Convert uploaded .xlsx files to csv artifacts server-side

Uploading a .xlsx file no longer becomes a generic downloadable binary -
it's read server-side with openpyxl and converted to a csv artifact,
rendering in the same interactive grid as a native CSV post. Detected by
content type or .xlsx extension (some clients send generic/incorrect
content types for multipart uploads). Only the active sheet converts;
formula cells use whatever cached result the spreadsheet software last
saved (openpyxl doesn't evaluate formulas itself). Corrupt/unreadable
files return 400.

Since the existing posting script already forwards any non-text file via
multipart with no special-casing, `daybook_post.py balance_sheet.xlsx`
already works with zero script changes - verified live.

Two real bugs found and fixed while building this, both now covered by
regression tests:
- {**a_querydict, ...} silently wrapped every value (including title) in
  a list - Python's ** spread on a dict subclass bypasses QueryDict's
  __getitem__ override and exposes its internal one-list-per-key storage
  directly. Fixed by building the delegated dict explicitly via .get().
- Pre-setting None for absent optional fields (e.g. source_prompt) defeated
  create_text_artifact's own data.get(field, default) calls, since .get()
  only falls through to its default when the key is *missing*, not when
  the value is None - caused a NOT NULL constraint failure. Fixed by only
  setting keys that were actually provided.
- csv.writer's default "\r\n" terminator was getting doubled by
  Path.write_text()'s own newline translation on Windows, producing a
  blank line between every row. Fixed with lineterminator="\n".

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

tests/api/test_artifacts.py

@@ -1,8 +1,10 @@
 import json
 import shutil
 import tempfile
+from io import BytesIO
 from pathlib import Path
 
+import openpyxl
 from django.contrib.auth import get_user_model
 from django.test import override_settings
 from django.urls import reverse
@@ -455,3 +457,75 @@ class ArtifactAPITestCase(APITestCase):
 
         artifact = Artifact.objects.get(pk=artifact_id)
         self.assertEqual(artifact.source_name, "test-agent")  # unchanged, not mangled
+
+    # -- xlsx upload converts to csv kind ------------------------------------
+
+    def make_xlsx(self, rows, filename="sheet.xlsx"):
+        workbook = openpyxl.Workbook()
+        sheet = workbook.active
+        for row in rows:
+            sheet.append(row)
+        buffer = BytesIO()
+        workbook.save(buffer)
+        buffer.seek(0)
+        buffer.name = filename
+        return buffer
+
+    def test_xlsx_upload_converts_to_csv_kind(self):
+        xlsx = self.make_xlsx([["Item", "Q1", "Q2"], ["Cash", 1000, 1200]])
+        response = self.client.post(
+            reverse("artifact-api"),
+            {"source": "test-agent", "title": "Balance Sheet", "file": xlsx},
+            format="multipart",
+            **self.auth_headers(self.token),
+        )
+        self.assertEqual(response.status_code, 201)
+        self.assertEqual(response.data["kind"], "csv")
+
+        artifact = Artifact.objects.get(pk=response.data["id"])
+        raw_path = (self.artifacts_root / artifact.path).with_suffix(".csv")
+        # read_text() normalizes line endings, so this is OS-portable rather
+        # than asserting the exact on-disk bytes (which are "\r\n" on Windows).
+        self.assertEqual(raw_path.read_text(encoding="utf-8"), "Item,Q1,Q2\nCash,1000,1200\n")
+
+        stored_html = (self.artifacts_root / artifact.path).read_text(encoding="utf-8")
+        self.assertIn('"columns": ["Item", "Q1", "Q2"]', stored_html)
+        self.assertIn('["Cash", "1000", "1200"]', stored_html)
+
+    def test_xlsx_upload_does_not_mangle_fields(self):
+        """Regression test: {**a_querydict} previously wrapped every value
+        (including title) in a list - QueryDict's __getitem__ override is
+        bypassed by the ** spread, exposing its internal list-per-key
+        storage directly. Separately, pre-setting None for absent optional
+        fields defeated create_text_artifact's own .get(field, default)
+        calls (.get only falls through when the key is missing, not None),
+        which previously caused a NOT NULL constraint failure on
+        source_prompt."""
+        xlsx = self.make_xlsx([["a", "b"]])
+        response = self.client.post(
+            reverse("artifact-api"),
+            {
+                "source": "test-agent",
+                "title": "Untitled Sheet",
+                "tags": "finance,q2",
+                "file": xlsx,
+            },
+            format="multipart",
+            **self.auth_headers(self.token),
+        )
+        self.assertEqual(response.status_code, 201)
+        self.assertEqual(response.data["title"], "Untitled Sheet")  # not ["Untitled Sheet"]
+        self.assertEqual(response.data["source_prompt"], "")  # defaulted, not None
+        self.assertEqual([t["name"] for t in response.data["tags"]], ["finance", "q2"])
+
+    def test_xlsx_upload_rejects_corrupt_file(self):
+        buffer = BytesIO(b"not a real xlsx file")
+        buffer.name = "broken.xlsx"
+        response = self.client.post(
+            reverse("artifact-api"),
+            {"source": "test-agent", "title": "Broken", "file": buffer},
+            format="multipart",
+            **self.auth_headers(self.token),
+        )
+        self.assertEqual(response.status_code, 400)
+

tests/unit/test_xlsx_conversion.py

@@ -0,0 +1,60 @@
+from io import BytesIO
+from types import SimpleNamespace
+
+import openpyxl
+from django.test import SimpleTestCase
+
+from artifacts.views import XLSX_CONTENT_TYPE, is_xlsx_upload, xlsx_to_csv_text
+
+
+class IsXlsxUploadTestCase(SimpleTestCase):
+    def test_matches_exact_content_type(self):
+        upload = SimpleNamespace(name="whatever.bin")
+        self.assertTrue(is_xlsx_upload(upload, XLSX_CONTENT_TYPE))
+
+    def test_matches_by_extension_regardless_of_content_type(self):
+        upload = SimpleNamespace(name="sheet.xlsx")
+        self.assertTrue(is_xlsx_upload(upload, "application/octet-stream"))
+        self.assertTrue(is_xlsx_upload(upload, ""))
+
+    def test_extension_match_is_case_insensitive(self):
+        upload = SimpleNamespace(name="SHEET.XLSX")
+        self.assertTrue(is_xlsx_upload(upload, ""))
+
+    def test_rejects_other_files(self):
+        upload = SimpleNamespace(name="report.pdf")
+        self.assertFalse(is_xlsx_upload(upload, "application/pdf"))
+
+
+class XlsxToCsvTextTestCase(SimpleTestCase):
+    def make_xlsx(self, rows):
+        workbook = openpyxl.Workbook()
+        sheet = workbook.active
+        for row in rows:
+            sheet.append(row)
+        buffer = BytesIO()
+        workbook.save(buffer)
+        buffer.seek(0)
+        return buffer
+
+    def test_converts_active_sheet_to_csv(self):
+        upload = self.make_xlsx([["Item", "Q1", "Q2"], ["Cash", 1000, 1200]])
+        content = xlsx_to_csv_text(upload)
+        self.assertEqual(content, "Item,Q1,Q2\nCash,1000,1200\n")
+
+    def test_uses_single_line_terminator_not_doubled(self):
+        """Regression test: csv.writer's default "\\r\\n" gets a second
+        newline translation from Path.write_text()'s text-mode handling when
+        this string is later saved to disk, producing a blank line between
+        every row on Windows - lineterminator="\\n" avoids embedding "\\r\\n"
+        in the string in the first place, leaving the OS-level text-mode
+        write as the only place "\\n" becomes "\\r\\n"."""
+        upload = self.make_xlsx([["a"], ["b"], ["c"]])
+        content = xlsx_to_csv_text(upload)
+        self.assertNotIn("\r", content)
+        self.assertEqual(content.count("\n"), 3)
+
+    def test_empty_cells_become_empty_strings(self):
+        upload = self.make_xlsx([["a", "b"], ["x", None]])
+        content = xlsx_to_csv_text(upload)
+        self.assertEqual(content, "a,b\nx,\n")