Fix loaders for M5 & ETT datasets (#3155)

*Description of changes:*
- Fix how `item_id` is obtained for M5 and ETT datasets
- Fix `lxml` dependency range


By submitting this pull request, I confirm that you can use, modify,
copy, and redistribute this contribution, under the terms of your
choice.


**Please tag this pr with at least one of these labels to make our
release process faster:** BREAKING, new feature, bug fix, other change,
dev setup

pyproject.toml

@@ -20,7 +20,7 @@ filterwarnings = "ignore"
 [tool.ruff]
 line-length = 79
 
-ignore = [
+lint.ignore = [
     # line-length is handled by black
     "E501",
 

requirements/requirements-docs.txt

@@ -3,6 +3,7 @@ ipykernel~=6.5
 nbconvert~=6.5.1
 nbsphinx~=0.8.8
 notedown
+lxml~=5.1.0
 pytest-runner~=2.11
 recommonmark
 sphinx~=4.0

src/gluonts/dataset/repository/_ett_small.py

@@ -39,29 +39,29 @@ def generate_ett_small_dataset(
         dfs.append(df)
 
     test = []
-    for df in dfs:
+    for region, df in enumerate(dfs):
         start = pd.Period(df["date"][0], freq=freq)
         for col in df.columns:
             if col in ["date"]:
                 continue
             test.append(
                 {
                     "start": start,
-                    "item_id": col,
+                    "item_id": f"{col}_{region}",
                     "target": df[col].values,
                 }
             )
 
     train = []
-    for df in dfs:
+    for region, df in enumerate(dfs):
         start = pd.Period(df["date"][0], freq=freq)
         for col in df.columns:
             if col in ["date"]:
                 continue
             train.append(
                 {
                     "start": start,
-                    "item_id": col,
+                    "item_id": f"{col}_{region}",
                     "target": df[col].values[:-prediction_length],
                 }
             )

src/gluonts/dataset/repository/_m5.py

@@ -113,7 +113,11 @@ def generate_m5_dataset(
     ]
 
     # Build target series
-    train_ids = sales_train_validation["item_id"]
+    train_ids = (
+        sales_train_validation["item_id"].str
+        + "_"
+        + sales_train_validation["store_id"].str
+    )
     train_df = sales_train_validation.drop(
         ["id", "item_id", "dept_id", "cat_id", "store_id", "state_id"],
         axis=1,

src/gluonts/json.py

@@ -26,7 +26,8 @@
 character if set to `True`.
 """
 
-__all__ = [  # noqa
+# ruff: noqa: F822
+__all__ = [
     "variant",
     "dump",
     "dumps",