spaCy/website/docs/api/language.md

title	teaser	tag	source
Language	A text-processing pipeline	class	spacy/language.py

Usually you'll load this once per process as nlp and pass the instance around your application. The Language class is created when you call spacy.load and contains the shared vocabulary and language data, optional binary weights, e.g. provided by a trained pipeline, and the processing pipeline containing components like the tagger or parser that are called on a document in order. You can also add your own processing pipeline components that take a Doc object, modify it and return it.

Language.__init__

Initialize a Language object. Note that the meta is only used for meta information in Language.meta and not to configure the nlp object or to override the config. To initialize from a config, use Language.from_config instead.

Example

# Construction from subclass
from spacy.lang.en import English
nlp = English()

# Construction from scratch
from spacy.vocab import Vocab
from spacy.language import Language
nlp = Language(Vocab())

Name	Description
vocab	A Vocab object. If True, a vocab is created using the default language data settings. Vocab
keyword-only
max_length	Maximum number of characters allowed in a single text. Defaults to 10 ** 6. int
meta	Meta data overrides. Dict[str, Any]
create_tokenizer	Optional function that receives the nlp object and returns a tokenizer. CallableLanguage], Callable[[str], Doc
batch_size	Default batch size for pipe and evaluate. Defaults to 1000. int

Language.from_config

Create a Language object from a loaded config. Will set up the tokenizer and language data, add pipeline components based on the pipeline and add pipeline components based on the definitions specified in the config. If no config is provided, the default config of the given language is used. This is also how spaCy loads a model under the hood based on its config.cfg.

Example

from thinc.api import Config
from spacy.language import Language

config = Config().from_disk("./config.cfg")
nlp = Language.from_config(config)

Name	Description
config	The loaded config. Union[Dict[str, Any], Config]
keyword-only
vocab	A Vocab object. If True, a vocab is created using the default language data settings. Vocab
disable	Names of pipeline components to disable. Disabled pipes will be loaded but they won't be run unless you explicitly enable them by calling nlp.enable_pipe. List[str]
exclude	Names of pipeline components to exclude. Excluded components won't be loaded. List[str]
meta	Meta data overrides. Dict[str, Any]
auto_fill	Whether to automatically fill in missing values in the config, based on defaults and function argument annotations. Defaults to True. bool
validate	Whether to validate the component config and arguments against the types expected by the factory. Defaults to True. bool
RETURNS	The initialized object. Language

Language.component

Register a custom pipeline component under a given name. This allows initializing the component by name using Language.add_pipe and referring to it in config files. This classmethod and decorator is intended for simple stateless functions that take a Doc and return it. For more complex stateful components that allow settings and need access to the shared nlp object, use the Language.factory decorator. For more details and examples, see the usage documentation.

Example

from spacy.language import Language

# Usage as a decorator
@Language.component("my_component")
def my_component(doc):
   # Do something to the doc
   return doc

# Usage as a function
Language.component("my_component2", func=my_component)

Name	Description
name	The name of the component factory. str
keyword-only
assigns	Doc or Token attributes assigned by this component, e.g. ["token.ent_id"]. Used for pipe analysis. Iterable[str]
requires	Doc or Token attributes required by this component, e.g. ["token.ent_id"]. Used for pipe analysis. Iterable[str]
retokenizes	Whether the component changes tokenization. Used for pipe analysis. bool
func	Optional function if not used as a decorator. Optional[CallableDoc], Doc

Language.factory

Register a custom pipeline component factory under a given name. This allows initializing the component by name using Language.add_pipe and referring to it in config files. The registered factory function needs to take at least two named arguments which spaCy fills in automatically: nlp for the current nlp object and name for the component instance name. This can be useful to distinguish multiple instances of the same component and allows trainable components to add custom losses using the component instance name. The default_config defines the default values of the remaining factory arguments. It's merged into the nlp.config. For more details and examples, see the usage documentation.

Example

from spacy.language import Language

# Usage as a decorator
@Language.factory(
   "my_component",
   default_config={"some_setting": True},
)
def create_my_component(nlp, name, some_setting):
     return MyComponent(some_setting)

# Usage as function
Language.factory(
    "my_component",
    default_config={"some_setting": True},
    func=create_my_component
)

Name	Description
name	The name of the component factory. str
keyword-only
default_config	The default config, describing the default values of the factory arguments. Dict[str, Any]
assigns	Doc or Token attributes assigned by this component, e.g. ["token.ent_id"]. Used for pipe analysis. Iterable[str]
requires	Doc or Token attributes required by this component, e.g. ["token.ent_id"]. Used for pipe analysis. Iterable[str]
retokenizes	Whether the component changes tokenization. Used for pipe analysis. bool
default_score_weights	The scores to report during training, and their default weight towards the final score used to select the best model. Weights should sum to 1.0 per component and will be combined and normalized for the whole pipeline. If a weight is set to None, the score will not be logged or weighted. Dict[str, Optional[float]]
func	Optional function if not used as a decorator. Optional[Callable...], Callable[[Doc], Doc]

Language.__call__

Apply the pipeline to some text. The text can span multiple sentences, and can contain arbitrary whitespace. Alignment into the original string is preserved.

Example

doc = nlp("An example sentence. Another sentence.")
assert (doc[0].text, doc[0].head.tag_) == ("An", "NN")

Name	Description
text	The text to be processed. str
keyword-only
disable	Names of pipeline components to disable. List[str]
component_cfg	Optional dictionary of keyword arguments for components, keyed by component names. Defaults to None. Optional[Dict[str, Dict[str, Any]]]
RETURNS	A container for accessing the annotations. Doc

Language.pipe

Process texts as a stream, and yield Doc objects in order. This is usually more efficient than processing texts one-by-one.

Example

texts = ["One document.", "...", "Lots of documents"]
for doc in nlp.pipe(texts, batch_size=50):
    assert doc.has_annotation("DEP")

Name	Description
texts	A sequence of strings. Iterable[str]
keyword-only
as_tuples	If set to True, inputs should be a sequence of (text, context) tuples. Output will then be a sequence of (doc, context) tuples. Defaults to False. bool
batch_size	The number of texts to buffer. Optional[int]
disable	Names of pipeline components to disable. List[str]
component_cfg	Optional dictionary of keyword arguments for components, keyed by component names. Defaults to None. Optional[Dict[str, Dict[str, Any]]]
n_process 2.2.2	Number of processors to use. Defaults to 1. int
YIELDS	Documents in the order of the original text. Doc

Language.set_error_handler

Define a callback that will be invoked when an error is thrown during processing of one or more documents. Specifically, this function will call set_error_handler on all the pipeline components that define that function. The error handler will be invoked with the original component's name, the component itself, the list of documents that was being processed, and the original error.

Example

def warn_error(proc_name, proc, docs, e):
    print(f"An error occurred when applying component {proc_name}.")

nlp.set_error_handler(warn_error)

Name	Description
error_handler	A function that performs custom error handling. Callable[[str, Callable[[Doc], Doc], List[Doc], Exception]

Language.initialize

Initialize the pipeline for training and return an Optimizer. Under the hood, it uses the settings defined in the [initialize] config block to set up the vocabulary, load in vectors and tok2vec weights and pass optional arguments to the initialize methods implemented by pipeline components or the tokenizer. This method is typically called automatically when you run spacy train. See the usage guide on the config lifecycle and initialization for details.

get_examples should be a function that returns an iterable of Example objects. The data examples can either be the full training data or a representative sample. They are used to initialize the models of trainable pipeline components and are passed each component's initialize method, if available. Initialization includes validating the network, inferring missing shapes and setting up the label scheme based on the data.

If no get_examples function is provided when calling nlp.initialize, the pipeline components will be initialized with generic data. In this case, it is crucial that the output dimension of each component has already been defined either in the config, or by calling pipe.add_label for each possible output label (e.g. for the tagger or textcat).

This method was previously called begin_training. It now also takes a function that is called with no arguments and returns a sequence of Example objects instead of tuples of Doc and GoldParse objects.

Example

get_examples = lambda: examples
optimizer = nlp.initialize(get_examples)

Name	Description
get_examples	Optional function that returns gold-standard annotations in the form of Example objects. Optional[Callable], Iterable[Example]
keyword-only
sgd	An optimizer. Will be created via create_optimizer if not set. Optional[Optimizer]
RETURNS	The optimizer. Optimizer

Language.resume_training

Continue training a trained pipeline. Create and return an optimizer, and initialize "rehearsal" for any pipeline component that has a rehearse method. Rehearsal is used to prevent models from "forgetting" their initialized "knowledge". To perform rehearsal, collect samples of text you want the models to retain performance on, and call nlp.rehearse with a batch of Example objects.

Example

optimizer = nlp.resume_training()
nlp.rehearse(examples, sgd=optimizer)

Name	Description
keyword-only
sgd	An optimizer. Will be created via create_optimizer if not set. Optional[Optimizer]
RETURNS	The optimizer. Optimizer

Language.update

Update the models in the pipeline.

The Language.update method now takes a batch of Example objects instead of the raw texts and annotations or Doc and GoldParse objects. An Example streamlines how data is passed around. It stores two Doc objects: one for holding the gold-standard reference data, and one for holding the predictions of the pipeline.

For most use cases, you shouldn't have to write your own training scripts anymore. Instead, you can use spacy train with a config file and custom registered functions if needed. See the training documentation for details.

Example

for raw_text, entity_offsets in train_data:
    doc = nlp.make_doc(raw_text)
    example = Example.from_dict(doc, {"entities": entity_offsets})
    nlp.update([example], sgd=optimizer)

Name	Description
examples	A batch of Example objects to learn from. Iterable[Example]
keyword-only
drop	The dropout rate. float
sgd	An optimizer. Will be created via create_optimizer if not set. Optional[Optimizer]
losses	Dictionary to update with the loss, keyed by pipeline component. Optional[Dict[str, float]]
component_cfg	Optional dictionary of keyword arguments for components, keyed by component names. Defaults to None. Optional[Dict[str, Dict[str, Any]]]
RETURNS	The updated losses dictionary. Dict[str, float]

Language.rehearse

Perform a "rehearsal" update from a batch of data. Rehearsal updates teach the current model to make predictions similar to an initial model, to try to address the "catastrophic forgetting" problem. This feature is experimental.

Example

optimizer = nlp.resume_training()
losses = nlp.rehearse(examples, sgd=optimizer)

Name	Description
examples	A batch of Example objects to learn from. Iterable[Example]
keyword-only
drop	The dropout rate. float
sgd	An optimizer. Will be created via create_optimizer if not set. Optional[Optimizer]
losses	Dictionary to update with the loss, keyed by pipeline component. Optional[Dict[str, float]]
RETURNS	The updated losses dictionary. Dict[str, float]

Language.evaluate

Evaluate a pipeline's components.

The Language.evaluate method now takes a batch of Example objects instead of tuples of Doc and GoldParse objects.

Example

scores = nlp.evaluate(examples)
print(scores)

Name	Description
examples	A batch of Example objects to learn from. Iterable[Example]
keyword-only
batch_size	The batch size to use. Optional[int]
scorer	Optional Scorer to use. If not passed in, a new one will be created. Optional[Scorer]
component_cfg	Optional dictionary of keyword arguments for components, keyed by component names. Defaults to None. Optional[Dict[str, Dict[str, Any]]]
scorer_cfg	Optional dictionary of keyword arguments for the Scorer. Defaults to None. Optional[Dict[str, Any]]
RETURNS	A dictionary of evaluation scores. Dict[str, Union[float, Dict[str, float]]]

Language.use_params

Replace weights of models in the pipeline with those provided in the params dictionary. Can be used as a context manager, in which case, models go back to their original weights after the block.

Example

with nlp.use_params(optimizer.averages):
    nlp.to_disk("/tmp/checkpoint")

Name	Description
params	A dictionary of parameters keyed by model ID. dict

Language.add_pipe

Add a component to the processing pipeline. Expects a name that maps to a component factory registered using @Language.component or @Language.factory. Components should be callables that take a Doc object, modify it and return it. Only one of before, after, first or last can be set. Default behavior is last=True.

As of v3.0, the Language.add_pipe method doesn't take callables anymore and instead expects the name of a component factory registered using @Language.component or @Language.factory. It now takes care of creating the component, adds it to the pipeline and returns it.

Example

@Language.component("component")
def component_func(doc):
    # modify Doc and return it
    return doc

nlp.add_pipe("component", before="ner")
component = nlp.add_pipe("component", name="custom_name", last=True)

# Add component from source pipeline
source_nlp = spacy.load("en_core_web_sm")
nlp.add_pipe("ner", source=source_nlp)

Name	Description
factory_name	Name of the registered component factory. str
name	Optional unique name of pipeline component instance. If not set, the factory name is used. An error is raised if the name already exists in the pipeline. Optional[str]
keyword-only
before	Component name or index to insert component directly before. Optional[Union[str, int]]
after	Component name or index to insert component directly after. Optional[Union[str, int]]
first	Insert component first / not first in the pipeline. Optional[bool]
last	Insert component last / not last in the pipeline. Optional[bool]
config 3	Optional config parameters to use for this component. Will be merged with the default_config specified by the component factory. Dict[str, Any]
source 3	Optional source pipeline to copy component from. If a source is provided, the factory_name is interpreted as the name of the component in the source pipeline. Make sure that the vocab, vectors and settings of the source pipeline match the target pipeline. Optional[Language]
validate 3	Whether to validate the component config and arguments against the types expected by the factory. Defaults to True. bool
RETURNS	The pipeline component. Callable[[Doc], Doc]

Language.create_pipe

Create a pipeline component from a factory.

As of v3.0, the Language.add_pipe method also takes the string name of the factory, creates the component, adds it to the pipeline and returns it. The Language.create_pipe method is now mostly used internally. To create a component and add it to the pipeline, you should always use Language.add_pipe.

Example

parser = nlp.create_pipe("parser")

Name	Description
factory_name	Name of the registered component factory. str
name	Optional unique name of pipeline component instance. If not set, the factory name is used. An error is raised if the name already exists in the pipeline. Optional[str]
keyword-only
config 3	Optional config parameters to use for this component. Will be merged with the default_config specified by the component factory. Dict[str, Any]
validate 3	Whether to validate the component config and arguments against the types expected by the factory. Defaults to True. bool
RETURNS	The pipeline component. Callable[[Doc], Doc]

Language.has_factory

Check whether a factory name is registered on the Language class or subclass. Will check for language-specific factories registered on the subclass, as well as general-purpose factories registered on the Language base class, available to all subclasses.

Example

from spacy.language import Language
from spacy.lang.en import English

@English.component("component")
def component(doc):
    return doc

assert English.has_factory("component")
assert not Language.has_factory("component")

Name	Description
name	Name of the pipeline factory to check. str
RETURNS	Whether a factory of that name is registered on the class. bool

Language.has_pipe

Check whether a component is present in the pipeline. Equivalent to name in nlp.pipe_names.

Example

@Language.component("component")
def component(doc):
    return doc

nlp.add_pipe("component", name="my_component")
assert "my_component" in nlp.pipe_names
assert nlp.has_pipe("my_component")

Name	Description
name	Name of the pipeline component to check. str
RETURNS	Whether a component of that name exists in the pipeline. bool

Language.get_pipe

Get a pipeline component for a given component name.

Example

parser = nlp.get_pipe("parser")
custom_component = nlp.get_pipe("custom_component")

Name	Description
name	Name of the pipeline component to get. str
RETURNS	The pipeline component. Callable[[Doc], Doc]

Language.replace_pipe

Replace a component in the pipeline and return the new component.

As of v3.0, the Language.replace_pipe method doesn't take callables anymore and instead expects the name of a component factory registered using @Language.component or @Language.factory.

Example

new_parser = nlp.replace_pipe("parser", "my_custom_parser")

Name	Description
name	Name of the component to replace. str
component	The factory name of the component to insert. str
keyword-only
config 3	Optional config parameters to use for the new component. Will be merged with the default_config specified by the component factory. Optional[Dict[str, Any]]
validate 3	Whether to validate the component config and arguments against the types expected by the factory. Defaults to True. bool
RETURNS	The new pipeline component. Callable[[Doc], Doc]

Language.rename_pipe

Rename a component in the pipeline. Useful to create custom names for pre-defined and pre-loaded components. To change the default name of a component added to the pipeline, you can also use the name argument on add_pipe.

Example

nlp.rename_pipe("parser", "spacy_parser")

Name	Description
old_name	Name of the component to rename. str
new_name	New name of the component. str

Language.remove_pipe

Remove a component from the pipeline. Returns the removed component name and component function.

Example

name, component = nlp.remove_pipe("parser")
assert name == "parser"

Name	Description
name	Name of the component to remove. str
RETURNS	A (name, component) tuple of the removed component. Tuple[str, CallableDoc], Doc

Language.disable_pipe

Temporarily disable a pipeline component so it's not run as part of the pipeline. Disabled components are listed in nlp.disabled and included in nlp.components, but not in nlp.pipeline, so they're not run when you process a Doc with the nlp object. If the component is already disabled, this method does nothing.

Example

nlp.add_pipe("ner")
nlp.add_pipe("textcat")
assert nlp.pipe_names == ["ner", "textcat"]
nlp.disable_pipe("ner")
assert nlp.pipe_names == ["textcat"]
assert nlp.component_names == ["ner", "textcat"]
assert nlp.disabled == ["ner"]

Name	Description
name	Name of the component to disable. str

Language.enable_pipe

Enable a previously disabled component (e.g. via Language.disable_pipes) so it's run as part of the pipeline, nlp.pipeline. If the component is already enabled, this method does nothing.

Example

nlp.disable_pipe("ner")
assert "ner" in nlp.disabled
assert not "ner" in nlp.pipe_names
nlp.enable_pipe("ner")
assert not "ner" in nlp.disabled
assert "ner" in nlp.pipe_names

Name	Description
name	Name of the component to enable. str

Language.select_pipes

Disable one or more pipeline components. If used as a context manager, the pipeline will be restored to the initial state at the end of the block. Otherwise, a DisabledPipes object is returned, that has a .restore() method you can use to undo your changes. You can specify either disable (as a list or string), or enable. In the latter case, all components not in the enable list will be disabled. Under the hood, this method calls into disable_pipe and enable_pipe.

Example

with nlp.select_pipes(disable=["tagger", "parser"]):
   nlp.initialize()

with nlp.select_pipes(enable="ner"):
    nlp.initialize()

disabled = nlp.select_pipes(disable=["tagger", "parser"])
nlp.initialize()
disabled.restore()

As of spaCy v3.0, the disable_pipes method has been renamed to select_pipes:

- nlp.disable_pipes(["tagger", "parser"])
+ nlp.select_pipes(disable=["tagger", "parser"])

Name	Description
keyword-only
disable	Name(s) of pipeline components to disable. Optional[Union[str, Iterable[str]]]
enable	Name(s) of pipeline components that will not be disabled. Optional[Union[str, Iterable[str]]]
RETURNS	The disabled pipes that can be restored by calling the object's .restore() method. DisabledPipes

Language.get_factory_meta

Get the factory meta information for a given pipeline component name. Expects the name of the component factory. The factory meta is an instance of the FactoryMeta dataclass and contains the information about the component and its default provided by the @Language.component or @Language.factory decorator.

Example

factory_meta = Language.get_factory_meta("ner")
assert factory_meta.factory == "ner"
print(factory_meta.default_config)

Name	Description
name	The factory name. str
RETURNS	The factory meta. FactoryMeta

Language.get_pipe_meta

Get the factory meta information for a given pipeline component name. Expects the name of the component instance in the pipeline. The factory meta is an instance of the FactoryMeta dataclass and contains the information about the component and its default provided by the @Language.component or @Language.factory decorator.

Example

nlp.add_pipe("ner", name="entity_recognizer")
factory_meta = nlp.get_pipe_meta("entity_recognizer")
assert factory_meta.factory == "ner"
print(factory_meta.default_config)

Name	Description
name	The pipeline component name. str
RETURNS	The factory meta. FactoryMeta

Language.analyze_pipes

Analyze the current pipeline components and show a summary of the attributes they assign and require, and the scores they set. The data is based on the information provided in the @Language.component and @Language.factory decorator. If requirements aren't met, e.g. if a component specifies a required property that is not set by a previous component, a warning is shown.

The pipeline analysis is static and does not actually run the components. This means that it relies on the information provided by the components themselves. If a custom component declares that it assigns an attribute but it doesn't, the pipeline analysis won't catch that.

Example

nlp = spacy.blank("en")
nlp.add_pipe("tagger")
nlp.add_pipe("entity_linker")
analysis = nlp.analyze_pipes()

### Structured
{
  "summary": {
    "tagger": {
      "assigns": ["token.tag"],
      "requires": [],
      "scores": ["tag_acc", "pos_acc", "lemma_acc"],
      "retokenizes": false
    },
    "entity_linker": {
      "assigns": ["token.ent_kb_id"],
      "requires": ["doc.ents", "doc.sents", "token.ent_iob", "token.ent_type"],
      "scores": [],
      "retokenizes": false
    }
  },
  "problems": {
    "tagger": [],
    "entity_linker": ["doc.ents", "doc.sents", "token.ent_iob", "token.ent_type"]
  },
  "attrs": {
    "token.ent_iob": { "assigns": [], "requires": ["entity_linker"] },
    "doc.ents": { "assigns": [], "requires": ["entity_linker"] },
    "token.ent_kb_id": { "assigns": ["entity_linker"], "requires": [] },
    "doc.sents": { "assigns": [], "requires": ["entity_linker"] },
    "token.tag": { "assigns": ["tagger"], "requires": [] },
    "token.ent_type": { "assigns": [], "requires": ["entity_linker"] }
  }
}

### Pretty
============================= Pipeline Overview =============================

#   Component       Assigns           Requires         Scores        Retokenizes
-   -------------   ---------------   --------------   -----------   -----------
0   tagger          token.tag                          tag_acc       False

1   entity_linker   token.ent_kb_id   doc.ents         nel_micro_f   False
                                      doc.sents        nel_micro_r
                                      token.ent_iob    nel_micro_p
                                      token.ent_type


================================ Problems (4) ================================
⚠ 'entity_linker' requirements not met: doc.ents, doc.sents,
token.ent_iob, token.ent_type

Name	Description
keyword-only
keys	The values to display in the table. Corresponds to attributes of the FactoryMeta. Defaults to ["assigns", "requires", "scores", "retokenizes"]. List[str]
pretty	Pretty-print the results as a table. Defaults to False. bool
RETURNS	Dictionary containing the pipe analysis, keyed by "summary" (component meta by pipe), "problems" (attribute names by pipe) and "attrs" (pipes that assign and require an attribute, keyed by attribute). Optional[Dict[str, Any]]

Language.replace_listeners

Find listener layers (connecting to a shared token-to-vector embedding component) of a given pipeline component model and replace them with a standalone copy of the token-to-vector layer. The listener layer allows other components to connect to a shared token-to-vector embedding component like Tok2Vec or Transformer. Replacing listeners can be useful when training a pipeline with components sourced from an existing pipeline: if multiple components (e.g. tagger, parser, NER) listen to the same token-to-vector component, but some of them are frozen and not updated, their performance may degrade significally as the token-to-vector component is updated with new data. To prevent this, listeners can be replaced with a standalone token-to-vector layer that is owned by the component and doesn't change if the component isn't updated.

This method is typically not called directly and only executed under the hood when loading a config with sourced components that define replace_listeners.

### Example
nlp = spacy.load("en_core_web_sm")
nlp.replace_listeners("tok2vec", "tagger", ["model.tok2vec"])

### config.cfg (excerpt)
[training]
frozen_components = ["tagger"]

[components]

[components.tagger]
source = "en_core_web_sm"
replace_listeners = ["model.tok2vec"]

Name	Description
tok2vec_name	Name of the token-to-vector component, typically "tok2vec" or "transformer".str
pipe_name	Name of pipeline component to replace listeners for. str
listeners	The paths to the listeners, relative to the component config, e.g. ["model.tok2vec"]. Typically, implementations will only connect to one tok2vec component, model.tok2vec, but in theory, custom models can use multiple listeners. The value here can either be an empty list to not replace any listeners, or a complete list of the paths to all listener layers used by the model that should be replaced.Iterable[str]

Language.meta

Meta data for the Language class, including name, version, data sources, license, author information and more. If a trained pipeline is loaded, this contains meta data of the pipeline. The Language.meta is also what's serialized as the meta.json when you save an nlp object to disk. See the meta data format for more details.

As of v3.0, the meta only contains meta information about the pipeline and isn't used to construct the language class and pipeline components. This information is expressed in the config.cfg.

Example

print(nlp.meta)

Name	Description
RETURNS	The meta data. Dict[str, Any]

Language.config

Export a trainable config.cfg for the current nlp object. Includes the current pipeline, all configs used to create the currently active pipeline components, as well as the default training config that can be used with spacy train. Language.config returns a Thinc Config object, which is a subclass of the built-in dict. It supports the additional methods to_disk (serialize the config to a file) and to_str (output the config as a string).

Example

nlp.config.to_disk("./config.cfg")
print(nlp.config.to_str())

Name	Description
RETURNS	The config. Config

Language.to_disk

Save the current state to a directory. Under the hood, this method delegates to the to_disk methods of the individual pipeline components, if available. This means that if a trained pipeline is loaded, all components and their weights will be saved to disk.

Example

nlp.to_disk("/path/to/pipeline")

Name	Description
path	A path to a directory, which will be created if it doesn't exist. Paths may be either strings or Path-like objects. Union[str, Path]
keyword-only
exclude	Names of pipeline components or serialization fields to exclude. Iterable[str]

Language.from_disk

Loads state from a directory, including all data that was saved with the Language object. Modifies the object in place and returns it.

Keep in mind that this method only loads the serialized state and doesn't set up the nlp object. This means that it requires the correct language class to be initialized and all pipeline components to be added to the pipeline. If you want to load a serialized pipeline from a directory, you should use spacy.load, which will set everything up for you.

Example

from spacy.language import Language
nlp = Language().from_disk("/path/to/pipeline")

# Using language-specific subclass
from spacy.lang.en import English
nlp = English().from_disk("/path/to/pipeline")

Name	Description
path	A path to a directory. Paths may be either strings or Path-like objects. Union[str, Path]
keyword-only
exclude	Names of pipeline components or serialization fields to exclude. Iterable[str]
RETURNS	The modified Language object. Language

Language.to_bytes

Serialize the current state to a binary string.

Example

nlp_bytes = nlp.to_bytes()

Name	Description
keyword-only
exclude	Names of pipeline components or serialization fields to exclude. iterable
RETURNS	The serialized form of the Language object. bytes

Language.from_bytes

Load state from a binary string. Note that this method is commonly used via the subclasses like English or German to make language-specific functionality like the lexical attribute getters available to the loaded object.

Note that if you want to serialize and reload a whole pipeline, using this alone won't work, you also need to handle the config. See "Serializing the pipeline" for details.

Example

from spacy.lang.en import English
nlp_bytes = nlp.to_bytes()
nlp2 = English()
nlp2.from_bytes(nlp_bytes)

Name	Description
bytes_data	The data to load from. bytes
keyword-only
exclude	Names of pipeline components or serialization fields to exclude. Iterable[str]
RETURNS	The Language object. Language

Attributes

Name	Description
vocab	A container for the lexical types. Vocab
tokenizer	The tokenizer. Tokenizer
make_doc	Callable that takes a string and returns a Doc. Callable[[str], Doc]
pipeline	List of (name, component) tuples describing the current processing pipeline, in order. List[Tuple[str, CallableDoc], Doc]
pipe_names 2	List of pipeline component names, in order. List[str]
pipe_labels 2.2	List of labels set by the pipeline components, if available, keyed by component name. Dict[str, List[str]]
pipe_factories 2.2	Dictionary of pipeline component names, mapped to their factory names. Dict[str, str]
factories	All available factory functions, keyed by name. Dict[str, Callable...], Callable[[Doc], Doc]
factory_names 3	List of all available factory names. List[str]
components 3	List of all available (name, component) tuples, including components that are currently disabled. List[Tuple[str, CallableDoc], Doc]
component_names 3	List of all available component names, including components that are currently disabled. List[str]
disabled 3	Names of components that are currently disabled and don't run as part of the pipeline. List[str]
path 2	Path to the pipeline data directory, if a pipeline is loaded from a path or package. Otherwise None. Optional[Path]

Class attributes

Name	Description
Defaults	Settings, data and factory methods for creating the nlp object and processing pipeline. Defaults
lang	IETF language tag, such as 'en' for English. str
default_config	Base config to use for Language.config. Defaults to default_config.cfg. Config

Defaults

The following attributes can be set on the Language.Defaults class to customize the default language data:

Example

from spacy.language import language
from spacy.lang.tokenizer_exceptions import URL_MATCH
from thinc.api import Config

DEFAULT_CONFIFG = """
[nlp.tokenizer]
@tokenizers = "MyCustomTokenizer.v1"
"""

class Defaults(Language.Defaults):
   stop_words = set()
   tokenizer_exceptions = {}
   prefixes = tuple()
   suffixes = tuple()
   infixes = tuple()
   token_match = None
   url_match = URL_MATCH
   lex_attr_getters = {}
   syntax_iterators = {}
   writing_system = {"direction": "ltr", "has_case": True, "has_letters": True}
   config = Config().from_str(DEFAULT_CONFIG)

Name	Description
stop_words	List of stop words, used for Token.is_stop. Example: stop_words.py Set[str]
tokenizer_exceptions	Tokenizer exception rules, string mapped to list of token attributes. Example: de/tokenizer_exceptions.py Dict[str, List[dict]]
prefixes, suffixes, infixes	Prefix, suffix and infix rules for the default tokenizer. Example: puncutation.py Optional[Sequence[Union[str, Pattern]]]
token_match	Optional regex for matching strings that should never be split, overriding the infix rules. Example: fr/tokenizer_exceptions.py Optional[Callable]
url_match	Regular expression for matching URLs. Prefixes and suffixes are removed before applying the match. Example: tokenizer_exceptions.py Optional[Callable]
lex_attr_getters	Custom functions for setting lexical attributes on tokens, e.g. like_num. Example: lex_attrs.py Dict[int, Callablestr], Any
syntax_iterators	Functions that compute views of a Doc object based on its syntax. At the moment, only used for noun chunks. Example: syntax_iterators.py. Dict[str, CallableUnion[Doc, Span, Iterator[Span]]]
writing_system	Information about the language's writing system, available via Vocab.writing_system. Defaults to: {"direction": "ltr", "has_case": True, "has_letters": True}.. Example: zh/__init__.py Dict[str, Any]
config	Default config added to nlp.config. This can include references to custom tokenizers or lemmatizers. Example: zh/__init__.py Config

Serialization fields

During serialization, spaCy will export several data fields used to restore different aspects of the object. If needed, you can exclude them from serialization by passing in the string names via the exclude argument.

Example

data = nlp.to_bytes(exclude=["tokenizer", "vocab"])
nlp.from_disk("/pipeline", exclude=["ner"])

Name	Description
vocab	The shared Vocab.
tokenizer	Tokenization rules and exceptions.
meta	The meta data, available as Language.meta.
...	String names of pipeline components, e.g. "ner".

FactoryMeta

The FactoryMeta contains the information about the component and its default provided by the @Language.component or @Language.factory decorator. It's created whenever a component is defined and stored on the Language class for each component instance and factory instance.

Name	Description
factory	The name of the registered component factory. str
default_config	The default config, describing the default values of the factory arguments. Dict[str, Any]
assigns	Doc or Token attributes assigned by this component, e.g. ["token.ent_id"]. Used for pipe analysis. Iterable[str]
requires	Doc or Token attributes required by this component, e.g. ["token.ent_id"]. Used for pipe analysis. Iterable[str]
retokenizes	Whether the component changes tokenization. Used for pipe analysis. bool
default_score_weights	The scores to report during training, and their default weight towards the final score used to select the best model. Weights should sum to 1.0 per component and will be combined and normalized for the whole pipeline. If a weight is set to None, the score will not be logged or weighted. Dict[str, Optional[float]]
scores	All scores set by the components if it's trainable, e.g. ["ents_f", "ents_r", "ents_p"]. Based on the default_score_weights and used for pipe analysis. Iterable[str]