precommit

Author: Victor49152

docs/dataset_manager/DESIGN.md

@@ -103,16 +103,16 @@ Transforms are composed in order; each receives the output of the previous.
 Registered in `dataset.py` under `Dataset.PREDEFINED`. Referenced by name in rulesets and YAML
 configs. Each predefined dataset ships with default transforms for supported model families.
 
-| Name                        | Source        | Notes                                      |
-| --------------------------- | ------------- | ------------------------------------------ |
-| `aime25`                    | AIME 2025     | Math reasoning                             |
-| `gpqa`                      | GPQA Diamond  | Science QA                                 |
-| `cnndailymail`              | CNN/DailyMail | Summarization                              |
-| `open_orca`                 | OpenOrca      | General instruction                        |
-| `livecodebench`             | LiveCodeBench | Code generation; requires additional setup |
-| `shopify_product_catalogue`    | Shopify       | E-commerce Q&A (q3vl)                      |
+| Name                           | Source        | Notes                                                 |
+| ------------------------------ | ------------- | ----------------------------------------------------- |
+| `aime25`                       | AIME 2025     | Math reasoning                                        |
+| `gpqa`                         | GPQA Diamond  | Science QA                                            |
+| `cnndailymail`                 | CNN/DailyMail | Summarization                                         |
+| `open_orca`                    | OpenOrca      | General instruction                                   |
+| `livecodebench`                | LiveCodeBench | Code generation; requires additional setup            |
+| `shopify_product_catalogue`    | Shopify       | E-commerce Q&A (q3vl)                                 |
 | `shopify_product_catalogue_8k` | Shopify       | 8k sample variant of Shopify product catalogue (q3vl) |
-| `random`                       | Synthetic     | Generated prompts for throughput testing   |
+| `random`                       | Synthetic     | Generated prompts for throughput testing              |
 
 ## Preset System
 

src/inference_endpoint/dataset_manager/__init__.py

@@ -27,7 +27,10 @@
 from .predefined.livecodebench import LiveCodeBench
 from .predefined.open_orca import OpenOrca
 from .predefined.random import RandomDataset
-from .predefined.shopify_product_catalogue import ShopifyProductCatalogue, ShopifyProductCatalogue8k
+from .predefined.shopify_product_catalogue import (
+    ShopifyProductCatalogue,
+    ShopifyProductCatalogue8k,
+)
 from .transforms import (
     AddStaticColumns,
     ColumnFilter,

src/inference_endpoint/dataset_manager/predefined/shopify_product_catalogue/__init__.py

@@ -24,9 +24,10 @@
 from typing import Any, ClassVar
 
 import pandas as pd
-from datasets import load_dataset
 from tqdm import tqdm
 
+from datasets import load_dataset
+
 from ...dataset import Dataset
 from . import presets
 from .metadata import ProductMetadata
@@ -139,9 +140,7 @@ def generate(
             load_options["revision"] = revision
 
         ds = load_dataset(cls.REPO_ID, split=split_key, **load_options)
-        logger.info(
-            f"Loaded {len(ds)} samples from {cls.REPO_ID} ({split_key})"
-        )
+        logger.info(f"Loaded {len(ds)} samples from {cls.REPO_ID} ({split_key})")
 
         all_rows: list[dict[str, Any]] = []
         for i in tqdm(

tests/unit/dataset_manager/test_shopify_product_catalogue.py

@@ -27,12 +27,12 @@
 from unittest.mock import patch
 
 import pandas as pd
+from inference_endpoint.dataset_manager.dataset import Dataset
 from inference_endpoint.dataset_manager.predefined.shopify_product_catalogue import (
     BaseShopifyProductCatalogue,
     ShopifyProductCatalogue,
     ShopifyProductCatalogue8k,
 )
-from inference_endpoint.dataset_manager.dataset import Dataset
 from inference_endpoint.dataset_manager.predefined.shopify_product_catalogue.presets import (
     ShopifyMultimodalFormatter,
     q3vl,
@@ -372,7 +372,9 @@ def test_class_inherits_from_base(self) -> None:
 
     def test_has_correct_repo_id(self) -> None:
         """REPO_ID points to nvidia/Shopify-product-catalogue-8k."""
-        assert ShopifyProductCatalogue8k.REPO_ID == "nvidia/Shopify-product-catalogue-8k"
+        assert (
+            ShopifyProductCatalogue8k.REPO_ID == "nvidia/Shopify-product-catalogue-8k"
+        )
 
     def test_has_correct_dataset_id(self) -> None:
         """DATASET_ID is shopify_product_catalogue_8k."""
@@ -381,11 +383,17 @@ def test_has_correct_dataset_id(self) -> None:
     def test_registered_in_dataset_predefined(self) -> None:
         """Class is auto-registered in Dataset.PREDEFINED."""
         assert "shopify_product_catalogue_8k" in Dataset.PREDEFINED
-        assert Dataset.PREDEFINED["shopify_product_catalogue_8k"] is ShopifyProductCatalogue8k
+        assert (
+            Dataset.PREDEFINED["shopify_product_catalogue_8k"]
+            is ShopifyProductCatalogue8k
+        )
 
     def test_shares_column_names_with_base(self) -> None:
         """Column names are identical to ShopifyProductCatalogue."""
-        assert ShopifyProductCatalogue8k.COLUMN_NAMES == ShopifyProductCatalogue.COLUMN_NAMES
+        assert (
+            ShopifyProductCatalogue8k.COLUMN_NAMES
+            == ShopifyProductCatalogue.COLUMN_NAMES
+        )
 
     def test_shares_presets_with_base(self) -> None:
         """Presets are shared with base class (q3vl works)."""
@@ -430,7 +438,12 @@ def test_generate_uses_correct_dataset_id_for_paths(
                 force=True,
             )
         # Verify cache path uses shopify_product_catalogue_8k
-        expected_path = tmp_path / "shopify_product_catalogue_8k" / "train" / "shopify_product_catalogue_8k_train.parquet"
+        expected_path = (
+            tmp_path
+            / "shopify_product_catalogue_8k"
+            / "train"
+            / "shopify_product_catalogue_8k_train.parquet"
+        )
         assert expected_path.exists()
 
     def test_get_dataloader_with_q3vl_preset(self, tmp_path: Path) -> None: