From c272c4f84411429e7b61cfb7dcc7c54c341bcb74 Mon Sep 17 00:00:00 2001
From: Roland Tannous <rolandtannous@gravityq.ai>
Date: Mon, 9 Mar 2026 22:00:20 +0000
Subject: [PATCH] fix: prefer tabular files over archives in Tier 1 dataset
 preview

Tier 1 check-format was picking images.zip over testmini.parquet,
causing wrong columns (image/label) and broken VLM mapping.
Also log first VLM conversion failure instead of swallowing silently.
---
 studio/backend/routes/datasets.py             | 32 ++++++++++++-------
 .../utils/datasets/format_conversion.py       |  9 ++++--
 2 files changed, 27 insertions(+), 14 deletions(-)

diff --git a/studio/backend/routes/datasets.py b/studio/backend/routes/datasets.py
index d21974d540..25965a847e 100644
--- a/studio/backend/routes/datasets.py
+++ b/studio/backend/routes/datasets.py
@@ -83,16 +83,12 @@ def _serialize_preview_rows(rows):
 # --- Endpoints ---
 
 # Recognized data-file extensions for the single-file fallback approach.
-DATA_EXTS = (
-    '.parquet',
-    '.json', '.jsonl',
-    '.csv', '.tsv',
-    '.txt',
-    '.arrow',
-    '.tar', '.tar.gz', '.tgz',
-    '.gz', '.zst',
-    '.zip',
-)
+# Tabular formats are preferred over archives for Tier 1 preview because
+# archives (e.g. images.zip) may be loaded as ImageFolder datasets with
+# synthetic columns (image/label) that don't match the real dataset schema.
+_TABULAR_EXTS = ('.parquet', '.json', '.jsonl', '.csv', '.tsv', '.arrow')
+_ARCHIVE_EXTS = ('.tar', '.tar.gz', '.tgz', '.gz', '.zst', '.zip', '.txt')
+DATA_EXTS = _TABULAR_EXTS + _ARCHIVE_EXTS
 LOCAL_FILE_EXTS = ('.json', '.jsonl', '.csv', '.parquet')
 LOCAL_UPLOAD_EXTS = {".csv", ".json", ".jsonl", ".parquet"}
 BACKEND_ROOT = Path(__file__).resolve().parents[1]
@@ -363,8 +359,20 @@ def check_format(
                 )
                 data_files = [f for f in repo_files if any(f.endswith(ext) for ext in DATA_EXTS)]
 
-                if data_files:
-                    first_file = data_files[0]
+                # Prefer tabular formats over archives (e.g. images.zip → ImageFolder
+                # with synthetic image/label columns that don't match the real schema).
+                tabular_files = [f for f in data_files if any(f.endswith(ext) for ext in _TABULAR_EXTS)]
+                candidates = tabular_files or data_files
+
+                # When a subset is specified, narrow to files whose name matches
+                # (e.g. subset="testmini" → prefer "testmini.parquet").
+                if request.subset and candidates:
+                    subset_matches = [f for f in candidates if request.subset in Path(f).stem]
+                    if subset_matches:
+                        candidates = subset_matches
+
+                if candidates:
+                    first_file = candidates[0]
                     logger.info(f"Tier 1: loading single file {first_file}")
                     load_kwargs = {
                         "path": request.dataset_name,
diff --git a/studio/backend/utils/datasets/format_conversion.py b/studio/backend/utils/datasets/format_conversion.py
index bb80dd8d6e..2d53db3a3e 100644
--- a/studio/backend/utils/datasets/format_conversion.py
+++ b/studio/backend/utils/datasets/format_conversion.py
@@ -443,8 +443,10 @@ def convert_to_vlm_format(
                     idx = futures[future]
                     try:
                         batch_results[idx] = future.result()
-                    except Exception:
+                    except Exception as e:
                         failed_count += 1
+                        if failed_count == 1:
+                            print(f"⚠️ First VLM conversion failure: {type(e).__name__}: {e}")
 
             converted_list.extend(r for r in batch_results if r is not None)
 
@@ -463,8 +465,11 @@ def convert_to_vlm_format(
         for sample in pbar:
             try:
                 converted_list.append(_convert_single_sample(sample))
-            except Exception:
+            except Exception as e:
                 failed_count += 1
+                if failed_count == 1:
+                    # Log the first failure to aid debugging
+                    print(f"⚠️ First VLM conversion failure: {type(e).__name__}: {e}")
             pbar.set_postfix(ok=len(converted_list), failed=failed_count, refresh=False)
         pbar.close()