mirror of
https://github.com/gradio-app/gradio.git
synced 2025-01-12 10:34:32 +08:00
Blocks-Components
- move audio component
This commit is contained in:
Ömer Faruk Özdemir
parent
cf32e0097c
commit
9e45418227
@ -1230,7 +1230,7 @@ class Video(Component):
|
||||
type: Optional[str] = None,
|
||||
source: str = "upload",
|
||||
label: Optional[str] = None,
|
||||
optional: bool = False,
|
||||
**kwargs
|
||||
):
|
||||
"""
|
||||
Parameters:
|
||||
@ -1242,7 +1242,7 @@ class Video(Component):
|
||||
"""
|
||||
self.type = type
|
||||
self.source = source
|
||||
super().__init__(label=label, optional=optional)
|
||||
super().__init__(label=label, **kwargs)
|
||||
|
||||
@classmethod
|
||||
def get_shortcut_implementations(cls):
|
||||
@ -1301,9 +1301,6 @@ class Video(Component):
|
||||
return self.save_flagged_file(
|
||||
dir, label, None if data is None else data["data"], encryption_key
|
||||
)
|
||||
# TODO: Might need to converge these two
|
||||
# Output save_flagged:
|
||||
# return self.save_flagged_file(dir, label, data["data"], encryption_key)
|
||||
|
||||
def generate_sample(self):
|
||||
return test_data.BASE64_VIDEO
|
||||
@ -1331,3 +1328,244 @@ class Video(Component):
|
||||
|
||||
def restore_flagged(self, dir, data, encryption_key):
|
||||
return self.restore_flagged_file(dir, data, encryption_key)
|
||||
|
||||
|
||||
class Audio(Component):
|
||||
"""
|
||||
Component accepts audio input files or creates an audio player that plays the output audio.
|
||||
|
||||
|
||||
Input type: Union[Tuple[int, numpy.array], file-object, numpy.array]
|
||||
Output type: Union[Tuple[int, numpy.array], str]
|
||||
Demos: main_note, generate_tone, reverse_audio, spectogram
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
default="",
|
||||
*,
|
||||
source: str = "upload",
|
||||
type: str = "numpy",
|
||||
label: str = None,
|
||||
**kwargs
|
||||
):
|
||||
"""
|
||||
Parameters:
|
||||
source (str): Source of audio. "upload" creates a box where user can drop an audio file, "microphone" creates a microphone input.
|
||||
type (str): Type of value to be returned by component. "numpy" returns a 2-set tuple with an integer sample_rate and the data numpy.array of shape (samples, 2), "file" returns a temporary file object whose path can be retrieved by file_obj.name, "filepath" returns the path directly.
|
||||
label (str): component name in interface.
|
||||
"""
|
||||
self.source = source
|
||||
requires_permissions = source == "microphone"
|
||||
self.type = type
|
||||
self.test_input = test_data.BASE64_AUDIO
|
||||
self.interpret_by_tokens = True
|
||||
super().__init__(label=label, requires_permissions=requires_permissions, **kwargs)
|
||||
|
||||
def get_template_context(self):
|
||||
return {
|
||||
"source": self.source, # TODO: This did not exist in output template, careful here if an error arrives
|
||||
**super().get_template_context(),
|
||||
}
|
||||
|
||||
@classmethod
|
||||
def get_shortcut_implementations(cls):
|
||||
return {
|
||||
"audio": {},
|
||||
"microphone": {"source": "microphone"},
|
||||
"mic": {"source": "microphone"},
|
||||
}
|
||||
|
||||
def preprocess_example(self, x):
|
||||
return {"name": x, "data": None, "is_example": True}
|
||||
|
||||
def preprocess(self, x: Dict[str, str] | None) -> Tuple[int, np.array] | str | None:
|
||||
"""
|
||||
Parameters:
|
||||
x (Dict[name: str, data: str]): JSON object with filename as 'name' property and base64 data as 'data' property
|
||||
Returns:
|
||||
(Union[Tuple[int, numpy.array], str, numpy.array]): audio in requested format
|
||||
"""
|
||||
if x is None:
|
||||
return x
|
||||
file_name, file_data, is_example = (
|
||||
x["name"],
|
||||
x["data"],
|
||||
x.get("is_example", False),
|
||||
)
|
||||
crop_min, crop_max = x.get("crop_min", 0), x.get("crop_max", 100)
|
||||
if is_example:
|
||||
file_obj = processing_utils.create_tmp_copy_of_file(file_name)
|
||||
else:
|
||||
file_obj = processing_utils.decode_base64_to_file(
|
||||
file_data, file_path=file_name
|
||||
)
|
||||
if crop_min != 0 or crop_max != 100:
|
||||
sample_rate, data = processing_utils.audio_from_file(
|
||||
file_obj.name, crop_min=crop_min, crop_max=crop_max
|
||||
)
|
||||
processing_utils.audio_to_file(sample_rate, data, file_obj.name)
|
||||
if self.type == "file":
|
||||
warnings.warn(
|
||||
"The 'file' type has been deprecated. Set parameter 'type' to 'filepath' instead.",
|
||||
DeprecationWarning,
|
||||
)
|
||||
return file_obj
|
||||
elif self.type == "filepath":
|
||||
return file_obj.name
|
||||
elif self.type == "numpy":
|
||||
return processing_utils.audio_from_file(file_obj.name)
|
||||
else:
|
||||
raise ValueError(
|
||||
"Unknown type: "
|
||||
+ str(self.type)
|
||||
+ ". Please choose from: 'numpy', 'filepath'."
|
||||
)
|
||||
|
||||
def serialize(self, x, called_directly):
|
||||
if x is None:
|
||||
return None
|
||||
if self.type == "filepath" or called_directly:
|
||||
name = x
|
||||
elif self.type == "file":
|
||||
warnings.warn(
|
||||
"The 'file' type has been deprecated. Set parameter 'type' to 'filepath' instead.",
|
||||
DeprecationWarning,
|
||||
)
|
||||
name = x.name
|
||||
elif self.type == "numpy":
|
||||
file = tempfile.NamedTemporaryFile(delete=False)
|
||||
name = file.name
|
||||
processing_utils.audio_to_file(x[0], x[1], name)
|
||||
else:
|
||||
raise ValueError(
|
||||
"Unknown type: "
|
||||
+ str(self.type)
|
||||
+ ". Please choose from: 'numpy', 'filepath'."
|
||||
)
|
||||
|
||||
file_data = processing_utils.encode_url_or_file_to_base64(name)
|
||||
return {"name": name, "data": file_data, "is_example": False}
|
||||
|
||||
def set_interpret_parameters(self, segments=8):
|
||||
"""
|
||||
Calculates interpretation score of audio subsections by splitting the audio into subsections, then using a "leave one out" method to calculate the score of each subsection by removing the subsection and measuring the delta of the output value.
|
||||
Parameters:
|
||||
segments (int): Number of interpretation segments to split audio into.
|
||||
"""
|
||||
self.interpretation_segments = segments
|
||||
return self
|
||||
|
||||
def tokenize(self, x):
|
||||
if x.get("is_example"):
|
||||
sample_rate, data = processing_utils.audio_from_file(x["name"])
|
||||
else:
|
||||
file_obj = processing_utils.decode_base64_to_file(x["data"])
|
||||
sample_rate, data = processing_utils.audio_from_file(file_obj.name)
|
||||
leave_one_out_sets = []
|
||||
tokens = []
|
||||
masks = []
|
||||
duration = data.shape[0]
|
||||
boundaries = np.linspace(0, duration, self.interpretation_segments + 1).tolist()
|
||||
boundaries = [round(boundary) for boundary in boundaries]
|
||||
for index in range(len(boundaries) - 1):
|
||||
start, stop = boundaries[index], boundaries[index + 1]
|
||||
masks.append((start, stop))
|
||||
|
||||
# Handle the leave one outs
|
||||
leave_one_out_data = np.copy(data)
|
||||
leave_one_out_data[start:stop] = 0
|
||||
file = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
|
||||
processing_utils.audio_to_file(sample_rate, leave_one_out_data, file.name)
|
||||
out_data = processing_utils.encode_file_to_base64(file.name)
|
||||
leave_one_out_sets.append(out_data)
|
||||
file.close()
|
||||
os.unlink(file.name)
|
||||
|
||||
# Handle the tokens
|
||||
token = np.copy(data)
|
||||
token[0:start] = 0
|
||||
token[stop:] = 0
|
||||
file = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
|
||||
processing_utils.audio_to_file(sample_rate, token, file.name)
|
||||
token_data = processing_utils.encode_file_to_base64(file.name)
|
||||
file.close()
|
||||
os.unlink(file.name)
|
||||
|
||||
tokens.append(token_data)
|
||||
tokens = [{"name": "token.wav", "data": token} for token in tokens]
|
||||
leave_one_out_sets = [
|
||||
{"name": "loo.wav", "data": loo_set} for loo_set in leave_one_out_sets
|
||||
]
|
||||
return tokens, leave_one_out_sets, masks
|
||||
|
||||
def get_masked_inputs(self, tokens, binary_mask_matrix):
|
||||
# create a "zero input" vector and get sample rate
|
||||
x = tokens[0]["data"]
|
||||
file_obj = processing_utils.decode_base64_to_file(x)
|
||||
sample_rate, data = processing_utils.audio_from_file(file_obj.name)
|
||||
zero_input = np.zeros_like(data, dtype="int16")
|
||||
# decode all of the tokens
|
||||
token_data = []
|
||||
for token in tokens:
|
||||
file_obj = processing_utils.decode_base64_to_file(token["data"])
|
||||
_, data = processing_utils.audio_from_file(file_obj.name)
|
||||
token_data.append(data)
|
||||
# construct the masked version
|
||||
masked_inputs = []
|
||||
for binary_mask_vector in binary_mask_matrix:
|
||||
masked_input = np.copy(zero_input)
|
||||
for t, b in zip(token_data, binary_mask_vector):
|
||||
masked_input = masked_input + t * int(b)
|
||||
file = tempfile.NamedTemporaryFile(delete=False)
|
||||
processing_utils.audio_to_file(sample_rate, masked_input, file.name)
|
||||
masked_data = processing_utils.encode_file_to_base64(file.name)
|
||||
file.close()
|
||||
os.unlink(file.name)
|
||||
masked_inputs.append(masked_data)
|
||||
return masked_inputs
|
||||
|
||||
def get_interpretation_scores(self, x, neighbors, scores, masks=None, tokens=None):
|
||||
"""
|
||||
Returns:
|
||||
(List[float]): Each value represents the interpretation score corresponding to an evenly spaced subsection of audio.
|
||||
"""
|
||||
return list(scores)
|
||||
|
||||
def save_flagged(self, dir, label, data, encryption_key):
|
||||
"""
|
||||
Returns: (str) path to audio file
|
||||
"""
|
||||
return self.save_flagged_file(
|
||||
dir, label, None if data is None else data["data"], encryption_key
|
||||
)
|
||||
|
||||
def generate_sample(self):
|
||||
return test_data.BASE64_AUDIO
|
||||
|
||||
def postprocess(self, y):
|
||||
"""
|
||||
Parameters:
|
||||
y (Union[Tuple[int, numpy.array], str]): audio data in requested format
|
||||
Returns:
|
||||
(str): base64 url data
|
||||
"""
|
||||
if self.type in ["numpy", "file", "auto"]:
|
||||
if self.type == "numpy" or (self.type == "auto" and isinstance(y, tuple)):
|
||||
sample_rate, data = y
|
||||
file = tempfile.NamedTemporaryFile(
|
||||
prefix="sample", suffix=".wav", delete=False
|
||||
)
|
||||
processing_utils.audio_to_file(sample_rate, data, file.name)
|
||||
y = file.name
|
||||
return processing_utils.encode_url_or_file_to_base64(y)
|
||||
else:
|
||||
raise ValueError(
|
||||
"Unknown type: " + self.type + ". Please choose from: 'numpy', 'file'."
|
||||
)
|
||||
|
||||
def deserialize(self, x):
|
||||
return processing_utils.decode_base64_to_file(x).name
|
||||
|
||||
def restore_flagged(self, dir, data, encryption_key):
|
||||
return self.restore_flagged_file(dir, data, encryption_key)["data"]
|
||||
|
||||
240
gradio/inputs.py
240
gradio/inputs.py
@ -28,6 +28,7 @@ from gradio.components import (
|
||||
Radio,
|
||||
Slider,
|
||||
Textbox,
|
||||
Audio
|
||||
)
|
||||
|
||||
if TYPE_CHECKING: # Only import for type checking (is False at runtime).
|
||||
@ -323,6 +324,34 @@ class Video(Component):
|
||||
super().__init__(type=type, source=source, label=label, optional=optional)
|
||||
|
||||
|
||||
class Audio(Audio):
|
||||
"""
|
||||
Component accepts audio input files.
|
||||
Input type: Union[Tuple[int, numpy.array], file-object, numpy.array]
|
||||
Demos: main_note, reverse_audio, spectogram
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
source: str = "upload",
|
||||
type: str = "numpy",
|
||||
label: str = None,
|
||||
optional: bool = False,
|
||||
):
|
||||
"""
|
||||
Parameters:
|
||||
source (str): Source of audio. "upload" creates a box where user can drop an audio file, "microphone" creates a microphone input.
|
||||
type (str): Type of value to be returned by component. "numpy" returns a 2-set tuple with an integer sample_rate and the data numpy.array of shape (samples, 2), "file" returns a temporary file object whose path can be retrieved by file_obj.name, "filepath" returns the path directly.
|
||||
label (str): component name in interface.
|
||||
optional (bool): If True, the interface can be submitted with no uploaded audio, in which case the input value is None.
|
||||
"""
|
||||
warnings.warn(
|
||||
"Usage of gradio.inputs is deprecated, and will not be supported in the future, please import your components from gradio.components",
|
||||
DeprecationWarning,
|
||||
)
|
||||
super().__init__(source=source, type=type, label=label, optional=optional)
|
||||
|
||||
|
||||
class InputComponent(Component):
|
||||
"""
|
||||
Input Component. All input components subclass this.
|
||||
@ -406,217 +435,6 @@ class InputComponent(Component):
|
||||
}
|
||||
|
||||
|
||||
class Audio(InputComponent):
|
||||
"""
|
||||
Component accepts audio input files.
|
||||
Input type: Union[Tuple[int, numpy.array], file-object, numpy.array]
|
||||
Demos: main_note, reverse_audio, spectogram
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
source: str = "upload",
|
||||
type: str = "numpy",
|
||||
label: str = None,
|
||||
optional: bool = False,
|
||||
):
|
||||
"""
|
||||
Parameters:
|
||||
source (str): Source of audio. "upload" creates a box where user can drop an audio file, "microphone" creates a microphone input.
|
||||
type (str): Type of value to be returned by component. "numpy" returns a 2-set tuple with an integer sample_rate and the data numpy.array of shape (samples, 2), "file" returns a temporary file object whose path can be retrieved by file_obj.name, "filepath" returns the path directly.
|
||||
label (str): component name in interface.
|
||||
optional (bool): If True, the interface can be submitted with no uploaded audio, in which case the input value is None.
|
||||
"""
|
||||
self.source = source
|
||||
requires_permissions = source == "microphone"
|
||||
self.type = type
|
||||
self.test_input = test_data.BASE64_AUDIO
|
||||
self.interpret_by_tokens = True
|
||||
super().__init__(label, requires_permissions, optional=optional)
|
||||
|
||||
def get_template_context(self):
|
||||
return {
|
||||
"source": self.source,
|
||||
"optional": self.optional,
|
||||
**super().get_template_context(),
|
||||
}
|
||||
|
||||
@classmethod
|
||||
def get_shortcut_implementations(cls):
|
||||
return {
|
||||
"audio": {},
|
||||
"microphone": {"source": "microphone"},
|
||||
"mic": {"source": "microphone"},
|
||||
}
|
||||
|
||||
def preprocess_example(self, x):
|
||||
return {"name": x, "data": None, "is_example": True}
|
||||
|
||||
def preprocess(self, x: Dict[str, str] | None) -> Tuple[int, np.array] | str | None:
|
||||
"""
|
||||
Parameters:
|
||||
x (Dict[name: str, data: str]): JSON object with filename as 'name' property and base64 data as 'data' property
|
||||
Returns:
|
||||
(Union[Tuple[int, numpy.array], str, numpy.array]): audio in requested format
|
||||
"""
|
||||
if x is None:
|
||||
return x
|
||||
file_name, file_data, is_example = (
|
||||
x["name"],
|
||||
x["data"],
|
||||
x.get("is_example", False),
|
||||
)
|
||||
crop_min, crop_max = x.get("crop_min", 0), x.get("crop_max", 100)
|
||||
if is_example:
|
||||
file_obj = processing_utils.create_tmp_copy_of_file(file_name)
|
||||
else:
|
||||
file_obj = processing_utils.decode_base64_to_file(
|
||||
file_data, file_path=file_name
|
||||
)
|
||||
if crop_min != 0 or crop_max != 100:
|
||||
sample_rate, data = processing_utils.audio_from_file(
|
||||
file_obj.name, crop_min=crop_min, crop_max=crop_max
|
||||
)
|
||||
processing_utils.audio_to_file(sample_rate, data, file_obj.name)
|
||||
if self.type == "file":
|
||||
warnings.warn(
|
||||
"The 'file' type has been deprecated. Set parameter 'type' to 'filepath' instead.",
|
||||
DeprecationWarning,
|
||||
)
|
||||
return file_obj
|
||||
elif self.type == "filepath":
|
||||
return file_obj.name
|
||||
elif self.type == "numpy":
|
||||
return processing_utils.audio_from_file(file_obj.name)
|
||||
else:
|
||||
raise ValueError(
|
||||
"Unknown type: "
|
||||
+ str(self.type)
|
||||
+ ". Please choose from: 'numpy', 'filepath'."
|
||||
)
|
||||
|
||||
def serialize(self, x, called_directly):
|
||||
if x is None:
|
||||
return None
|
||||
if self.type == "filepath" or called_directly:
|
||||
name = x
|
||||
elif self.type == "file":
|
||||
warnings.warn(
|
||||
"The 'file' type has been deprecated. Set parameter 'type' to 'filepath' instead.",
|
||||
DeprecationWarning,
|
||||
)
|
||||
name = x.name
|
||||
elif self.type == "numpy":
|
||||
file = tempfile.NamedTemporaryFile(delete=False)
|
||||
name = file.name
|
||||
processing_utils.audio_to_file(x[0], x[1], name)
|
||||
else:
|
||||
raise ValueError(
|
||||
"Unknown type: "
|
||||
+ str(self.type)
|
||||
+ ". Please choose from: 'numpy', 'filepath'."
|
||||
)
|
||||
|
||||
file_data = processing_utils.encode_url_or_file_to_base64(name)
|
||||
return {"name": name, "data": file_data, "is_example": False}
|
||||
|
||||
def set_interpret_parameters(self, segments=8):
|
||||
"""
|
||||
Calculates interpretation score of audio subsections by splitting the audio into subsections, then using a "leave one out" method to calculate the score of each subsection by removing the subsection and measuring the delta of the output value.
|
||||
Parameters:
|
||||
segments (int): Number of interpretation segments to split audio into.
|
||||
"""
|
||||
self.interpretation_segments = segments
|
||||
return self
|
||||
|
||||
def tokenize(self, x):
|
||||
if x.get("is_example"):
|
||||
sample_rate, data = processing_utils.audio_from_file(x["name"])
|
||||
else:
|
||||
file_obj = processing_utils.decode_base64_to_file(x["data"])
|
||||
sample_rate, data = processing_utils.audio_from_file(file_obj.name)
|
||||
leave_one_out_sets = []
|
||||
tokens = []
|
||||
masks = []
|
||||
duration = data.shape[0]
|
||||
boundaries = np.linspace(0, duration, self.interpretation_segments + 1).tolist()
|
||||
boundaries = [round(boundary) for boundary in boundaries]
|
||||
for index in range(len(boundaries) - 1):
|
||||
start, stop = boundaries[index], boundaries[index + 1]
|
||||
masks.append((start, stop))
|
||||
|
||||
# Handle the leave one outs
|
||||
leave_one_out_data = np.copy(data)
|
||||
leave_one_out_data[start:stop] = 0
|
||||
file = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
|
||||
processing_utils.audio_to_file(sample_rate, leave_one_out_data, file.name)
|
||||
out_data = processing_utils.encode_file_to_base64(file.name)
|
||||
leave_one_out_sets.append(out_data)
|
||||
file.close()
|
||||
os.unlink(file.name)
|
||||
|
||||
# Handle the tokens
|
||||
token = np.copy(data)
|
||||
token[0:start] = 0
|
||||
token[stop:] = 0
|
||||
file = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
|
||||
processing_utils.audio_to_file(sample_rate, token, file.name)
|
||||
token_data = processing_utils.encode_file_to_base64(file.name)
|
||||
file.close()
|
||||
os.unlink(file.name)
|
||||
|
||||
tokens.append(token_data)
|
||||
tokens = [{"name": "token.wav", "data": token} for token in tokens]
|
||||
leave_one_out_sets = [
|
||||
{"name": "loo.wav", "data": loo_set} for loo_set in leave_one_out_sets
|
||||
]
|
||||
return tokens, leave_one_out_sets, masks
|
||||
|
||||
def get_masked_inputs(self, tokens, binary_mask_matrix):
|
||||
# create a "zero input" vector and get sample rate
|
||||
x = tokens[0]["data"]
|
||||
file_obj = processing_utils.decode_base64_to_file(x)
|
||||
sample_rate, data = processing_utils.audio_from_file(file_obj.name)
|
||||
zero_input = np.zeros_like(data, dtype="int16")
|
||||
# decode all of the tokens
|
||||
token_data = []
|
||||
for token in tokens:
|
||||
file_obj = processing_utils.decode_base64_to_file(token["data"])
|
||||
_, data = processing_utils.audio_from_file(file_obj.name)
|
||||
token_data.append(data)
|
||||
# construct the masked version
|
||||
masked_inputs = []
|
||||
for binary_mask_vector in binary_mask_matrix:
|
||||
masked_input = np.copy(zero_input)
|
||||
for t, b in zip(token_data, binary_mask_vector):
|
||||
masked_input = masked_input + t * int(b)
|
||||
file = tempfile.NamedTemporaryFile(delete=False)
|
||||
processing_utils.audio_to_file(sample_rate, masked_input, file.name)
|
||||
masked_data = processing_utils.encode_file_to_base64(file.name)
|
||||
file.close()
|
||||
os.unlink(file.name)
|
||||
masked_inputs.append(masked_data)
|
||||
return masked_inputs
|
||||
|
||||
def get_interpretation_scores(self, x, neighbors, scores, masks=None, tokens=None):
|
||||
"""
|
||||
Returns:
|
||||
(List[float]): Each value represents the interpretation score corresponding to an evenly spaced subsection of audio.
|
||||
"""
|
||||
return list(scores)
|
||||
|
||||
def save_flagged(self, dir, label, data, encryption_key):
|
||||
"""
|
||||
Returns: (str) path to audio file
|
||||
"""
|
||||
return self.save_flagged_file(
|
||||
dir, label, None if data is None else data["data"], encryption_key
|
||||
)
|
||||
|
||||
def generate_sample(self):
|
||||
return test_data.BASE64_AUDIO
|
||||
|
||||
|
||||
class File(InputComponent):
|
||||
"""
|
||||
Component accepts generic file uploads.
|
||||
|
||||
@ -21,7 +21,7 @@ import PIL
|
||||
from ffmpy import FFmpeg
|
||||
|
||||
from gradio import processing_utils
|
||||
from gradio.components import Component, Image, Textbox, Video
|
||||
from gradio.components import Component, Image, Textbox, Video, Audio
|
||||
|
||||
if TYPE_CHECKING: # Only import for type checking (is False at runtime).
|
||||
from gradio import Interface
|
||||
@ -83,6 +83,25 @@ class Video(Video):
|
||||
DeprecationWarning,
|
||||
)
|
||||
super().__init__(label=label, type=type)
|
||||
class Audio(Audio):
|
||||
"""
|
||||
Creates an audio player that plays the output audio.
|
||||
Output type: Union[Tuple[int, numpy.array], str]
|
||||
Demos: generate_tone, reverse_audio
|
||||
"""
|
||||
|
||||
def __init__(self, type: str = "auto", label: Optional[str] = None):
|
||||
"""
|
||||
Parameters:
|
||||
type (str): Type of value to be passed to component. "numpy" returns a 2-set tuple with an integer sample_rate and the data as 16-bit int numpy.array of shape (samples, 2), "file" returns a temporary file path to the saved wav audio file, "auto" detects return type.
|
||||
label (str): component name in interface.
|
||||
"""
|
||||
warnings.warn(
|
||||
"Usage of gradio.outputs is deprecated, and will not be supported in the future, please import your components from gradio.components",
|
||||
DeprecationWarning,
|
||||
)
|
||||
super().__init__(type=type, label=label)
|
||||
|
||||
|
||||
|
||||
class OutputComponent(Component):
|
||||
@ -309,60 +328,6 @@ class HighlightedText(OutputComponent):
|
||||
return json.loads(data)
|
||||
|
||||
|
||||
class Audio(OutputComponent):
|
||||
"""
|
||||
Creates an audio player that plays the output audio.
|
||||
Output type: Union[Tuple[int, numpy.array], str]
|
||||
Demos: generate_tone, reverse_audio
|
||||
"""
|
||||
|
||||
def __init__(self, type: str = "auto", label: Optional[str] = None):
|
||||
"""
|
||||
Parameters:
|
||||
type (str): Type of value to be passed to component. "numpy" returns a 2-set tuple with an integer sample_rate and the data as 16-bit int numpy.array of shape (samples, 2), "file" returns a temporary file path to the saved wav audio file, "auto" detects return type.
|
||||
label (str): component name in interface.
|
||||
"""
|
||||
self.type = type
|
||||
super().__init__(label)
|
||||
|
||||
def get_template_context(self):
|
||||
return {**super().get_template_context()}
|
||||
|
||||
@classmethod
|
||||
def get_shortcut_implementations(cls):
|
||||
return {
|
||||
"audio": {},
|
||||
}
|
||||
|
||||
def postprocess(self, y):
|
||||
"""
|
||||
Parameters:
|
||||
y (Union[Tuple[int, numpy.array], str]): audio data in requested format
|
||||
Returns:
|
||||
(str): base64 url data
|
||||
"""
|
||||
if self.type in ["numpy", "file", "auto"]:
|
||||
if self.type == "numpy" or (self.type == "auto" and isinstance(y, tuple)):
|
||||
sample_rate, data = y
|
||||
file = tempfile.NamedTemporaryFile(
|
||||
prefix="sample", suffix=".wav", delete=False
|
||||
)
|
||||
processing_utils.audio_to_file(sample_rate, data, file.name)
|
||||
y = file.name
|
||||
return processing_utils.encode_url_or_file_to_base64(y)
|
||||
else:
|
||||
raise ValueError(
|
||||
"Unknown type: " + self.type + ". Please choose from: 'numpy', 'file'."
|
||||
)
|
||||
|
||||
def deserialize(self, x):
|
||||
return processing_utils.decode_base64_to_file(x).name
|
||||
|
||||
def save_flagged(self, dir, label, data, encryption_key):
|
||||
return self.save_flagged_file(dir, label, data, encryption_key)
|
||||
|
||||
def restore_flagged(self, dir, data, encryption_key):
|
||||
return self.restore_flagged_file(dir, data, encryption_key)["data"]
|
||||
|
||||
|
||||
class JSON(OutputComponent):
|
||||
|
||||
Loading…
Reference in New Issue
Block a user