Add cli client

README.md

@@ -116,6 +116,13 @@ plot_duplicates(image_dir='path/to/image/directory',
                 duplicate_map=duplicates,
                 filename='ukbench00120.jpg')
 ```
+
+Alternatively, you can also use our library in the CLI:
+```bash
+imagededup find-duplicates --image_dir path/to/image/directory --method PHash
+``` 
+Use the `--help` flag to get more information on the available options.
+
 For more examples, refer [this](https://github.com/idealo/imagededup/tree/master/examples) part of the
 repository.
 

imagededup/client/client.py

@@ -0,0 +1,48 @@
+import importlib
+import click
+
+from typing import Optional
+from pathlib import PosixPath
+
+
+@click.group()
+def cli():
+    pass
+
+
+@cli.command()
+@click.option('--image_dir', help='Path to the directory containing all the images.', type=str, required=True)
+@click.option('--method', help='Select which algorithm to use.',
+              type=click.Choice(['PHash', 'DHash', 'WHash', 'AHash', 'CNN']), required=True)
+@click.option('--outfile', help='Name of the file the results should be written to.', type=str)
+@click.option('--min_similarity_threshold',
+              help='For CNN only: threshold value (must be float between -1.0 and 1.0). Default is 0.9.',
+              type=click.FloatRange(-1.0, 1.0),
+              default=0.9)
+@click.option('--max_distance_threshold',
+              help='For hashing methods only: threshold value (must be integer between 0 and 64). Default is 10.',
+              type=click.IntRange(0, 64), default=10)
+@click.option('--scores',
+              help='Boolean indicating whether scores are to be returned along with retrieved duplicates.',
+              type=bool)
+def find_duplicates(image_dir: PosixPath,
+                    method: str,
+                    outfile: Optional[str],
+                    min_similarity_threshold: float,
+                    max_distance_threshold: int,
+                    scores: bool) -> None:
+    selected_method = getattr(importlib.import_module('imagededup.methods'), method)()
+    encodings = selected_method.encode_images(image_dir)
+
+    if method == 'CNN':
+        duplicates = selected_method.find_duplicates(encoding_map=encodings,
+                                                     outfile=outfile,
+                                                     min_similarity_threshold=min_similarity_threshold,
+                                                     scores=scores)
+    else:
+        duplicates = selected_method.find_duplicates(encoding_map=encodings,
+                                                     outfile=outfile,
+                                                     max_distance_threshold=max_distance_threshold,
+                                                     scores=scores)
+    if outfile is None:
+        click.echo(duplicates)

setup.py

@@ -36,7 +36,8 @@
         'tensorflow~=2.0.0',
         'tqdm',
         'scikit-learn',
-        'matplotlib'
+        'matplotlib',
+        'Click'
     ],
     extras_require={
         'tests': ['pytest', 'pytest-cov', 'pytest-mock', 'codecov'],
@@ -55,4 +56,5 @@
         'Topic :: Software Development :: Libraries :: Python Modules',
     ],
     packages=find_packages(exclude=('tests',)),
+    entry_points={'console_scripts': ['imagededup=imagededup.client.client:cli']},
 )

tests/test_client.py

@@ -0,0 +1,153 @@
+import os
+
+from click.testing import CliRunner
+from imagededup.client.client import find_duplicates
+
+PATH_IMAGE_DIR = 'tests/data/mixed_images'
+FILENAME = 'tests/test_output.json'
+
+
+def test_no_image_dir_given():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates, ['--image_dir', ''])
+    assert result.exit_code == 2
+
+
+def test_image_dir_given_but_no_method():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates, ['--image_dir', PATH_IMAGE_DIR])
+    assert result.exit_code == 2
+
+
+def test_image_dir_given_and_method():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates, ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash'])
+    assert result.exit_code == 0
+
+
+def test_image_dir_given_but_wrong_method():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates, ['--image_dir', PATH_IMAGE_DIR, '--method', 'LHash'])
+    assert result.exit_code == 2
+
+
+def test_file_is_created():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--outfile', FILENAME])
+    assert result.exit_code == 0
+    assert os.path.isfile(FILENAME) is True
+    # cleanup
+    os.remove(FILENAME)
+
+
+def test_hash_max_distance_threshold_int():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--max_distance_threshold', '20'])
+    assert result.exit_code == 0
+
+
+def test_hash_max_distance_threshold_no_int():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--max_distance_threshold', '0.5'])
+    assert result.exit_code == 2
+
+
+def test_hash_max_distance_threshold_in_range_left_interval():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--max_distance_threshold', '0'])
+    assert result.exit_code == 0
+
+
+def test_hash_max_distance_threshold_in_range_right_interval():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--max_distance_threshold', '64'])
+    assert result.exit_code == 0
+
+
+def test_hash_max_distance_threshold_out_of_range_negative():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--max_distance_threshold', '-30'])
+    assert result.exit_code == 2
+
+
+def test_hash_max_distance_threshold_out_of_range_positive():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--max_distance_threshold', '900'])
+    assert result.exit_code == 2
+
+
+def test_hash_min_similarity_threshold_has_no_effect():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--min_similarity_threshold', '0.5'])
+    assert result.exit_code == 0
+
+
+def test_cnn_min_similarity_threshold_float():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'CNN', '--min_similarity_threshold', '0.5'])
+    assert result.exit_code == 0
+
+
+def test_cnn_min_similarity_threshold_no_float():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'CNN', '--min_similarity_threshold', '10'])
+    assert result.exit_code == 2
+
+
+def test_cnn_min_similarity_threshold_in_range_left_interval():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'CNN', '--min_similarity_threshold', '-1.0'])
+    assert result.exit_code == 0
+
+
+def test_cnn_min_similarity_threshold_in_range_right_interval():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'CNN', '--min_similarity_threshold', '1.0'])
+    assert result.exit_code == 0
+
+
+def test_cnn_min_similarity_threshold_out_of_range_negative():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'CNN', '--min_similarity_threshold', '-1.5'])
+    assert result.exit_code == 2
+
+
+def test_cnn_min_similarity_threshold_out_of_range_positive():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'CNN', '--min_similarity_threshold', '1.5'])
+    assert result.exit_code == 2
+
+
+def test_cnn_max_distance_threshold_has_no_effect():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--max_distance_threshold', '10'])
+    assert result.exit_code == 0
+
+
+def test_scores_boolean():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--scores', 'False'])
+    assert result.exit_code == 0
+
+
+def test_scores_no_boolean():
+    runner = CliRunner()
+    result = runner.invoke(find_duplicates,
+                           ['--image_dir', PATH_IMAGE_DIR, '--method', 'PHash', '--scores', 'hello'])
+    assert result.exit_code == 2