Migrate ArrayRecord parsing off TF

.pre-commit-config.yaml

@@ -8,7 +8,7 @@ repos:
       - id: codespell
         args:
           - '-w'
-          - '--skip="*.txt,pylintrc,.*,src/maxtext/assets/*"'
+          - '--skip="*.txt,pylintrc,.*,src/maxtext/assets/*,src/maxtext/input_pipeline/protos/*"'
           - '-L ND,nd,sems,TE,ROUGE,rouge,astroid,ags,dout'
           - '.'
         additional_dependencies:
@@ -30,6 +30,7 @@ repos:
         args:
           - '--disable=R0401,R0917,W0201,W0613'
           - "--ignore-patterns='.pytype,.*pyi$'"
+          - '--ignore-paths=src/maxtext/input_pipeline/protos'
           - 'benchmarks'
           - 'src'
           - 'tests'
@@ -47,6 +48,7 @@ repos:
     rev: 24.10.1
     hooks:
       - id: pyink
+        exclude: src/maxtext/input_pipeline/protos/
         args:
           - '--pyink-indentation=2'
           - '--line-length=122'

src/maxtext/input_pipeline/input_pipeline_utils.py

@@ -21,17 +21,17 @@
 
 if TYPE_CHECKING:
   import datasets
+  import tensorflow as tf
 
 import grain.python as grain
 import numpy as np
-import tensorflow as tf
+from maxtext.input_pipeline.protos import example_pb2
 from maxtext.input_pipeline import tokenizer
 from maxtext.multimodal import processor as mm_processor
 from maxtext.multimodal import utils as mm_utils
 from maxtext.utils import max_logging
 
-Features = dict[str, tf.Tensor]
-AUTOTUNE = tf.data.experimental.AUTOTUNE
+Features = dict[str, Any]
 INPUT_TOKENS_KEY = "input_ids"
 
 ########## Functions used by TFDS pipeline
@@ -58,6 +58,8 @@ def shift_data_by_truncation(x):
 
 
 def add_segmentation_and_position(x, data_columns, padding_token=0):
+  import tensorflow as tf  # pylint: disable=import-outside-toplevel
+
   for data_column in data_columns:
     x[f"{data_column}_segmentation"] = tf.cast(x[data_column] != padding_token, tf.int32)
     x[f"{data_column}_position"] = tf.broadcast_to(
@@ -68,6 +70,7 @@ def add_segmentation_and_position(x, data_columns, padding_token=0):
 
 def TokenizeOp(tokenizer_model, features: Features, data_keys: Iterable[str] = ("inputs", "targets")) -> Features:
   """Op for tokenization"""
+  import tensorflow as tf  # pylint: disable=import-outside-toplevel
 
   def _process_string(string_tensor):
     # Extract string value and decode it if necessary
@@ -421,20 +424,23 @@ class ParseFeatures(grain.MapTransform):
 
   def __init__(self, data_columns, tokenize):
     self.data_columns = data_columns
-    if tokenize:
-      self.dtype = tf.string
-    else:
-      self.dtype = tf.int64
+    self.tokenize = tokenize
 
   def map(self, element):
-    def _parse(example):
-      parsed = tf.io.parse_example(
-          example,
-          {col: tf.io.FixedLenSequenceFeature([], dtype=self.dtype, allow_missing=True) for col in self.data_columns},
-      )
-      return parsed
-
-    return _parse(element)
+    """Parse a serialized tf.train.Example proto and extract features."""
+    example = example_pb2.Example()
+    example.ParseFromString(element)
+    features = example.features.feature
+
+    parsed = {}
+    for col in self.data_columns:
+      if col in features:
+        f = features[col]
+        if self.tokenize:
+          parsed[col] = np.array(f.bytes_list.value, dtype=object)
+        else:
+          parsed[col] = np.array(f.int64_list.value, dtype=np.int32)
+    return parsed
 
 
 @dataclasses.dataclass
@@ -447,9 +453,9 @@ def __init__(self, column_names, tokenize):
 
   def map(self, element):
     if self.tokenize:
-      return {col: element[col].numpy()[0].decode() for col in self.column_names}
+      return {col: element[col][0].decode() for col in self.column_names}
     else:
-      return {col: element[col].numpy() for col in self.column_names}
+      return {col: element[col] for col in self.column_names}
 
 
 @dataclasses.dataclass

src/maxtext/input_pipeline/protos/__init__.py

@@ -0,0 +1,13 @@
+# Copyright 2023–2026 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.