File renames:

  input_maker.py + input_maker_lib.py -> pack_nbrs.py
  input_maker_test.py -> pack_nbrs_test.py
  graph_builder.py + graph_builder_lib.py -> build_graph.py
  graph_builder_test.py -> build_graph_test.py

This change does NOT change the names of these functions:
  nsl.tools.build_graph
  nsl.tools.pack_nbrs

PiperOrigin-RevId: 274638570

Author: Neural-Link Team

neural_structured_learning/tools/BUILD

@@ -30,9 +30,9 @@ py_library(
     srcs = ["__init__.py"],
     srcs_version = "PY2AND3",
     deps = [
-        ":graph_builder_lib",
+        ":build_graph_lib",
         ":graph_utils",
-        ":input_maker_lib",
+        ":pack_nbrs_lib",
     ],
 )
 
@@ -57,11 +57,13 @@ py_test(
 )
 
 py_library(
-    name = "graph_builder_lib",
-    srcs = ["graph_builder_lib.py"],
+    name = "build_graph_lib",
+    srcs = ["build_graph.py"],
     srcs_version = "PY2AND3",
     deps = [
         ":graph_utils",
+        # package absl:app
+        # package absl/flags
         # package absl/logging
         # package numpy
         # package six
@@ -70,11 +72,11 @@ py_library(
 )
 
 py_test(
-    name = "graph_builder_lib_test",
-    srcs = ["graph_builder_lib_test.py"],
+    name = "build_graph_test",
+    srcs = ["build_graph_test.py"],
     srcs_version = "PY2AND3",
     deps = [
-        ":graph_builder_lib",
+        ":build_graph_lib",
         ":graph_utils",
         # package protobuf,
         # package absl/testing:absltest
@@ -83,50 +85,47 @@ py_test(
 )
 
 py_binary(
-    name = "graph_builder",
-    srcs = ["graph_builder.py"],
+    name = "build_graph",
+    srcs = ["build_graph.py"],
     python_version = "PY3",
     deps = [
-        ":graph_builder_lib",
-        # package absl:app
-        # package absl/flags
-        # package tensorflow
+        ":build_graph_lib",
     ],
 )
 
 py_library(
-    name = "input_maker_lib",
-    srcs = ["input_maker_lib.py"],
+    name = "pack_nbrs_lib",
+    srcs = ["pack_nbrs.py"],
     srcs_version = "PY2AND3",
     deps = [
         ":graph_utils",
+        # package absl:app
+        # package absl/flags
         # package absl/logging
         # package six
         # package tensorflow
     ],
 )
 
 py_test(
-    name = "input_maker_lib_test",
-    srcs = ["input_maker_lib_test.py"],
+    name = "pack_nbrs_test",
+    srcs = ["pack_nbrs_test.py"],
     srcs_version = "PY2AND3",
     deps = [
         ":graph_utils",
-        ":input_maker_lib",
+        ":pack_nbrs_lib",
         # package protobuf,
         # package absl/testing:absltest
         # package tensorflow
     ],
 )
 
 py_binary(
-    name = "input_maker",
-    srcs = ["input_maker.py"],
+    name = "pack_nbrs",
+    srcs = ["pack_nbrs.py"],
     python_version = "PY3",
     deps = [
-        ":input_maker_lib",
-        # package absl:app
-        # package absl/flags
+        ":pack_nbrs_lib",
         # package tensorflow
     ],
 )

neural_structured_learning/tools/__init__.py

@@ -1,8 +1,20 @@
-"""Tools and APIs for preparing data for Neural Structured Learning."""
+"""Tools and APIs for preparing data for Neural Structured Learning.
 
-from neural_structured_learning.tools.graph_builder_lib import build_graph
+In addition to the functions exported here, two of the modules can be invoked
+from the command-line as follows:
+
+```sh
+$ python -m neural_structured_learning.tools.build_graph.py ...
+$ python -m neural_structured_learning.tools.pack_nbrs.py ...
+```
+
+For details on the command-line usage for these programs, see the
+`nsl.tools.build_graph` and `nsl.tools.pack_nbrs` documentation.
+"""
+
+from neural_structured_learning.tools.build_graph import build_graph
 from neural_structured_learning.tools.graph_utils import add_edge
 from neural_structured_learning.tools.graph_utils import add_undirected_edges
 from neural_structured_learning.tools.graph_utils import read_tsv_graph
 from neural_structured_learning.tools.graph_utils import write_tsv_graph
-from neural_structured_learning.tools.input_maker_lib import pack_nbrs
+from neural_structured_learning.tools.pack_nbrs import pack_nbrs

neural_structured_learning/tools/graph_builder_lib.py renamed to neural_structured_learning/tools/build_graph.py

@@ -12,11 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-r"""Library to build a graph based on dense input features (embeddings).
-
-A python-based program for graph building also exists on
-[GitHub](https://github.com/tensorflow/neural-structured-learning/tree/master/neural_structured_learning/tools/graph_builder.py).
-"""
+r"""Program & library to build a graph from dense features (embeddings)."""
 
 from __future__ import absolute_import
 from __future__ import division
@@ -26,6 +22,8 @@
 import itertools
 import time
 
+from absl import app
+from absl import flags
 from absl import logging
 from neural_structured_learning.tools import graph_utils
 import numpy as np
@@ -165,6 +163,18 @@ def build_graph(embedding_files,
   All edges in the output will be symmetric (i.e., if edge `A--w-->B` exists in
   the output, then so will edge `B--w-->A`).
 
+  Note that this function can also be invoked as a binary from a shell. Sample
+  usage:
+
+  `python -m neural_structured_learning.tools.build_graph` [*flags*]
+  *embedding_file.tfr... output_graph.tsv*
+
+  For details about this program's flags, run:
+
+  ```
+  python -m neural_structured_learning.tools.build_graph --help
+  ```
+
   Args:
     embedding_files: A list of names of TFRecord files containing
       `tf.train.Example` objects, which in turn contain dense embeddings.
@@ -182,3 +192,35 @@ def build_graph(embedding_files,
   graph = collections.defaultdict(dict)
   _add_edges(embeddings, similarity_threshold, graph)
   graph_utils.write_tsv_graph(output_graph_path, graph)
+
+
+def _main(argv):
+  """Main function for invoking the `nsl.tools.build_graph` function."""
+  flag = flags.FLAGS
+  flag.showprefixforinfo = False
+  if len(argv) < 3:
+    raise app.UsageError(
+        'Invalid number of arguments; expected 2 or more, got %d' %
+        (len(argv) - 1))
+
+  build_graph(argv[1:-1], argv[-1], flag.similarity_threshold,
+              flag.id_feature_name, flag.embedding_feature_name)
+
+
+if __name__ == '__main__':
+  flags.DEFINE_string(
+      'id_feature_name', 'id',
+      """Name of the singleton bytes_list feature in each input Example
+      whose value is the Example's ID.""")
+  flags.DEFINE_string(
+      'embedding_feature_name', 'embedding',
+      """Name of the float_list feature in each input Example
+      whose value is the Example's (dense) embedding.""")
+  flags.DEFINE_float(
+      'similarity_threshold', 0.8,
+      """Lower bound on the cosine similarity required for an edge
+      to be created between two nodes.""")
+
+  # Ensure TF 2.0 behavior even if TF 1.X is installed.
+  tf.compat.v1.enable_v2_behavior()
+  app.run(_main)

neural_structured_learning/tools/graph_builder_lib_test.py renamed to neural_structured_learning/tools/build_graph_test.py

@@ -11,21 +11,21 @@
 # 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.
-"""Tests for neural_structured_learning.tools.graph_builder_lib."""
+"""Tests for neural_structured_learning.tools.build_graph."""
 
 from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
 
 from absl.testing import absltest
-from neural_structured_learning.tools import graph_builder_lib
+from neural_structured_learning.tools import build_graph as build_graph_lib
 from neural_structured_learning.tools import graph_utils
 import tensorflow as tf
 
 from google.protobuf import text_format
 
 
-class BuildGraphLibTest(absltest.TestCase):
+class BuildGraphTest(absltest.TestCase):
 
   def _create_embedding_file(self):
     return self.create_tempfile('embeddings.tfr').full_path
@@ -82,9 +82,9 @@ def testGraphBuildingNoThresholding(self):
     embedding_path = self._create_embedding_file()
     self._write_embeddings(embedding_path)
     graph_path = self._create_graph_file()
-    graph_builder_lib.build_graph([embedding_path],
-                                  graph_path,
-                                  similarity_threshold=0)
+    build_graph_lib.build_graph([embedding_path],
+                                graph_path,
+                                similarity_threshold=0)
     g_actual = graph_utils.read_tsv_graph(graph_path)
     self.assertDictEqual(
         g_actual, {
@@ -107,9 +107,9 @@ def testGraphBuildingWithThresholding(self):
     embedding_path = self._create_embedding_file()
     self._write_embeddings(embedding_path)
     graph_path = self._create_graph_file()
-    graph_builder_lib.build_graph([embedding_path],
-                                  graph_path,
-                                  similarity_threshold=0.51)
+    build_graph_lib.build_graph([embedding_path],
+                                graph_path,
+                                similarity_threshold=0.51)
     g_actual = graph_utils.read_tsv_graph(graph_path)
     self.assertDictEqual(g_actual, {})