Abderrahim-Mouhcine/aesthetic-immich-starter
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Initial upload: aesthetic-immich-starter

Browse Source
This commit is contained in:
Abderrahim-Mouhcine
2026-08-16 19:59:33 +01:00
commit 8938106e99
4777 changed files with 2176899 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
backend/.venv/Lib/site-packages/mediapipe/tasks/python/text/__init__.py
+35
View File
@@ -0,0 +1,35 @@
# Copyright 2022 The MediaPipe Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# 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.
"""MediaPipe Tasks Text API."""
import mediapipe.tasks.python.text.language_detector
import mediapipe.tasks.python.text.text_classifier
import mediapipe.tasks.python.text.text_embedder
LanguageDetector = language_detector.LanguageDetector
LanguageDetectorOptions = language_detector.LanguageDetectorOptions
LanguageDetectorResult = language_detector.LanguageDetectorResult
TextClassifier = text_classifier.TextClassifier
TextClassifierOptions = text_classifier.TextClassifierOptions
TextClassifierResult = text_classifier.TextClassifierResult
TextEmbedder = text_embedder.TextEmbedder
TextEmbedderOptions = text_embedder.TextEmbedderOptions
TextEmbedderResult = text_embedder.TextEmbedderResult
# Remove unnecessary modules to avoid duplication in API docs.
del mediapipe
del language_detector
del text_classifier
del text_embedder
backend/.venv/Lib/site-packages/mediapipe/tasks/python/text/language_detector.py
+290
View File
@@ -0,0 +1,290 @@
# Copyright 2023 The MediaPipe Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# 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.
"""MediaPipe language detector task."""
import ctypes
import dataclasses
from typing import List, Optional
from mediapipe.tasks.python.components.containers import category as category_module
from mediapipe.tasks.python.components.containers import classification_result as classification_result_module
from mediapipe.tasks.python.components.processors import classifier_options as classifier_options_module
from mediapipe.tasks.python.components.processors import classifier_options_c as classifier_options_c_module
from mediapipe.tasks.python.core import base_options as base_options_module
from mediapipe.tasks.python.core import base_options_c as base_options_c_module
from mediapipe.tasks.python.core import mediapipe_c_bindings as mediapipe_c_bindings_module
from mediapipe.tasks.python.core import mediapipe_c_utils
from mediapipe.tasks.python.core import serial_dispatcher
_BaseOptions = base_options_module.BaseOptions
Category = category_module.Category
Classifications = classification_result_module.Classifications
ClassifierOptions = classifier_options_module.ClassifierOptions
@dataclasses.dataclass
class LanguageDetectorResult:
@dataclasses.dataclass
class Detection:
"""A language code and its probability."""
# An i18n language / locale code, e.g. "en" for English, "uz" for Uzbek,
# "ja"-Latn for Japanese (romaji).
language_code: str
probability: float
detections: List[Detection]
class LanguageDetectorPredictionC(ctypes.Structure):
"""A language code and its probability."""
_fields_ = [
("language_code", ctypes.c_char_p),
("probability", ctypes.c_float),
]
class LanguageDetectorResultC(ctypes.Structure):
"""Language detector output."""
_fields_ = [
("predictions", ctypes.POINTER(LanguageDetectorPredictionC)),
("predictions_count", ctypes.c_uint32),
]
def _convert_to_python_language_detector_result(
c_result: LanguageDetectorResultC,
) -> LanguageDetectorResult:
"""Converts a C LanguageDetectorResult to a Python LanguageDetectorResult."""
py_result = LanguageDetectorResult(detections=[])
for i in range(c_result.predictions_count):
c_prediction = c_result.predictions[i]
py_prediction = LanguageDetectorResult.Detection(
language_code=c_prediction.language_code.decode("utf-8"),
probability=c_prediction.probability,
)
py_result.detections.append(py_prediction)
return py_result
class LanguageDetectorOptionsC(ctypes.Structure):
_fields_ = [
("base_options", base_options_c_module.BaseOptionsC),
("classifier_options", classifier_options_c_module.ClassifierOptionsC),
]
@dataclasses.dataclass
class LanguageDetectorOptions:
"""Options for the language detector task.
Attributes:
base_options: Base options for the language detector task.
display_names_locale: The locale to use for display names specified through
the TFLite Model Metadata.
max_results: The maximum number of top-scored classification results to
return.
score_threshold: Overrides the ones provided in the model metadata. Results
below this value are rejected.
category_allowlist: Allowlist of category names. If non-empty,
classification results whose category name is not in this set will be
filtered out. Duplicate or unknown category names are ignored. Mutually
exclusive with `category_denylist`.
category_denylist: Denylist of category names. If non-empty, classification
results whose category name is in this set will be filtered out. Duplicate
or unknown category names are ignored. Mutually exclusive with
`category_allowlist`.
"""
base_options: _BaseOptions
display_names_locale: Optional[str] = None
max_results: Optional[int] = None
score_threshold: Optional[float] = None
category_allowlist: Optional[List[str]] = None
category_denylist: Optional[List[str]] = None
def to_ctypes(self) -> LanguageDetectorOptionsC:
"""Generates a ctypes LanguageDetectorOptionsC."""
base_options_c = self.base_options.to_ctypes()
classifier_options_c = (
classifier_options_c_module.convert_to_classifier_options_c(
classifier_options_module.ClassifierOptions(
display_names_locale=self.display_names_locale,
max_results=self.max_results,
score_threshold=self.score_threshold,
category_allowlist=self.category_allowlist,
category_denylist=self.category_denylist,
)
)
)
c_options = LanguageDetectorOptionsC()
c_options.base_options = base_options_c
c_options.classifier_options = classifier_options_c
return c_options
_CTYPES_SIGNATURES = (
mediapipe_c_utils.CStatusFunction(
"MpLanguageDetectorCreate",
(
ctypes.POINTER(LanguageDetectorOptionsC),
ctypes.POINTER(ctypes.c_void_p),
),
),
mediapipe_c_utils.CStatusFunction(
"MpLanguageDetectorDetect",
(
ctypes.c_void_p,
ctypes.c_char_p,
ctypes.POINTER(LanguageDetectorResultC),
),
),
mediapipe_c_utils.CStatusFunction(
"MpLanguageDetectorClose",
(
ctypes.c_void_p,
),
),
mediapipe_c_utils.CFunction(
"MpLanguageDetectorCloseResult",
[ctypes.POINTER(LanguageDetectorResultC)],
None,
),
)
class LanguageDetector:
"""Class that predicts the language of an input text.
This API expects a TFLite model with TFLite Model Metadata that contains the
mandatory (described below) input tensors, output tensor, and the language
codes in an AssociatedFile.
Input tensors:
(kTfLiteString)
- 1 input tensor that is scalar or has shape [1] containing the input
string.
Output tensor:
(kTfLiteFloat32)
- 1 output tensor of shape`[1 x N]` where `N` is the number of languages.
"""
_lib: serial_dispatcher.SerialDispatcher
_handle: ctypes.c_void_p
def __init__(
self, lib: serial_dispatcher.SerialDispatcher, handle: ctypes.c_void_p
):
self._lib = lib
self._detector_handle = handle
@classmethod
def create_from_model_path(cls, model_path: str) -> "LanguageDetector":
"""Creates an `LanguageDetector` object from a TensorFlow Lite model and the default `LanguageDetectorOptions`.
Args:
model_path: Path to the model.
Returns:
`LanguageDetector` object that's created from the model file and the
default `LanguageDetectorOptions`.
Raises:
ValueError: If failed to create `LanguageDetector` object from the
provided
file such as invalid file path.
RuntimeError: If other types of error occurred.
"""
return cls.create_from_options(
LanguageDetectorOptions(
base_options=_BaseOptions(model_asset_path=model_path)
)
)
@classmethod
def create_from_options(
cls, options: LanguageDetectorOptions
) -> "LanguageDetector":
"""Creates the `LanguageDetector` object from language detector options.
Args:
options: Options for the language detector task.
Returns:
`LanguageDetector` object that's created from `options`.
Raises:
ValueError: If failed to create `LanguageDetector` object from
`LanguageDetectorOptions` such as missing the model.
RuntimeError: If other types of error occurred.
"""
lib = mediapipe_c_bindings_module.load_shared_library(_CTYPES_SIGNATURES)
ctypes_options = options.to_ctypes()
detector_handle = ctypes.c_void_p()
lib.MpLanguageDetectorCreate(
ctypes.byref(ctypes_options),
ctypes.byref(detector_handle),
)
return LanguageDetector(lib=lib, handle=detector_handle)
def detect(self, text: str) -> LanguageDetectorResult:
"""Predicts the language of the input `text`.
Args:
text: The input text.
Returns:
A `LanguageDetectorResult` object that contains a list of languages and
scores.
Raises:
ValueError: If any of the input arguments is invalid.
RuntimeError: If language detection failed to run.
"""
ctypes_result = LanguageDetectorResultC()
self._lib.MpLanguageDetectorDetect(
self._detector_handle,
text.encode("utf-8"),
ctypes.byref(ctypes_result),
)
python_result = _convert_to_python_language_detector_result(ctypes_result)
self._lib.MpLanguageDetectorCloseResult(ctypes.byref(ctypes_result))
return python_result
def close(self):
"""Shuts down the MediaPipe task instance."""
if not self._detector_handle:
return
self._lib.MpLanguageDetectorClose(self._detector_handle)
self._detector_handle = None
self._lib.close()
def __enter__(self):
"""Returns `self` upon entering the runtime context."""
return self
def __exit__(self, unused_exc_type, unused_exc_value, unused_traceback):
"""Shuts down the MediaPipe task instance on exit of the context manager.
Raises:
RuntimeError: If the MediaPipe LanguageDetector task failed to close.
"""
self.close()
def __del__(self):
self.close()
backend/.venv/Lib/site-packages/mediapipe/tasks/python/text/text_classifier.py
+264
View File
@@ -0,0 +1,264 @@
# Copyright 2022 The MediaPipe Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# 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.
"""MediaPipe text classifier task."""
import ctypes
import dataclasses
from typing import List, Optional
from mediapipe.tasks.python.components.containers import category as category_module
from mediapipe.tasks.python.components.containers import classification_result as classification_result_module
from mediapipe.tasks.python.components.containers import classification_result_c as classification_result_c_module
from mediapipe.tasks.python.components.processors import classifier_options as classifier_options_module
from mediapipe.tasks.python.components.processors import classifier_options_c as classifier_options_c_module
from mediapipe.tasks.python.core import base_options as base_options_module
from mediapipe.tasks.python.core import base_options_c as base_options_c_module
from mediapipe.tasks.python.core import mediapipe_c_bindings
from mediapipe.tasks.python.core import mediapipe_c_utils
from mediapipe.tasks.python.core import serial_dispatcher
_BaseOptions = base_options_module.BaseOptions
Category = category_module.Category
Classifications = classification_result_module.Classifications
TextClassifierResult = classification_result_module.ClassificationResult
ClassifierOptions = classifier_options_module.ClassifierOptions
class TextClassifierOptionsC(ctypes.Structure):
_fields_ = [
("base_options", base_options_c_module.BaseOptionsC),
("classifier_options", classifier_options_c_module.ClassifierOptionsC),
]
@dataclasses.dataclass
class TextClassifierOptions:
"""Options for the text classifier task.
Attributes:
base_options: Base options for the text classifier task.
display_names_locale: The locale to use for display names specified through
the TFLite Model Metadata.
max_results: The maximum number of top-scored classification results to
return.
score_threshold: Overrides the ones provided in the model metadata. Results
below this value are rejected.
category_allowlist: Allowlist of category names. If non-empty,
classification results whose category name is not in this set will be
filtered out. Duplicate or unknown category names are ignored. Mutually
exclusive with `category_denylist`.
category_denylist: Denylist of category names. If non-empty, classification
results whose category name is in this set will be filtered out. Duplicate
or unknown category names are ignored. Mutually exclusive with
`category_allowlist`.
"""
base_options: _BaseOptions
display_names_locale: Optional[str] = None
max_results: Optional[int] = None
score_threshold: Optional[float] = None
category_allowlist: Optional[List[str]] = None
category_denylist: Optional[List[str]] = None
def to_ctypes(self) -> TextClassifierOptionsC:
"""Generates a ctypes TextClassifierOptionsC."""
base_options_c = self.base_options.to_ctypes()
classifier_options_c = (
classifier_options_c_module.convert_to_classifier_options_c(
classifier_options_module.ClassifierOptions(
display_names_locale=self.display_names_locale,
max_results=self.max_results,
score_threshold=self.score_threshold,
category_allowlist=self.category_allowlist,
category_denylist=self.category_denylist,
)
)
)
c_options = TextClassifierOptionsC()
c_options.base_options = base_options_c
c_options.classifier_options = classifier_options_c
return c_options
_CTYPES_SIGNATURES = (
mediapipe_c_utils.CStatusFunction(
"MpTextClassifierCreate",
[
ctypes.POINTER(TextClassifierOptionsC),
ctypes.POINTER(ctypes.c_void_p),
],
),
mediapipe_c_utils.CStatusFunction(
"MpTextClassifierClassify",
[
ctypes.c_void_p,
ctypes.c_char_p,
ctypes.POINTER(
classification_result_c_module.ClassificationResultC
),
],
),
mediapipe_c_utils.CStatusFunction(
"MpTextClassifierClose",
[
ctypes.c_void_p,
],
),
mediapipe_c_utils.CFunction(
"MpTextClassifierCloseResult",
[ctypes.POINTER(classification_result_c_module.ClassificationResultC)],
None,
),
)
class TextClassifier:
"""Class that performs classification on text.
This API expects a TFLite model with (optional) TFLite Model Metadata that
contains the mandatory (described below) input tensors, output tensor,
and the optional (but recommended) category labels as AssociatedFiles with
type
TENSOR_AXIS_LABELS per output classification tensor. Metadata is required for
models with int32 input tensors because it contains the input process unit
for the model's Tokenizer. No metadata is required for models with string
input tensors.
Input tensors:
(kTfLiteInt32)
- 3 input tensors of size `[batch_size x bert_max_seq_len]` representing
the input ids, segment ids, and mask ids
- or 1 input tensor of size `[batch_size x max_seq_len]` representing the
input ids
or (kTfLiteString)
- 1 input tensor that is shapeless or has shape [1] containing the input
string
At least one output tensor with:
(kTfLiteFloat32/kBool)
- `[1 x N]` array with `N` represents the number of categories.
- optional (but recommended) category labels as AssociatedFiles with type
TENSOR_AXIS_LABELS, containing one label per line. The first such
AssociatedFile (if any) is used to fill the `category_name` field of the
results. The `display_name` field is filled from the AssociatedFile (if
any) whose locale matches the `display_names_locale` field of the
`TextClassifierOptions` used at creation time ("en" by default, i.e.
English). If none of these are available, only the `index` field of the
results will be filled.
"""
_lib: serial_dispatcher.SerialDispatcher
_handle: ctypes.c_void_p
def __init__(
self,
lib: serial_dispatcher.SerialDispatcher,
handle: ctypes.c_void_p,
):
self._lib = lib
self._classifier_handle = handle
@classmethod
def create_from_model_path(cls, model_path: str) -> "TextClassifier":
"""Creates an `TextClassifier` object from a TensorFlow Lite model and the default `TextClassifierOptions`.
Args:
model_path: Path to the model.
Returns:
`TextClassifier` object that's created from the model file and the
default `TextClassifierOptions`.
Raises:
ValueError: If failed to create `TextClassifier` object from the provided
file such as invalid file path.
RuntimeError: If other types of error occurred.
"""
return cls.create_from_options(
TextClassifierOptions(
base_options=_BaseOptions(model_asset_path=model_path)
)
)
@classmethod
def create_from_options(
cls, options: TextClassifierOptions
) -> "TextClassifier":
"""Creates the `TextClassifier` object from text classifier options.
Args:
options: Options for the text classifier task.
Returns:
`TextClassifier` object that's created from `options`.
Raises:
ValueError: If failed to create `TextClassifier` object from
`TextClassifierOptions` such as missing the model.
RuntimeError: If other types of error occurred.
"""
lib = mediapipe_c_bindings.load_shared_library(_CTYPES_SIGNATURES)
ctypes_options = options.to_ctypes()
classifier_handle = ctypes.c_void_p()
lib.MpTextClassifierCreate(
ctypes.byref(ctypes_options),
ctypes.byref(classifier_handle),
)
return TextClassifier(lib=lib, handle=classifier_handle)
def classify(self, text: str) -> TextClassifierResult:
"""Performs classification on the input `text`.
Args:
text: The input text.
Returns:
A `TextClassifierResult` object that contains a list of text
classifications.
Raises:
ValueError: If any of the input arguments is invalid.
RuntimeError: If text classification failed to run.
"""
ctypes_result = classification_result_c_module.ClassificationResultC()
self._lib.MpTextClassifierClassify(
self._classifier_handle,
text.encode("utf-8"),
ctypes.byref(ctypes_result),
)
python_result = TextClassifierResult.from_ctypes(ctypes_result)
self._lib.MpTextClassifierCloseResult(ctypes.byref(ctypes_result))
return python_result
def close(self):
"""Shuts down the MediaPipe task instance."""
if self._classifier_handle:
self._lib.MpTextClassifierClose(self._classifier_handle)
self._classifier_handle = None
self._lib.close()
def __enter__(self):
"""Returns `self` upon entering the runtime context."""
return self
def __exit__(self, unused_exc_type, unused_exc_value, unused_traceback):
"""Shuts down the MediaPipe task instance on exit of the context manager.
Raises:
RuntimeError: If the MediaPipe TextClassifier task failed to close.
"""
self.close()
def __del__(self):
self.close()
backend/.venv/Lib/site-packages/mediapipe/tasks/python/text/text_embedder.py
+260
View File
@@ -0,0 +1,260 @@
# Copyright 2022 The MediaPipe Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# 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.
"""MediaPipe text embedder task."""
import ctypes
import dataclasses
from typing import Optional
from mediapipe.tasks.python.components.containers import embedding_result as embedding_result_module
from mediapipe.tasks.python.components.containers import embedding_result_c as embedding_result_c_module
from mediapipe.tasks.python.components.utils import cosine_similarity
from mediapipe.tasks.python.core import base_options as base_options_module
from mediapipe.tasks.python.core import base_options_c as base_options_c_module
from mediapipe.tasks.python.core import mediapipe_c_bindings
from mediapipe.tasks.python.core import mediapipe_c_utils
from mediapipe.tasks.python.core import serial_dispatcher
TextEmbedderResult = embedding_result_module.EmbeddingResult
class _EmbedderOptionsC(ctypes.Structure):
"""C struct for embedder options."""
_fields_ = [
('l2_normalize', ctypes.c_bool),
('quantize', ctypes.c_bool),
]
class _TextEmbedderOptionsC(ctypes.Structure):
"""C struct for text embedder options."""
_fields_ = [
('base_options', base_options_c_module.BaseOptionsC),
('embedder_options', _EmbedderOptionsC),
]
_CTYPES_SIGNATURES = (
mediapipe_c_utils.CStatusFunction(
'MpTextEmbedderCreate',
[
ctypes.POINTER(_TextEmbedderOptionsC),
ctypes.POINTER(ctypes.c_void_p),
],
),
mediapipe_c_utils.CStatusFunction(
'MpTextEmbedderEmbed',
[
ctypes.c_void_p,
ctypes.c_char_p,
ctypes.POINTER(embedding_result_c_module.EmbeddingResultC),
],
),
mediapipe_c_utils.CStatusFunction(
'MpTextEmbedderClose',
[
ctypes.c_void_p,
],
),
mediapipe_c_utils.CFunction(
'MpTextEmbedderCloseResult',
[ctypes.POINTER(embedding_result_c_module.EmbeddingResultC)],
None,
),
)
@dataclasses.dataclass
class TextEmbedderOptions:
"""Options for the text embedder task.
Attributes:
base_options: Base options for the text embedder task.
l2_normalize: Whether to normalize the returned feature vector with L2 norm.
Use this option only if the model does not already contain a native
L2_NORMALIZATION TF Lite Op. In most cases, this is already the case and
L2 norm is thus achieved through TF Lite inference.
quantize: Whether the returned embedding should be quantized to bytes via
scalar quantization. Embeddings are implicitly assumed to be unit-norm and
therefore any dimension is guaranteed to have a value in [-1.0, 1.0]. Use
the l2_normalize option if this is not the case.
"""
base_options: base_options_module.BaseOptions
l2_normalize: Optional[bool] = None
quantize: Optional[bool] = None
def to_ctypes(self) -> _TextEmbedderOptionsC:
"""Generates a ctypes TextEmbedderOptionsC object."""
base_options_c = self.base_options.to_ctypes()
embedder_options_c = _EmbedderOptionsC(
l2_normalize=self.l2_normalize
if self.l2_normalize is not None
else False,
quantize=self.quantize if self.quantize is not None else False,
)
return _TextEmbedderOptionsC(
base_options=base_options_c, embedder_options=embedder_options_c
)
class TextEmbedder:
"""Class that performs embedding extraction on text.
This API expects a TFLite model with TFLite Model Metadata that contains the
mandatory (described below) input tensors and output tensors. Metadata should
contain the input process unit for the model's Tokenizer as well as input /
output tensor metadata.
Input tensors:
(kTfLiteInt32)
- 3 input tensors of size `[batch_size x bert_max_seq_len]` with names
"ids", "mask", and "segment_ids" representing the input ids, mask ids, and
segment ids respectively.
- or 1 input tensor of size `[batch_size x max_seq_len]` representing the
input ids.
At least one output tensor with:
(kTfLiteFloat32)
- `N` components corresponding to the `N` dimensions of the returned
feature vector for this output layer.
- Either 2 or 4 dimensions, i.e. `[1 x N]` or `[1 x 1 x 1 x N]`.
"""
_lib: serial_dispatcher.SerialDispatcher
_handle: ctypes.c_void_p
def __init__(
self,
lib: serial_dispatcher.SerialDispatcher,
handle: ctypes.c_void_p,
):
self._lib = lib
self._embedder_handle = handle
@classmethod
def create_from_model_path(cls, model_path: str) -> 'TextEmbedder':
"""Creates an `TextEmbedder` object from a TensorFlow Lite model and the default `TextEmbedderOptions`.
Args:
model_path: Path to the model.
Returns:
`TextEmbedder` object that's created from the model file and the default
`TextEmbedderOptions`.
Raises:
ValueError: If failed to create `TextEmbedder` object from the provided
file such as invalid file path.
RuntimeError: If other types of error occurred.
"""
base_options = base_options_module.BaseOptions(model_asset_path=model_path)
options = TextEmbedderOptions(base_options=base_options)
return cls.create_from_options(options)
@classmethod
def create_from_options(cls, options: TextEmbedderOptions) -> 'TextEmbedder':
"""Creates the `TextEmbedder` object from text embedder options.
Args:
options: Options for the text embedder task.
Returns:
`TextEmbedder` object that's created from `options`.
Raises:
ValueError: If failed to create `TextEmbedder` object from
`TextEmbedderOptions` such as missing the model.
RuntimeError: If other types of error occurred.
"""
lib = mediapipe_c_bindings.load_shared_library(_CTYPES_SIGNATURES)
ctypes_options = options.to_ctypes()
embedder_handle = ctypes.c_void_p()
lib.MpTextEmbedderCreate(
ctypes.byref(ctypes_options), ctypes.byref(embedder_handle)
)
return TextEmbedder(lib=lib, handle=embedder_handle)
def embed(
self,
text: str,
) -> embedding_result_module.EmbeddingResult:
"""Performs text embedding extraction on the provided text.
Args:
text: The input text.
Returns:
An embedding result object that contains a list of embeddings.
Raises:
ValueError: If any of the input arguments is invalid.
RuntimeError: If text embedder failed to run.
"""
ctypes_result = embedding_result_c_module.EmbeddingResultC()
self._lib.MpTextEmbedderEmbed(
self._embedder_handle,
text.encode('utf-8'),
ctypes.byref(ctypes_result),
)
python_result = (
embedding_result_module.EmbeddingResult.from_ctypes(
ctypes_result
)
)
self._lib.MpTextEmbedderCloseResult(ctypes.byref(ctypes_result))
return python_result
@classmethod
def cosine_similarity(
cls,
u: embedding_result_module.Embedding,
v: embedding_result_module.Embedding,
) -> float:
"""Utility function to compute cosine similarity between two embedding entries.
May return an InvalidArgumentError if e.g. the feature vectors are
of different types (quantized vs. float), have different sizes, or have a
an L2-norm of 0.
Args:
u: An embedding entry.
v: An embedding entry.
Returns:
The cosine similarity for the two embeddings.
Raises:
ValueError: May return an error if e.g. the feature vectors are of
different types (quantized vs. float), have different sizes, or have
an L2-norm of 0.
"""
return cosine_similarity.cosine_similarity(u, v)
def close(self):
"""Shuts down the MediaPipe task instance."""
if self._embedder_handle:
self._lib.MpTextEmbedderClose(self._embedder_handle)
self._embedder_handle = None
self._lib.close()
def __enter__(self):
"""Returns `self` upon entering the runtime context."""
return self
def __exit__(self, unused_exc_type, unused_exc_value, unused_traceback):
"""Shuts down the MediaPipe task instance on exit of the context manager."""
self.close()
def __del__(self):
self.close()