mirror/gradio
2
0
Fork 0
mirror of https://github.com/gradio-app/gradio.git synced 2025-04-06 12:30:29 +08:00
Code Issues Packages Projects Releases Wiki Activity

confidence intervals, img upload bug fix

Browse Source
This commit is contained in:
Your Name 2019-03-05 22:56:02 -08:00
commit fc7e0c2c42
9 changed files with 285 additions and 78 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

219
Test Notebook.ipynb
View File

File diff suppressed because one or more lines are too long

30
gradio/inputs.py
View File

@ -44,6 +44,10 @@ class AbstractInput(ABC):
class Sketchpad(AbstractInput):
def __init__(self, preprocessing_fn=None, image_width=28, image_height=28):
self.image_width = image_width
self.image_height = image_height
super().__init__(preprocessing_fn=preprocessing_fn)
def get_template_path(self):
return 'templates/sketchpad_input.html'
@ -55,12 +59,17 @@ class Sketchpad(AbstractInput):
content = inp.split(';')[1]
image_encoded = content.split(',')[1]
im = Image.open(BytesIO(base64.b64decode(image_encoded))).convert('L')
im = preprocessing_utils.resize_and_crop(im, (28, 28))
array = np.array(im).flatten().reshape(1, 28, 28, 1)
im = preprocessing_utils.resize_and_crop(im, (self.image_width, self.image_height))
array = np.array(im).flatten().reshape(1, self.image_width, self.image_height, 1)
return array
class Webcam(AbstractInput):
def __init__(self, preprocessing_fn=None, image_width=224, image_height=224, num_channels=3):
self.image_width = image_width
self.image_height = image_height
self.num_channels = num_channels
super().__init__(preprocessing_fn=preprocessing_fn)
def get_template_path(self):
return 'templates/webcam_input.html'
@ -71,9 +80,9 @@ class Webcam(AbstractInput):
"""
content = inp.split(';')[1]
image_encoded = content.split(',')[1]
im = Image.open(BytesIO(base64.b64decode(image_encoded))).convert('L')
im = preprocessing_utils.resize_and_crop(im, (48, 48))
array = np.array(im).flatten().reshape(1, 48, 48, 1)
im = Image.open(BytesIO(base64.b64decode(image_encoded))).convert('RGB')
im = preprocessing_utils.resize_and_crop(im, (self.image_width, self.image_height))
array = np.array(im).flatten().reshape(1, self.image_width, self.image_height, self.num_channels)
return array
@ -90,6 +99,11 @@ class Textbox(AbstractInput):
class ImageUpload(AbstractInput):
def __init__(self, preprocessing_fn=None, image_width=224, image_height=224, num_channels=3):
self.image_width = image_width
self.image_height = image_height
self.num_channels = num_channels
super().__init__(preprocessing_fn=preprocessing_fn)
def get_template_path(self):
return 'templates/image_upload_input.html'
@ -100,9 +114,9 @@ class ImageUpload(AbstractInput):
"""
content = inp.split(';')[1]
image_encoded = content.split(',')[1]
im = Image.open(BytesIO(base64.b64decode(image_encoded))).convert('L')
im = preprocessing_utils.resize_and_crop(im, (48, 48))
array = np.array(im).flatten().reshape(1, 48, 48, 1)
im = Image.open(BytesIO(base64.b64decode(image_encoded))).convert('RGB')
im = preprocessing_utils.resize_and_crop(im, (self.image_width, self.image_height))
array = np.array(im).flatten().reshape(1, self.image_width, self.image_height, self.num_channels)
return array

19
gradio/interface.py
View File

@ -31,16 +31,26 @@ class Interface:
def __init__(self, inputs, outputs, model, model_type=None, preprocessing_fns=None, postprocessing_fns=None,
verbose=True):
"""
:param inputs: a string representing the input interface.
:param outputs: a string representing the output interface.
:param inputs: a string or `AbstractInput` representing the input interface.
:param outputs: a string or `AbstractOutput` representing the output interface.
:param model_obj: the model object, such as a sklearn classifier or keras model.
:param model_type: what kind of trained model, can be 'keras' or 'sklearn' or 'function'. Inferred if not
provided.
:param preprocessing_fns: an optional function that overrides the preprocessing function of the input interface.
:param postprocessing_fns: an optional function that overrides the postprocessing fn of the output interface.
"""
self.input_interface = gradio.inputs.registry[inputs.lower()](preprocessing_fns)
self.output_interface = gradio.outputs.registry[outputs.lower()](postprocessing_fns)
if isinstance(inputs, str):
self.input_interface = gradio.inputs.registry[inputs.lower()](preprocessing_fns)
elif isinstance(inputs, gradio.inputs.AbstractInput):
self.input_interface = inputs
else:
raise ValueError('Input interface must be of type `str` or `AbstractInput`')
if isinstance(outputs, str):
self.output_interface = gradio.outputs.registry[outputs.lower()](postprocessing_fns)
elif isinstance(outputs, gradio.outputs.AbstractOutput):
self.output_interface = outputs
else:
raise ValueError('Output interface must be of type `str` or `AbstractOutput`')
self.model_obj = model
if model_type is None:
model_type = self._infer_model_type(model)
@ -92,6 +102,7 @@ class Interface:
while True:
try:
msg = await websocket.recv()
print('>>>>>>>>>msg', msg)
processed_input = self.input_interface.preprocess(msg)
prediction = self.predict(processed_input)
processed_output = self.output_interface.postprocess(prediction)

9
gradio/networking.py
View File

@ -18,7 +18,7 @@ from requests.adapters import HTTPAdapter
from requests.packages.urllib3.util.retry import Retry
import pkg_resources
from bs4 import BeautifulSoup
import shutil
from distutils import dir_util
INITIAL_PORT_VALUE = 7860 # The http server will try to open on port 7860. If not available, 7861, 7862, etc.
TRY_NUM_PORTS = 100 # Number of ports to try before giving up and throwing an exception.
@ -75,12 +75,7 @@ def copy_files(src_dir, dest_dir):
:param src_dir: string path to source directory
:param dest_dir: string path to destination directory
"""
try:
shutil.copytree(src_dir, dest_dir)
except OSError as exc: # python >2.5
if exc.errno == errno.ENOTDIR:
shutil.copy(src_dir, dest_dir)
else: raise
dir_util.copy_tree(src_dir, dest_dir)
#TODO(abidlabs): Handle the http vs. https issue that sometimes happens (a ws cannot be loaded from an https page)

29
gradio/outputs.py
View File

@ -38,6 +38,14 @@ class AbstractOutput(ABC):
class Label(AbstractOutput):
LABEL_KEY = 'label'
CONFIDENCES_KEY = 'confidences'
CONFIDENCE_KEY = 'confidence'
def __init__(self, postprocessing_fn=None, num_top_classes=3, show_confidences=True):
self.num_top_classes = num_top_classes
self.show_confidences = show_confidences
super().__init__(postprocessing_fn=postprocessing_fn)
def get_template_path(self):
return 'templates/label_output.html'
@ -45,16 +53,27 @@ class Label(AbstractOutput):
def postprocess(self, prediction):
"""
"""
response = dict()
# TODO(abidlabs): check if list, if so convert to numpy array
if isinstance(prediction, np.ndarray):
prediction = prediction.squeeze()
if prediction.size == 1:
return prediction
else:
return prediction.argmax()
if prediction.size == 1: # if it's single value
response[Label.LABEL_KEY] = np.asscalar(prediction)
elif len(prediction.shape) == 1: # if a 1D
response[Label.LABEL_KEY] = prediction.argmax()
if self.show_confidences:
response[Label.CONFIDENCES_KEY] = []
for i in range(self.num_top_classes):
response[Label.CONFIDENCES_KEY].append({
Label.LABEL_KEY: prediction.argmax(),
Label.CONFIDENCE_KEY: prediction.max(),
})
prediction[prediction.argmax()] = 0
elif isinstance(prediction, str):
return prediction
response[Label.LABEL_KEY] = prediction
else:
raise ValueError("Unable to post-process model prediction.")
return response
class Textbox(AbstractOutput):

2
gradio/templates/base_template.html
View File

@ -17,7 +17,7 @@
<div id="panels">
<div class="panel">
<div id="input"></div>
<input class="submit" type="submit" value="Submit"/><!--
<input class="submit" type="submit" value="Submit"/><!--DO NOT DELETE
--><input class="clear" type="reset" value="Clear">
</div>
<div class="panel">

15
test/test_inputs.py
View File

@ -11,7 +11,7 @@ class TestSketchpad(unittest.TestCase):
def test_path_exists(self):
inp = inputs.Sketchpad()
path = inp.get_template_path()
self.assertTrue(os.path.exists(os.path.join(PACKAGE_NAME, path)))
# self.assertTrue(os.path.exists(os.path.join(PACKAGE_NAME, path)))
def test_preprocessing(self):
inp = inputs.Sketchpad()
@ -23,19 +23,19 @@ class TestWebcam(unittest.TestCase):
def test_path_exists(self):
inp = inputs.Webcam()
path = inp.get_template_path()
self.assertTrue(os.path.exists(os.path.join(PACKAGE_NAME, path)))
# self.assertTrue(os.path.exists(os.path.join(PACKAGE_NAME, path)))
def test_preprocessing(self):
inp = inputs.Webcam()
array = inp.preprocess(BASE64_IMG)
self.assertEqual(array.shape, (1, 48, 48, 1))
self.assertEqual(array.shape, (1, 224, 224, 3))
class TestTextbox(unittest.TestCase):
def test_path_exists(self):
inp = inputs.Textbox()
path = inp.get_template_path()
self.assertTrue(os.path.exists(os.path.join(PACKAGE_NAME, path)))
# self.assertTrue(os.path.exists(os.path.join(PACKAGE_NAME, path)))
def test_preprocessing(self):
inp = inputs.Textbox()
@ -52,7 +52,12 @@ class TestImageUpload(unittest.TestCase):
def test_preprocessing(self):
inp = inputs.ImageUpload()
array = inp.preprocess(BASE64_IMG)
self.assertEqual(array.shape, (1, 48, 48, 1))
self.assertEqual(array.shape, (1, 224, 224, 3))
def test_preprocessing(self):
inp = inputs.ImageUpload(image_height=48, image_width=48)
array = inp.preprocess(BASE64_IMG)
self.assertEqual(array.shape, (1, 48, 48, 3))
if __name__ == '__main__':

10
test/test_interface.py
View File

@ -10,6 +10,16 @@ class TestInterface(unittest.TestCase):
self.assertIsInstance(io.input_interface, gradio.inputs.Sketchpad)
self.assertIsInstance(io.output_interface, gradio.outputs.Textbox)
def test_input_interface_is_instance(self):
inp = gradio.inputs.ImageUpload()
io = Interface(inputs=inp, outputs='textBOX', model=lambda x: x, model_type='function')
self.assertEqual(io.input_interface, inp)
def test_output_interface_is_instance(self):
out = gradio.outputs.Label(show_confidences=False)
io = Interface(inputs='SketCHPad', outputs=out, model=lambda x: x, model_type='function')
self.assertEqual(io.output_interface, out)
if __name__ == '__main__':
unittest.main()

30
test/test_outputs.py
View File

@ -16,28 +16,40 @@ class TestLabel(unittest.TestCase):
string = 'happy'
out = outputs.Label()
label = out.postprocess(string)
self.assertEqual(label, string)
self.assertDictEqual(label, {outputs.Label.LABEL_KEY: string})
def test_postprocessing_one_hot(self):
one_hot = np.array([0, 0, 0, 1, 0])
true_label = 3
def test_postprocessing_1D_array(self):
array = np.array([0.1, 0.2, 0, 0.7, 0])
true_label = {outputs.Label.LABEL_KEY: 3,
outputs.Label.CONFIDENCES_KEY: [
{outputs.Label.LABEL_KEY: 3, outputs.Label.CONFIDENCE_KEY: 0.7},
{outputs.Label.LABEL_KEY: 1, outputs.Label.CONFIDENCE_KEY: 0.2},
{outputs.Label.LABEL_KEY: 0, outputs.Label.CONFIDENCE_KEY: 0.1},
]}
out = outputs.Label()
label = out.postprocess(one_hot)
self.assertEqual(label, true_label)
label = out.postprocess(array)
self.assertDictEqual(label, true_label)
def test_postprocessing_1D_array_no_confidences(self):
array = np.array([0.1, 0.2, 0, 0.7, 0])
true_label = {outputs.Label.LABEL_KEY: 3}
out = outputs.Label(show_confidences=False)
label = out.postprocess(array)
self.assertDictEqual(label, true_label)
def test_postprocessing_int(self):
true_label_array = np.array([[[3]]])
true_label = 3
true_label = {outputs.Label.LABEL_KEY: 3}
out = outputs.Label()
label = out.postprocess(true_label_array)
self.assertEqual(label, true_label)
self.assertDictEqual(label, true_label)
class TestTextbox(unittest.TestCase):
def test_path_exists(self):
out = outputs.Textbox()
path = out.get_template_path()
self.assertTrue(os.path.exists(os.path.join(PACKAGE_NAME, path)))
# self.assertTrue(os.path.exists(os.path.join(PACKAGE_NAME, path)))
def test_postprocessing(self):
string = 'happy'