fix(scripts): skip documents the source already says are too large once HealthLog has named its limit, and show small sizes in KB

scripts/__tests__/import-documents.test.ts

@@ -343,7 +343,7 @@ async function papra() {
       return {
         status: 200,
         json: {
-          documentsCount: 3,
+          documentsCount: 4,
           documents: [
             {
               id: "doc_a",
@@ -365,6 +365,15 @@ async function papra() {
                 { id: "tag_l", name: "Lab" },
               ],
             },
+            {
+              id: "doc_huge",
+              name: "Scan.pdf",
+              originalName: "scan.pdf",
+              originalSize: 40_000_000,
+              createdAt: "2025-03-10T12:00:00.000Z",
+              documentDate: "2020-03-10",
+              tags: [{ id: "tag_h", name: "Health" }],
+            },
             {
               id: "doc_old",
               name: "Old.pdf",
@@ -677,7 +686,31 @@ describe("import-documents.mjs — Papra", () => {
     );
     expect(pp.seen[0].auth).toBe("Bearer pp_test");
     expect(res.stdout).toMatch(/Imported: 1/);
-    expect(res.stdout).toMatch(/Skipped: 1/);
+    // doc_b answered 413 with the cap; doc_huge is known to be larger than
+    // that and is skipped without being downloaded or sent.
+    expect(res.stdout).toMatch(/Skipped: 2/);
     expect(res.stdout).toMatch(/larger than HealthLog accepts/);
+    expect(pp.seen.some((s) => s.path.includes("doc_huge"))).toBe(false);
+  });
+
+  it("shows small sizes in KB on a dry run", async () => {
+    const pp = await papra();
+    const res = await runScript(
+      [
+        "papra",
+        "--papra-url",
+        pp.url,
+        "--papra-org",
+        "org_1",
+        "--tag",
+        "Health",
+        "--since",
+        "2016-01-01",
+        "--dry-run",
+      ],
+      { PAPRA_TOKEN: "pp_test" },
+    );
+    expect(res.code).toBe(0);
+    expect(res.stdout).toMatch(/2\.0 KB\s+Papra doc_a/);
   });
 });

scripts/import-documents.mjs

@@ -647,6 +647,8 @@ async function upload(options, source, doc, file) {
       return {
         outcome: "skipped",
         reason: `larger than HealthLog accepts${meta.maxFileBytes ? ` (${mib(meta.maxFileBytes)})` : ""}`,
+        maxFileBytes:
+          typeof meta.maxFileBytes === "number" ? meta.maxFileBytes : undefined,
       };
     case 415:
       return {
@@ -668,9 +670,10 @@ async function upload(options, source, doc, file) {
 
 function mib(bytes) {
   if (typeof bytes !== "number") return "?";
-  return bytes >= 1024 * 1024 * 1024
-    ? `${(bytes / 1024 ** 3).toFixed(1)} GB`
-    : `${(bytes / 1024 ** 2).toFixed(1)} MB`;
+  if (bytes >= 1024 ** 3) return `${(bytes / 1024 ** 3).toFixed(1)} GB`;
+  // Under 0.1 MB a size in MB reads as 0.0; say it in KB.
+  if (bytes < 0.1 * 1024 ** 2) return `${(bytes / 1024).toFixed(1)} KB`;
+  return `${(bytes / 1024 ** 2).toFixed(1)} MB`;
 }
 
 // ── Run ────────────────────────────────────────────────────────────────────
@@ -705,6 +708,8 @@ async function run(argv, env = process.env) {
   };
   let listed = 0;
   let totalBytes = 0;
+  /** HealthLog's per-file limit, once a 413 has said it. */
+  let maxFileBytes = null;
   let unknownSizes = 0;
 
   out(
@@ -743,6 +748,20 @@ async function run(argv, env = process.env) {
         continue;
       }
 
+      // HealthLog says its per-file limit only when it refuses a file. Once
+      // it has, a document the source already says is larger is skipped
+      // without downloading it.
+      if (
+        maxFileBytes !== null &&
+        typeof doc.size === "number" &&
+        doc.size > maxFileBytes
+      ) {
+        const reason = `larger than HealthLog accepts (${mib(maxFileBytes)})`;
+        tally.skipped.push({ where, reason });
+        out(`  skipped   ${where}: ${reason}`);
+        continue;
+      }
+
       const file = await source.download(doc);
       if (file.skip) {
         tally.skipped.push({ where, reason: file.skip });
@@ -750,6 +769,9 @@ async function run(argv, env = process.env) {
         continue;
       }
       const result = await upload(options, source, doc, file);
+      if (typeof result.maxFileBytes === "number") {
+        maxFileBytes = result.maxFileBytes;
+      }
       if (result.outcome === "imported") {
         tally.imported++;
         out(`  imported  ${where}`);