Fix error message when kafka is not installed

python/hopsworks_common/decorators.py

@@ -20,8 +20,10 @@
 import os
 
 from hopsworks_common.core.constants import (
+    HAS_CONFLUENT_KAFKA,
     HAS_GREAT_EXPECTATIONS,
     HAS_POLARS,
+    confluent_kafka_not_installed_message,
     great_expectations_not_installed_message,
     polars_not_installed_message,
 )
@@ -96,3 +98,12 @@ def g(*args, **kwds):
         return f(*args, **kwds)
 
     return g
+
+def uses_kafka(f):
+    @functools.wraps(f)
+    def g(*args, **kwds):
+        if not HAS_CONFLUENT_KAFKA:
+            raise ModuleNotFoundError(confluent_kafka_not_installed_message)
+        return f(*args, **kwds)
+
+    return g

python/hsfs/core/kafka_engine.py

@@ -20,17 +20,26 @@
 from io import BytesIO
 from typing import TYPE_CHECKING, Any, Callable, Dict, Literal, Optional, Tuple, Union
 
-import pandas as pd
 from hopsworks_common import client
 from hopsworks_common.core.constants import HAS_NUMPY
+from hopsworks_common.decorators import uses_kafka
 from hsfs.core import storage_connector_api
-from hsfs.core.constants import HAS_AVRO, HAS_CONFLUENT_KAFKA, HAS_FAST_AVRO
+from hsfs.core.constants import (
+    HAS_AVRO,
+    HAS_CONFLUENT_KAFKA,
+    HAS_FAST_AVRO,
+    HAS_PANDAS,
+    avro_not_installed_message,
+)
 from tqdm import tqdm
 
 
 if HAS_NUMPY:
     import numpy as np
 
+if HAS_PANDAS:
+    import pandas as pd
+
 if HAS_CONFLUENT_KAFKA:
     from confluent_kafka import Consumer, KafkaError, Producer, TopicPartition
 
@@ -46,6 +55,7 @@
     from hsfs.feature_group import ExternalFeatureGroup, FeatureGroup
 
 
+@uses_kafka
 def init_kafka_consumer(
     feature_store_id: int,
     offline_write_options: Dict[str, Any],
@@ -112,6 +122,7 @@ def _init_kafka_resources(
     return producer, headers, feature_writers, writer
 
 
+@uses_kafka
 def init_kafka_producer(
     feature_store_id: int,
     offline_write_options: Dict[str, Any],
@@ -120,6 +131,7 @@ def init_kafka_producer(
     return Producer(get_kafka_config(feature_store_id, offline_write_options))
 
 
+@uses_kafka
 def kafka_get_offsets(
     topic_name: str,
     feature_store_id: int,
@@ -192,6 +204,9 @@ def get_encoder_func(writer_schema: str) -> callable:
         parsed_schema = parse_schema(schema)
         return lambda record, outf: schemaless_writer(outf, parsed_schema, record)
 
+    if not HAS_AVRO:
+        raise ModuleNotFoundError(avro_not_installed_message)
+
     parsed_schema = avro.schema.parse(writer_schema)
     writer = avro.io.DatumWriter(parsed_schema)
     return lambda record, outf: writer.write(record, avro.io.BinaryEncoder(outf))
@@ -207,11 +222,11 @@ def encode_row(complex_feature_writers, writer, row):
             # for avro to be able to serialize them, they need to be python data types
             if HAS_NUMPY and isinstance(row[k], np.ndarray):
                 row[k] = row[k].tolist()
-            if isinstance(row[k], pd.Timestamp):
+            if HAS_PANDAS and isinstance(row[k], pd.Timestamp):
                 row[k] = row[k].to_pydatetime()
             if isinstance(row[k], datetime) and row[k].tzinfo is None:
                 row[k] = row[k].replace(tzinfo=timezone.utc)
-            if isinstance(row[k], pd._libs.missing.NAType):
+            if HAS_PANDAS and isinstance(row[k], pd._libs.missing.NAType):
                 row[k] = None
     # encode complex features
     row = encode_complex_features(complex_feature_writers, row)
@@ -244,6 +259,7 @@ def get_kafka_config(
     return config
 
 
+@uses_kafka
 def build_ack_callback_and_optional_progress_bar(
     n_rows: int, is_multi_part_insert: bool, offline_write_options: Dict[str, Any]
 ) -> Tuple[Callable, Optional[tqdm]]: