staging/main/0.12.0 (#1145)

* refactor: Upgrade the models to use keras 3.0 (#1138)

* Replace snappy with cramjam (#1091)

* add downloads tile (#1085)

* Replace snappy with cramjam

* Delete test_no_snappy

---------

Co-authored-by: Taylor Turner <[email protected]>

* pre-commit fix (#1122)

* Bug fix for float precision calculation using categorical data with trailing zeros. (#1125)

* Revert "Bug fix for float precision calculation using categorical data with t…" (#1133)

This reverts commit d3159bd.

* refactor: move layers outside of class

* refactor: update model to keras 3.0

* fix: manifest

* fix: bugs in compile and train

* fix: bug in load_from_library

* fix: bugs in CharCNN

* refactor: loading tf model labeler

* fix: bug in data_labeler identification

* fix: update model to use proper softmax layer names

* fix: formatting

* fix: remove unused line

* refactor: drop support for 3.8

* fix: comments

* fix: comment

---------

Co-authored-by: Gábor Lipták <[email protected]>
Co-authored-by: Taylor Turner <[email protected]>
Co-authored-by: James Schadt <[email protected]>

* Fix Tox (#1143)

* tox new

* update

* update

* update

* update

* update

* update

* update

* update tox.ini

* update

* update

* remove docs

* empty retrigger

* update (#1146)

* bump version

* update 3.11

* remove dist/

---------

Co-authored-by: JGSweets <[email protected]>
Co-authored-by: Gábor Lipták <[email protected]>
Co-authored-by: James Schadt <[email protected]>

Author: 4 people

.github/workflows/publish-python-package.yml

@@ -20,7 +20,7 @@ jobs:
     - name: Set up Python
       uses: actions/setup-python@v5
       with:
-        python-version: '3.10'
+        python-version: '3.11'
     - name: Install dependencies
       run: |
         python -m pip install --upgrade pip

.github/workflows/test-python-package.yml

@@ -16,7 +16,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.8, 3.9, "3.10"]
+        python-version: [3.9, "3.10", "3.11"]
 
     steps:
     - uses: actions/checkout@v4

MANIFEST.in

@@ -1,4 +1,5 @@
 global-exclude .DS_Store
+global-exclude */__pycache__/*
 
 include *.txt
 include CODEOWNERS

dataprofiler/labelers/char_load_tf_model.py

@@ -237,7 +237,8 @@ def _construct_model(self) -> None:
         model_loc = self._parameters["model_path"]
 
         self._model: tf.keras.Model = tf.keras.models.load_model(model_loc)
-        softmax_output_layer_name = self._model.outputs[0].name.split("/")[0]
+        self._model = tf.keras.Model(self._model.inputs, self._model.outputs)
+        softmax_output_layer_name = self._model.output_names[0]
         softmax_layer_ind = cast(
             int,
             labeler_utils.get_tf_layer_index_from_name(
@@ -252,21 +253,28 @@ def _construct_model(self) -> None:
                 num_labels, activation="softmax", name="softmax_output"
             )(self._model.layers[softmax_layer_ind - 1].output)
 
-        # Output the model into a .pb file for TensorFlow
-        argmax_layer = tf.keras.backend.argmax(new_softmax_layer)
+        # Add argmax layer to get labels directly as an output
+        argmax_layer = tf.keras.ops.argmax(new_softmax_layer, axis=2)
 
         argmax_outputs = [new_softmax_layer, argmax_layer]
         self._model = tf.keras.Model(self._model.inputs, argmax_outputs)
+        self._model = tf.keras.Model(self._model.inputs, self._model.outputs)
 
         # Compile the model w/ metrics
-        softmax_output_layer_name = self._model.outputs[0].name.split("/")[0]
+        softmax_output_layer_name = self._model.output_names[0]
         losses = {softmax_output_layer_name: "categorical_crossentropy"}
 
         # use f1 score metric
         f1_score_training = labeler_utils.F1Score(
             num_classes=num_labels, average="micro"
         )
-        metrics = {softmax_output_layer_name: ["acc", f1_score_training]}
+        metrics = {
+            softmax_output_layer_name: [
+                "categorical_crossentropy",
+                "acc",
+                f1_score_training,
+            ]
+        }
 
         self._model.compile(loss=losses, optimizer="adam", metrics=metrics)
 
@@ -294,30 +302,33 @@ def _reconstruct_model(self) -> None:
         num_labels = self.num_labels
         default_ind = self.label_mapping[self._parameters["default_label"]]
 
-        # Remove the 2 output layers ('softmax', 'tf_op_layer_ArgMax')
-        for _ in range(2):
-            self._model.layers.pop()
-
         # Add the final Softmax layer to the previous spot
+        # self._model.layers[-2] to skip: original softmax
         final_softmax_layer = tf.keras.layers.Dense(
             num_labels, activation="softmax", name="softmax_output"
-        )(self._model.layers[-4].output)
+        )(self._model.layers[-2].output)
 
-        # Output the model into a .pb file for TensorFlow
-        argmax_layer = tf.keras.backend.argmax(final_softmax_layer)
+        # Add argmax layer to get labels directly as an output
+        argmax_layer = tf.keras.ops.argmax(final_softmax_layer, axis=2)
 
         argmax_outputs = [final_softmax_layer, argmax_layer]
         self._model = tf.keras.Model(self._model.inputs, argmax_outputs)
 
         # Compile the model
-        softmax_output_layer_name = self._model.outputs[0].name.split("/")[0]
+        softmax_output_layer_name = self._model.output_names[0]
         losses = {softmax_output_layer_name: "categorical_crossentropy"}
 
         # use f1 score metric
         f1_score_training = labeler_utils.F1Score(
             num_classes=num_labels, average="micro"
         )
-        metrics = {softmax_output_layer_name: ["acc", f1_score_training]}
+        metrics = {
+            softmax_output_layer_name: [
+                "categorical_crossentropy",
+                "acc",
+                f1_score_training,
+            ]
+        }
 
         self._model.compile(loss=losses, optimizer="adam", metrics=metrics)
 
@@ -370,7 +381,7 @@ def fit(
         f1_report: dict = {}
 
         self._model.reset_metrics()
-        softmax_output_layer_name = self._model.outputs[0].name.split("/")[0]
+        softmax_output_layer_name = self._model.output_names[0]
 
         start_time = time.time()
         batch_id = 0