spark-nlp 2.6.3rc1__py2.py3-none-any.whl → 6.2.1__py2.py3-none-any.whl

sparknlp/annotator/embeddings/distil_bert_embeddings.py

@@ -0,0 +1,221 @@
+#  Copyright 2017-2022 John Snow Labs
+#
+#  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.
+"""Contains classes for DistilBertEmbeddings."""
+
+from sparknlp.common import *
+
+
+class DistilBertEmbeddings(AnnotatorModel,
+                           HasEmbeddingsProperties,
+                           HasCaseSensitiveProperties,
+                           HasStorageRef,
+                           HasBatchedAnnotate,
+                           HasEngine,
+                           HasMaxSentenceLengthLimit):
+    """DistilBERT is a small, fast, cheap and light Transformer model trained by
+    distilling BERT base. It has 40% less parameters than ``bert-base-uncased``,
+    runs 60% faster while preserving over 95% of BERT's performances as measured
+    on the GLUE language understanding benchmark.
+
+    Pretrained models can be loaded with :meth:`.pretrained` of the companion
+    object:
+
+    >>> embeddings = DistilBertEmbeddings.pretrained() \\
+    ...     .setInputCols(["document", "token"]) \\
+    ...     .setOutputCol("embeddings")
+
+
+    The default model is ``"distilbert_base_cased"``, if no name is provided.
+    For available pretrained models please see the
+    `Models Hub <https://sparknlp.org/models?task=Embeddings>`__.
+
+    For extended examples of usage, see the `Examples
+    <https://github.com/JohnSnowLabs/spark-nlp/blob/master/examples/python/transformers/HuggingFace%20in%20Spark%20NLP%20-%20DistilBERT.ipynb>`__.
+    To see which models are compatible and how to import them see
+    `Import Transformers into Spark NLP 🚀
+    <https://github.com/JohnSnowLabs/spark-nlp/discussions/5669>`_.
+
+    ====================== ======================
+    Input Annotation types Output Annotation type
+    ====================== ======================
+    ``DOCUMENT, TOKEN``    ``WORD_EMBEDDINGS``
+    ====================== ======================
+
+    Parameters
+    ----------
+    batchSize
+        Size of every batch, by default 8
+    dimension
+        Number of embedding dimensions, by default 768
+    caseSensitive
+        Whether to ignore case in tokens for embeddings matching, by default
+        False
+    maxSentenceLength
+        Max sentence length to process, by default 128
+    configProtoBytes
+        ConfigProto from tensorflow, serialized into byte array.
+
+    Notes
+    -----
+    - DistilBERT doesn't have ``token_type_ids``, you don't need to
+      indicate which token belongs to which segment. Just separate your segments
+      with the separation token ``tokenizer.sep_token`` (or ``[SEP]``).
+    - DistilBERT doesn't have options to select the input positions
+      (``position_ids`` input). This could be added if necessary though,
+      just let us know if you need this option.
+
+    References
+    ----------
+    The DistilBERT model was proposed in the paper
+    `DistilBERT, a distilled version of BERT: smaller, faster, cheaper and
+    lighter <https://arxiv.org/abs/1910.01108>`__.
+
+    **Paper Abstract:**
+
+    *As Transfer Learning from large-scale pre-trained models becomes more
+    prevalent  in Natural Language Processing (NLP), operating these
+    large models in on-the- edge and/or under constrained computational
+    training or inference budgets  remains challenging. In this work, we
+    propose a method to pre-train a smaller  general-purpose language
+    representation model, called DistilBERT, which can  then be
+    fine-tuned with good performances on a wide range of tasks like its
+    larger counterparts. While most prior work investigated the use of
+    distillation  for building task-specific models, we leverage
+    knowledge distillation during  the pretraining phase and show that it
+    is possible to reduce the size of a BERT  model by 40%, while
+    retaining 97% of its language understanding capabilities  and being
+    60% faster. To leverage the inductive biases learned by larger
+    models  during pretraining, we introduce a triple loss combining
+    language modeling,  distillation and cosine-distance losses. Our
+    smaller, faster and lighter model  is cheaper to pre-train and we
+    demonstrate its capabilities for on-device  computations in a
+    proof-of-concept experiment and a comparative on-device study.*
+
+    Examples
+    --------
+    >>> import sparknlp
+    >>> from sparknlp.base import *
+    >>> from sparknlp.annotator import *
+    >>> from pyspark.ml import Pipeline
+    >>> documentAssembler = DocumentAssembler() \\
+    ...     .setInputCol("text") \\
+    ...     .setOutputCol("document")
+    >>> tokenizer = Tokenizer() \\
+    ...     .setInputCols(["document"]) \\
+    ...     .setOutputCol("token")
+    >>> embeddings = DistilBertEmbeddings.pretrained() \\
+    ...     .setInputCols(["document", "token"]) \\
+    ...     .setOutputCol("embeddings") \\
+    ...     .setCaseSensitive(True)
+    >>> embeddingsFinisher = EmbeddingsFinisher() \\
+    ...     .setInputCols(["embeddings"]) \\
+    ...     .setOutputCols("finished_embeddings") \\
+    ...     .setOutputAsVector(True) \\
+    ...     .setCleanAnnotations(False)
+    >>> pipeline = Pipeline() \\
+    ...     .setStages([
+    ...       documentAssembler,
+    ...       tokenizer,
+    ...       embeddings,
+    ...       embeddingsFinisher
+    ...     ])
+    >>> data = spark.createDataFrame([["This is a sentence."]]).toDF("text")
+    >>> result = pipeline.fit(data).transform(data)
+    >>> result.selectExpr("explode(finished_embeddings) as result").show(5, 80)
+    +--------------------------------------------------------------------------------+
+    |                                                                          result|
+    +--------------------------------------------------------------------------------+
+    |[0.1127224713563919,-0.1982710212469101,0.5360898375511169,-0.272536993026733...|
+    |[0.35534414649009705,0.13215228915214539,0.40981462597846985,0.14036104083061...|
+    |[0.328085333108902,-0.06269335001707077,-0.017595693469047546,-0.024373905733...|
+    |[0.15617232024669647,0.2967822253704071,0.22324979305267334,-0.04568954557180...|
+    |[0.45411425828933716,0.01173491682857275,0.190129816532135,0.1178255230188369...|
+    +--------------------------------------------------------------------------------+
+    """
+
+    name = "DistilBertEmbeddings"
+
+    inputAnnotatorTypes = [AnnotatorType.DOCUMENT, AnnotatorType.TOKEN]
+
+    outputAnnotatorType = AnnotatorType.WORD_EMBEDDINGS
+
+    configProtoBytes = Param(Params._dummy(),
+                             "configProtoBytes",
+                             "ConfigProto from tensorflow, serialized into byte array. Get with config_proto.SerializeToString()",
+                             TypeConverters.toListInt)
+
+    def setConfigProtoBytes(self, b):
+        """Sets configProto from tensorflow, serialized into byte array.
+
+        Parameters
+        ----------
+        b : List[int]
+            ConfigProto from tensorflow, serialized into byte array
+        """
+        return self._set(configProtoBytes=b)
+
+    @keyword_only
+    def __init__(self, classname="com.johnsnowlabs.nlp.embeddings.DistilBertEmbeddings", java_model=None):
+        super(DistilBertEmbeddings, self).__init__(
+            classname=classname,
+            java_model=java_model
+        )
+        self._setDefault(
+            dimension=768,
+            batchSize=8,
+            maxSentenceLength=128,
+            caseSensitive=False
+        )
+
+    @staticmethod
+    def loadSavedModel(folder, spark_session):
+        """Loads a locally saved model.
+
+        Parameters
+        ----------
+        folder : str
+            Folder of the saved model
+        spark_session : pyspark.sql.SparkSession
+            The current SparkSession
+
+        Returns
+        -------
+        DistilBertEmbeddings
+            The restored model
+        """
+        from sparknlp.internal import _DistilBertLoader
+        jModel = _DistilBertLoader(folder, spark_session._jsparkSession)._java_obj
+        return DistilBertEmbeddings(java_model=jModel)
+
+    @staticmethod
+    def pretrained(name="distilbert_base_cased", lang="en", remote_loc=None):
+        """Downloads and loads a pretrained model.
+
+        Parameters
+        ----------
+        name : str, optional
+            Name of the pretrained model, by default "distilbert_base_cased"
+        lang : str, optional
+            Language of the pretrained model, by default "en"
+        remote_loc : str, optional
+            Optional remote address of the resource, by default None. Will use
+            Spark NLPs repositories otherwise.
+
+        Returns
+        -------
+        DistilBertEmbeddings
+            The restored model
+        """
+        from sparknlp.pretrained import ResourceDownloader
+        return ResourceDownloader.downloadModel(DistilBertEmbeddings, name, lang, remote_loc)

sparknlp/annotator/embeddings/doc2vec.py

@@ -0,0 +1,352 @@
+#  Copyright 2017-2022 John Snow Labs
+#
+#  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.
+"""Contains classes for Doc2Vec."""
+
+from sparknlp.common import *
+
+
+class Doc2VecApproach(AnnotatorApproach, HasStorageRef, HasEnableCachingProperties):
+    """Trains a Word2Vec model that creates vector representations of words in a
+    text corpus.
+
+    The algorithm first constructs a vocabulary from the corpus and then learns
+    vector representation of words in the vocabulary. The vector representation
+    can be used as features in natural language processing and machine learning
+    algorithms.
+
+    We use Word2Vec implemented in Spark ML. It uses skip-gram model in our
+    implementation and a hierarchical softmax method to train the model. The
+    variable names in the implementation match the original C implementation.
+
+    For instantiated/pretrained models, see :class:`.Doc2VecModel`.
+
+    For available pretrained models please see the `Models Hub <https://sparknlp.org/models>`__.
+
+    ====================== =======================
+    Input Annotation types Output Annotation type
+    ====================== =======================
+    ``TOKEN``              ``SENTENCE_EMBEDDINGS``
+    ====================== =======================
+
+    Parameters
+    ----------
+    vectorSize
+        The dimension of codes after transforming from words (> 0), by default
+        100
+    windowSize
+        The window size (context words from [-window, window]) (> 0), by default
+        5
+    numPartitions
+        Number of partitions for sentences of words (> 0), by default 1
+    minCount
+        The minimum number of times a token must appear to be included in the
+        word2vec model's vocabulary (>= 0), by default 1
+    maxSentenceLength
+        The window size (Maximum length (in words) of each sentence in the input
+        data. Any sentence longer than this threshold will be divided into
+        chunks up to the size (> 0), by default 1000
+    stepSize
+        Step size (learning rate) to be used for each iteration of optimization
+        (> 0), by default 0.025
+    maxIter
+        Maximum number of iterations (>= 0), by default 1
+    seed
+        Random seed, by default 44
+
+
+    References
+    ----------
+    For the original C implementation, see https://code.google.com/p/word2vec/
+
+    For the research paper, see `Efficient Estimation of Word Representations in
+    Vector Space <https://arxiv.org/abs/1301.3781>`__ and `Distributed
+    Representations of Words and Phrases and their Compositionality
+    <https://arxiv.org/pdf/1310.4546v1.pdf>`__.
+
+    Examples
+    --------
+    >>> import sparknlp
+    >>> from sparknlp.base import *
+    >>> from sparknlp.annotator import *
+    >>> from pyspark.ml import Pipeline
+    >>> documentAssembler = DocumentAssembler() \\
+    ...     .setInputCol("text") \\
+    ...     .setOutputCol("document")
+    >>> tokenizer = Tokenizer() \\
+    ...     .setInputCols(["document"]) \\
+    ...     .setOutputCol("token")
+    >>> embeddings = Doc2VecApproach() \\
+    ...     .setInputCols(["token"]) \\
+    ...     .setOutputCol("embeddings")
+    >>> pipeline = Pipeline() \\
+    ...     .setStages([
+    ...       documentAssembler,
+    ...       tokenizer,
+    ...       embeddings
+    ...     ])
+    >>> path = "sherlockholmes.txt"
+    >>> dataset = spark.read.text(path).toDF("text")
+    >>> pipelineModel = pipeline.fit(dataset)
+    """
+    inputAnnotatorTypes = [AnnotatorType.TOKEN]
+
+    outputAnnotatorType = AnnotatorType.SENTENCE_EMBEDDINGS
+
+    vectorSize = Param(Params._dummy(),
+                       "vectorSize",
+                       "the dimension of codes after transforming from words (> 0)",
+                       typeConverter=TypeConverters.toInt)
+
+    windowSize = Param(Params._dummy(),
+                       "windowSize",
+                       "the window size (context words from [-window, window]) (> 0)",
+                       typeConverter=TypeConverters.toInt)
+
+    numPartitions = Param(Params._dummy(),
+                          "numPartitions",
+                          "number of partitions for sentences of words (> 0)",
+                          typeConverter=TypeConverters.toInt)
+
+    minCount = Param(Params._dummy(),
+                     "minCount",
+                     "the minimum number of times a token must " +
+                     "appear to be included in the word2vec model's vocabulary (>= 0)",
+                     typeConverter=TypeConverters.toInt)
+
+    maxSentenceLength = Param(Params._dummy(),
+                              "maxSentenceLength",
+                              "the window size (Maximum length (in words) of each sentence in the input data. Any sentence longer than this threshold will " +
+                              "be divided into chunks up to the size (> 0)",
+                              typeConverter=TypeConverters.toInt)
+
+    stepSize = Param(Params._dummy(),
+                     "stepSize",
+                     "Step size (learning rate) to be used for each iteration of optimization (> 0)",
+                     typeConverter=TypeConverters.toFloat)
+
+    maxIter = Param(Params._dummy(),
+                    "maxIter",
+                    "maximum number of iterations (>= 0)",
+                    typeConverter=TypeConverters.toInt)
+
+    seed = Param(Params._dummy(),
+                 "seed",
+                 "Random seed",
+                 typeConverter=TypeConverters.toInt)
+
+    def setVectorSize(self, vectorSize):
+        """
+        Sets vector size (default: 100).
+        """
+        return self._set(vectorSize=vectorSize)
+
+    def setWindowSize(self, windowSize):
+        """
+        Sets window size (default: 5).
+        """
+        return self._set(windowSize=windowSize)
+
+    def setStepSize(self, stepSize):
+        """
+        Sets initial learning rate (default: 0.025).
+        """
+        return self._set(stepSize=stepSize)
+
+    def setNumPartitions(self, numPartitions):
+        """
+        Sets number of partitions (default: 1). Use a small number for
+        accuracy.
+        """
+        return self._set(numPartitions=numPartitions)
+
+    def setMaxIter(self, numIterations):
+        """
+        Sets number of iterations (default: 1), which should be smaller
+        than or equal to number of partitions.
+        """
+        return self._set(maxIter=numIterations)
+
+    def setSeed(self, seed):
+        """
+        Sets random seed.
+        """
+        return self._set(seed=seed)
+
+    def setMinCount(self, minCount):
+        """
+        Sets minCount, the minimum number of times a token must appear
+        to be included in the word2vec model's vocabulary (default: 5).
+        """
+        return self._set(minCount=minCount)
+
+    def setMaxSentenceLength(self, maxSentenceLength):
+        """
+        Maximum length (in words) of each sentence in the input data.
+        Any sentence longer than this threshold will be divided into
+        chunks up to the size (> 0)
+        """
+        return self._set(maxSentenceLength=maxSentenceLength)
+
+    @keyword_only
+    def __init__(self):
+        super(Doc2VecApproach, self).__init__(classname="com.johnsnowlabs.nlp.embeddings.Doc2VecApproach")
+        self._setDefault(
+            vectorSize=100,
+            windowSize=5,
+            numPartitions=1,
+            minCount=1,
+            maxSentenceLength=1000,
+            stepSize=0.025,
+            maxIter=1,
+            seed=44
+        )
+
+    def _create_model(self, java_model):
+        return Doc2VecModel(java_model=java_model)
+
+
+class Doc2VecModel(AnnotatorModel, HasStorageRef, HasEmbeddingsProperties):
+    """Word2Vec model that creates vector representations of words in a text
+    corpus.
+
+    The algorithm first constructs a vocabulary from the corpus and then learns
+    vector representation of words in the vocabulary. The vector representation
+    can be used as features in natural language processing and machine learning
+    algorithms.
+
+    We use Word2Vec implemented in Spark ML. It uses skip-gram model in our
+    implementation and a hierarchical softmax method to train the model. The
+    variable names in the implementation match the original C implementation.
+
+    This is the instantiated model of the :class:`.Doc2VecApproach`. For
+    training your own model, please see the documentation of that class.
+
+    Pretrained models can be loaded with :meth:`.pretrained` of the companion
+    object:
+
+    >>> embeddings = Doc2VecModel.pretrained() \\
+    ...     .setInputCols(["token"]) \\
+    ...     .setOutputCol("embeddings")
+
+    The default model is `"doc2vec_gigaword_300"`, if no name is provided.
+
+    ====================== =======================
+    Input Annotation types Output Annotation type
+    ====================== =======================
+    ``TOKEN``              ``SENTENCE_EMBEDDINGS``
+    ====================== =======================
+
+    Parameters
+    ----------
+    vectorSize
+        The dimension of codes after transforming from words (> 0) , by default
+        100
+
+    References
+    ----------
+    For the original C implementation, see https://code.google.com/p/word2vec/
+
+    For the research paper, see `Efficient Estimation of Word Representations in
+    Vector Space <https://arxiv.org/abs/1301.3781>`__ and `Distributed
+    Representations of Words and Phrases and their Compositionality
+    <https://arxiv.org/pdf/1310.4546v1.pdf>`__.
+
+    Examples
+    --------
+    >>> import sparknlp
+    >>> from sparknlp.base import *
+    >>> from sparknlp.annotator import *
+    >>> from pyspark.ml import Pipeline
+    >>> documentAssembler = DocumentAssembler() \\
+    ...     .setInputCol("text") \\
+    ...     .setOutputCol("document")
+    >>> tokenizer = Tokenizer() \\
+    ...     .setInputCols(["document"]) \\
+    ...     .setOutputCol("token")
+    >>> embeddings = Doc2VecModel.pretrained() \\
+    ...     .setInputCols(["token"]) \\
+    ...     .setOutputCol("embeddings")
+    >>> embeddingsFinisher = EmbeddingsFinisher() \\
+    ...     .setInputCols(["embeddings"]) \\
+    ...     .setOutputCols("finished_embeddings") \\
+    ...     .setOutputAsVector(True)
+    >>> pipeline = Pipeline().setStages([
+    ...     documentAssembler,
+    ...     tokenizer,
+    ...     embeddings,
+    ...     embeddingsFinisher
+    ... ])
+    >>> data = spark.createDataFrame([["This is a sentence."]]).toDF("text")
+    >>> result = pipeline.fit(data).transform(data)
+    >>> result.selectExpr("explode(finished_embeddings) as result").show(1, 80)
+    +--------------------------------------------------------------------------------+
+    |                                                                          result|
+    +--------------------------------------------------------------------------------+
+    |[0.06222493574023247,0.011579325422644615,0.009919632226228714,0.109361454844...|
+    +--------------------------------------------------------------------------------+
+    """
+    name = "Doc2VecModel"
+
+    inputAnnotatorTypes = [AnnotatorType.TOKEN]
+
+    outputAnnotatorType = AnnotatorType.SENTENCE_EMBEDDINGS
+
+    vectorSize = Param(Params._dummy(),
+                       "vectorSize",
+                       "the dimension of codes after transforming from words (> 0)",
+                       typeConverter=TypeConverters.toInt)
+
+    def setVectorSize(self, vectorSize):
+        """
+        Sets vector size (default: 100).
+        """
+        return self._set(vectorSize=vectorSize)
+
+    def __init__(self, classname="com.johnsnowlabs.nlp.embeddings.Doc2VecModel", java_model=None):
+        super(Doc2VecModel, self).__init__(
+            classname=classname,
+            java_model=java_model
+        )
+        self._setDefault(
+            vectorSize=100
+        )
+
+    @staticmethod
+    def pretrained(name="doc2vec_gigaword_300", lang="en", remote_loc=None):
+        """Downloads and loads a pretrained model.
+
+        Parameters
+        ----------
+        name : str, optional
+            Name of the pretrained model, by default "doc2vec_wiki"
+        lang : str, optional
+            Language of the pretrained model, by default "en"
+        remote_loc : str, optional
+            Optional remote address of the resource, by default None. Will use
+            Spark NLPs repositories otherwise.
+
+        Returns
+        -------
+        Doc2VecModel
+            The restored model
+        """
+        from sparknlp.pretrained import ResourceDownloader
+        return ResourceDownloader.downloadModel(Doc2VecModel, name, lang, remote_loc)
+
+    def getVectors(self):
+        """
+        Returns the vector representation of the words as a dataframe
+        with two fields, word and vector.
+        """
+        return self._call_java("getVectors")