}`;
+}
+
+async function listExtractorResults() {
+  let entries: string[];
+  try {
+    entries = await readdir(OUTPUTS_DIR);
+  } catch {
+    return [];
+  }
+
+  const results = [];
+  for (const stem of entries.toSorted()) {
+    if (!isSafeStem(stem)) continue;
+    const jsonFile = `${stem}.json`;
+    const pdfFile = `${stem}.pdf`;
+    const jsonPath = join(OUTPUTS_DIR, stem, jsonFile);
+    const pdfPath = join(OUTPUTS_DIR, stem, pdfFile);
+    if (!(await Bun.file(jsonPath).exists())) continue;
+    results.push({
+      stem,
+      files: {
+        json: (await Bun.file(jsonPath).exists()) ? jsonFile : undefined,
+        pdf: (await Bun.file(pdfPath).exists()) ? pdfFile : undefined,
+      },
+      downloads: [
+        { label: "JSON herunterladen", url: extractorDownloadUrl(stem, jsonFile) },
+        ...((await Bun.file(pdfPath).exists()) ? [{ label: "PDF herunterladen", url: extractorDownloadUrl(stem, pdfFile) }] : []),
+      ],
+    });
+  }
+  return results;
+}
+
+async function serveExtractorDownload(stem: string, file: string): Promise<Response> {
+  if (!isSafeStem(stem) || (file !== `${stem}.json` && file !== `${stem}.pdf`)) {
+    return new Response("Not found", { status: 404, headers: securityHeaders });
+  }
+  const path = join(OUTPUTS_DIR, stem, file);
+  const output = Bun.file(path);
+  if (!(await output.exists())) {
+    return new Response("Not found", { status: 404, headers: securityHeaders });
+  }
+  return new Response(output, {
+    headers: {
+      ...securityHeaders,
+      "Content-Type": file.endsWith(".json") ? "application/json; charset=utf-8" : "application/pdf",
+      "Cache-Control": "private, no-store",
+      "Content-Disposition": `attachment; filename="${file.replaceAll('"', "")}"`,
+    },
+  });
+}
+
 export const routes = {
   "/extractor": {
     GET: withAuth(async () => new Response(Bun.file(UPLOAD_PAGE_PATH), {
@@ -67,6 +121,41 @@ export const routes = {
     ),
   },
 
+  "/api/extractor/results": {
+    GET: withAuth(async () => Response.json({ results: await listExtractorResults() })),
+  },
+
+  "/api/extractor/results/:stem": {
+    GET: withAuth(async (req: Request) => {
+      const stem = (req as Request & { params: Record<string, string | undefined> }).params.stem ?? "";
+      if (!isSafeStem(stem)) {
+        return jsonError("Invalid result id", 400);
+      }
+      const jsonPath = join(OUTPUTS_DIR, stem, `${stem}.json`);
+      const pdfPath = join(OUTPUTS_DIR, stem, `${stem}.pdf`);
+      const jsonFile = Bun.file(jsonPath);
+      if (!(await jsonFile.exists())) {
+        return Response.json({ error: "Extractor result not found" }, { status: 404 });
+      }
+      return Response.json({
+        ok: true,
+        stem,
+        result: await jsonFile.json(),
+        downloads: [
+          { label: "JSON herunterladen", url: extractorDownloadUrl(stem, `${stem}.json`) },
+          ...((await Bun.file(pdfPath).exists()) ? [{ label: "PDF herunterladen", url: extractorDownloadUrl(stem, `${stem}.pdf`) }] : []),
+        ],
+      });
+    }),
+  },
+
+  "/downloads/extractor/:stem/:file": {
+    GET: withAuth(async (req: Request) => {
+      const params = (req as Request & { params: Record<string, string | undefined> }).params;
+      return serveExtractorDownload(params.stem ?? "", params.file ?? "");
+    }),
+  },
+
   "/extract": {
     POST: withAuth(async (req: Request) => {
       const contentLength = Number(req.headers.get("content-length") ?? 0);