explosion--spacy/spacy_cli/cli_manifest.json

{
  "command": "python -m spacy",
  "command_help": {
    "apply": "Usage: python -m spacy apply [OPTIONS] MODEL DATA_PATH OUTPUT_FILE\n\n  Apply a trained pipeline to documents to get predictions. Expects a loadable\n  spaCy pipeline and path to the data, which can be a directory or a file. The\n  data files can be provided in multiple formats:     1. .spacy files     2.\n  .jsonl files with a specified \"field\" to read the text from.     3. Files with\n  any other extension are assumed to be containing        a single document.\n  DOCS: https://spacy.io/api/cli#apply\n\nArguments:\n  MODEL        Model name or path  [required]\n  DATA_PATH    Location of the documents to predict on. Can be a single file in\n               .spacy format or a .jsonl file. Files with other extensions are\n               treated as single plain text documents. If a directory is\n               provided it is traversed recursively to grab all files to be\n               processed. The files can be a mixture of .spacy, .jsonl and text\n               files. If .jsonl is provided the specified field is going to be\n               grabbed (\"text\" by default).  [required]\n  OUTPUT_FILE  Path to save the resulting .spacy file  [required]\n\nOptions:\n  -c, --code PATH           Path to Python file with additional code (registered\n                            functions) to be imported\n  -tk, --text-key TEXT      Key containing text string for JSONL  [default:\n                            text]\n  -F, --force               Force overwriting the output file\n  -g, --gpu-id INTEGER      GPU ID or -1 for CPU.  [default: -1]\n  -b, --batch-size INTEGER  Batch size.  [default: 1]\n  -n, --n-process INTEGER   number of processors to use.  [default: 1]\n  --help                    Show this message and exit.\n",
    "assemble": "Usage: python -m spacy assemble [OPTIONS] CONFIG_PATH OUTPUT_PATH\n\n  Assemble a spaCy pipeline from a config file. The config file includes all\n  settings for initializing the pipeline. To override settings in the config,\n  e.g. settings that point to local paths or that you want to experiment with,\n  you can override them as command line options. The --code argument lets you\n  pass in a Python file that can be used to register custom functions that are\n  referenced in the config.\n\n  DOCS: https://spacy.io/api/cli#assemble\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n  OUTPUT_PATH  Output directory to store assembled pipeline in  [required]\n\nOptions:\n  -c, --code PATH     Path to Python file with additional code (registered\n                      functions) to be imported\n  -V, -VV, --verbose  Display more information for debugging purposes\n  --help              Show this message and exit.\n",
    "benchmark accuracy": "Usage: python -m spacy benchmark accuracy [OPTIONS] MODEL DATA_PATH\n\n  Evaluate a trained pipeline. Expects a loadable spaCy pipeline and evaluation\n  data in the binary .spacy format. The --gold-preproc option sets up the\n  evaluation examples with gold-standard sentences and tokens for the\n  predictions. Gold preprocessing helps the annotations align to the\n  tokenization, and may result in sequences of more consistent length. However,\n  it may reduce runtime accuracy due to train/test skew. To render a sample of\n  dependency parses in a HTML file, set as output directory as the displacy_path\n  argument.\n\n  DOCS: https://spacy.io/api/cli#benchmark-accuracy\n\nArguments:\n  MODEL      Model name or path  [required]\n  DATA_PATH  Location of binary evaluation data in .spacy format  [required]\n\nOptions:\n  -o, --output FILE               Output JSON file for metrics\n  -c, --code PATH                 Path to Python file with additional code\n                                  (registered functions) to be imported\n  -g, --gpu-id INTEGER            GPU ID or -1 for CPU  [default: -1]\n  -G, --gold-preproc              Use gold preprocessing\n  -dp, --displacy-path DIRECTORY  Directory to output rendered parses as HTML\n  -dl, --displacy-limit INTEGER   Limit of parses to render as HTML  [default:\n                                  25]\n  -P, --per-component             Return scores per component, only applicable\n                                  when an output JSON file is specified.\n  -sk, --spans-key TEXT           Spans key to use when evaluating Doc.spans\n                                  [default: sc]\n  --help                          Show this message and exit.\n",
    "benchmark speed": "Usage: python -m spacy benchmark speed [OPTIONS] MODEL DATA_PATH\n\n  Benchmark a pipeline. Expects a loadable spaCy pipeline and benchmark data in\n  the binary .spacy format.\n\nArguments:\n  MODEL      Model name or path  [required]\n  DATA_PATH  Location of binary evaluation data in .spacy format  [required]\n\nOptions:\n  -b, --batch-size INTEGER RANGE  Override the pipeline batch size  [x>=1]\n  --no-shuffle                    Do not shuffle benchmark data\n  -g, --gpu-id INTEGER            GPU ID or -1 for CPU  [default: -1]\n  --batches INTEGER RANGE         Minimum number of batches to benchmark\n                                  [default: 50; x>=30]\n  -w, --warmup INTEGER RANGE      Number of iterations over the data for warmup\n                                  [default: 3; x>=0]\n  -c, --code PATH                 Path to Python file with additional code\n                                  (registered functions) to be imported\n  --help                          Show this message and exit.\n",
    "convert": "Usage: python -m spacy convert [OPTIONS] INPUT_PATH [OUTPUT_DIR]\n\n  Convert files into json or DocBin format for training. The resulting .spacy\n  file can be used with the train command and other experiment management\n  functions.\n\n  If no output_dir is specified and the output format is JSON, the data is\n  written to stdout, so you can pipe them forward to a JSON file: $ spacy\n  convert some_file.conllu --file-type json > some_file.json\n\n  DOCS: https://spacy.io/api/cli#convert\n\nArguments:\n  INPUT_PATH    Input file or directory  [required]\n  [OUTPUT_DIR]  Output directory. '-' for stdout.  [default: -]\n\nOptions:\n  -t, --file-type [json|spacy]  Type of data to produce  [default: spacy]\n  -n, --n-sents INTEGER         Number of sentences per doc (0 to disable)\n                                [default: 1]\n  -s, --seg-sents               Segment sentences (for -c ner)\n  -b, --model, --base TEXT      Trained spaCy pipeline for sentence segmentation\n                                to use as base (for --seg-sents)\n  -m, --morphology              Enable appending morphology to tags\n  -T, --merge-subtokens         Merge CoNLL-U subtokens\n  -c, --converter TEXT          Converter: ('conllubio', 'conllu', 'conll',\n                                'ner', 'iob', 'json')  [default: auto]\n  -nm, --ner-map PATH           NER tag mapping (as JSON-encoded dict of entity\n                                types)\n  -l, --lang TEXT               Language (if tokenizer required)\n  -C, --concatenate             Concatenate output to a single file\n  --help                        Show this message and exit.\n",
    "debug config": "Usage: python -m spacy debug config [OPTIONS] CONFIG_PATH\n\n  Debug a config file and show validation errors. The command will create all\n  objects in the tree and validate them. Note that some config validation errors\n  are blocking and will prevent the rest of the config from being resolved. This\n  means that you may not see all validation errors at once and some issues are\n  only shown once previous errors have been fixed. Similar as with the 'train'\n  command, you can override settings from the config as command line options.\n  For instance, --training.batch_size 128 overrides the value of \"batch_size\" in\n  the block \"[training]\".\n\n  DOCS: https://spacy.io/api/cli#debug-config\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n\nOptions:\n  -c, --code-path, --code PATH  Path to Python file with additional code\n                                (registered functions) to be imported\n  -F, --show-functions          Show an overview of all registered functions\n                                used in the config and where they come from\n                                (modules, files etc.)\n  -V, --show-variables          Show an overview of all variables referenced in\n                                the config and their values. This will also\n                                reflect variables overwritten on the CLI.\n  --help                        Show this message and exit.\n",
    "debug data": "Usage: python -m spacy debug data [OPTIONS] CONFIG_PATH\n\n  Analyze, debug and validate your training and development data. Outputs useful\n  stats, and can help you find problems like invalid entity annotations, cyclic\n  dependencies, low data labels and more.\n\n  DOCS: https://spacy.io/api/cli#debug-data\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n\nOptions:\n  -c, --code-path, --code PATH  Path to Python file with additional code\n                                (registered functions) to be imported\n  -IW, --ignore-warnings        Ignore warnings, only show stats and errors\n  -V, --verbose                 Print additional information and explanations\n  -NF, --no-format              Don't pretty-print the results\n  --help                        Show this message and exit.\n",
    "debug diff-config": "Usage: python -m spacy debug diff-config [OPTIONS] CONFIG_PATH\n\n  Show a diff of a config file with respect to spaCy's defaults or another\n  config file. If additional settings were used in the creation of the config\n  file, then you must supply these as extra parameters to the command when\n  comparing to the default settings. The generated diff can also be used when\n  posting to the discussion forum to provide more information for the\n  maintainers.\n\n  The `optimize`, `gpu`, and `pretraining` options are only relevant when\n  comparing against the default configuration (or specifically when `compare_to`\n  is None).\n\n  DOCS: https://spacy.io/api/cli#debug-diff\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n\nOptions:\n  --compare-to PATH               Path to a config file to diff against, or\n                                  `None` to compare against default settings\n  -o, --optimize [efficiency|accuracy]\n                                  Whether the user config was optimized for\n                                  efficiency or accuracy. Only relevant when\n                                  comparing against the default config.\n                                  [default: efficiency]\n  -G, --gpu                       Whether the original config can run on a GPU.\n                                  Only relevant when comparing against the\n                                  default config.\n  --pretraining, --pt             Whether to compare on a config with\n                                  pretraining involved. Only relevant when\n                                  comparing against the default config.\n  -md, --markdown                 Generate Markdown for GitHub issues\n  --help                          Show this message and exit.\n",
    "debug model": "Usage: python -m spacy debug model [OPTIONS] CONFIG_PATH COMPONENT\n\n  Analyze a Thinc model implementation. Includes checks for internal structure\n  and activations during training.\n\n  DOCS: https://spacy.io/api/cli#debug-model\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n  COMPONENT    Name of the pipeline component of which the model should be\n               analysed  [required]\n\nOptions:\n  -l, --layers TEXT     Comma-separated names of layer IDs to print\n  -DIM, --dimensions    Show dimensions\n  -PAR, --parameters    Show parameters\n  -GRAD, --gradients    Show gradients\n  -ATTR, --attributes   Show attributes\n  -P0, --print-step0    Print model before training\n  -P1, --print-step1    Print model after initialization\n  -P2, --print-step2    Print model after training\n  -P3, --print-step3    Print final predictions\n  -g, --gpu-id INTEGER  GPU ID or -1 for CPU  [default: -1]\n  --help                Show this message and exit.\n",
    "debug profile": "Usage: python -m spacy debug profile [OPTIONS] MODEL [INPUTS]\n\n  Profile which functions take the most time in a spaCy pipeline. Input should\n  be formatted as one JSON object per line with a key \"text\". It can either be\n  provided as a JSONL file, or be read from sys.sytdin. If no input file is\n  specified, the IMDB dataset is loaded via Thinc.\n\n  DOCS: https://spacy.io/api/cli#debug-profile\n\nArguments:\n  MODEL     Trained pipeline to load  [required]\n  [INPUTS]  Location of input file. '-' for stdin.\n\nOptions:\n  -n, --n-texts INTEGER  Maximum number of texts to use if available  [default:\n                         10000]\n  --help                 Show this message and exit.\n",
    "debug-data": "Usage: python -m spacy debug-data [OPTIONS] CONFIG_PATH\n\n  Analyze, debug and validate your training and development data. Outputs useful\n  stats, and can help you find problems like invalid entity annotations, cyclic\n  dependencies, low data labels and more.\n\n  DOCS: https://spacy.io/api/cli#debug-data\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n\nOptions:\n  -c, --code-path, --code PATH  Path to Python file with additional code\n                                (registered functions) to be imported\n  -IW, --ignore-warnings        Ignore warnings, only show stats and errors\n  -V, --verbose                 Print additional information and explanations\n  -NF, --no-format              Don't pretty-print the results\n  --help                        Show this message and exit.\n",
    "download": "Usage: python -m spacy download [OPTIONS] MODEL\n\n  Download compatible trained pipeline from the default download path using pip.\n  If --direct flag is set, the command expects the full package name with\n  version. For direct downloads, the compatibility check will be skipped. All\n  additional arguments provided to this command will be passed to `pip install`\n  on package installation.\n\n  DOCS: https://spacy.io/api/cli#download AVAILABLE PACKAGES:\n  https://spacy.io/models\n\nArguments:\n  MODEL  Name of pipeline package to download  [required]\n\nOptions:\n  -d, -D, --direct  Force direct download of name + version\n  -S, --sdist       Download sdist (.tar.gz) archive instead of pre-built binary\n                    wheel\n  -U, --url TEXT    Download from given url\n  --help            Show this message and exit.\n",
    "evaluate": "Usage: python -m spacy evaluate [OPTIONS] MODEL DATA_PATH\n\n  Evaluate a trained pipeline. Expects a loadable spaCy pipeline and evaluation\n  data in the binary .spacy format. The --gold-preproc option sets up the\n  evaluation examples with gold-standard sentences and tokens for the\n  predictions. Gold preprocessing helps the annotations align to the\n  tokenization, and may result in sequences of more consistent length. However,\n  it may reduce runtime accuracy due to train/test skew. To render a sample of\n  dependency parses in a HTML file, set as output directory as the displacy_path\n  argument.\n\n  DOCS: https://spacy.io/api/cli#benchmark-accuracy\n\nArguments:\n  MODEL      Model name or path  [required]\n  DATA_PATH  Location of binary evaluation data in .spacy format  [required]\n\nOptions:\n  -o, --output FILE               Output JSON file for metrics\n  -c, --code PATH                 Path to Python file with additional code\n                                  (registered functions) to be imported\n  -g, --gpu-id INTEGER            GPU ID or -1 for CPU  [default: -1]\n  -G, --gold-preproc              Use gold preprocessing\n  -dp, --displacy-path DIRECTORY  Directory to output rendered parses as HTML\n  -dl, --displacy-limit INTEGER   Limit of parses to render as HTML  [default:\n                                  25]\n  -P, --per-component             Return scores per component, only applicable\n                                  when an output JSON file is specified.\n  -sk, --spans-key TEXT           Spans key to use when evaluating Doc.spans\n                                  [default: sc]\n  --help                          Show this message and exit.\n",
    "find-function": "Usage: python -m spacy find-function [OPTIONS] FUNC_NAME\n\n  Find the module, path and line number to the file the registered function is\n  defined in, if available.\n\n  func_name (str): Name of the registered function. registry_name\n  (Optional[str]): Name of the catalogue registry.\n\n  DOCS: https://spacy.io/api/cli#find-function\n\nArguments:\n  FUNC_NAME  Name of the registered function.  [required]\n\nOptions:\n  -r, --registry TEXT  Name of the catalogue registry.\n  --help               Show this message and exit.\n",
    "find-threshold": "Usage: python -m spacy find-threshold [OPTIONS] MODEL DATA_PATH PIPE_NAME\n                                      THRESHOLD_KEY SCORES_KEY\n\n  Runs prediction trials for a trained model with varying thresholds to maximize\n  the specified metric. The search space for the threshold is traversed linearly\n  from 0 to 1 in `n_trials` steps. Results are displayed in a table on `stdout`\n  (the corresponding API call to `spacy.cli.find_threshold.find_threshold()`\n  returns all results).\n\n  This is applicable only for components whose predictions are influenced by\n  thresholds - e.g. `textcat_multilabel` and `spancat`, but not `textcat`. Note\n  that the full path to the corresponding threshold attribute in the config has\n  to be provided.\n\n  DOCS: https://spacy.io/api/cli#find-threshold\n\nArguments:\n  MODEL          Model name or path  [required]\n  DATA_PATH      Location of binary evaluation data in .spacy format  [required]\n  PIPE_NAME      Name of pipe to examine thresholds for  [required]\n  THRESHOLD_KEY  Key of threshold attribute in component's configuration\n                 [required]\n  SCORES_KEY     Metric to optimize  [required]\n\nOptions:\n  -n, --n_trials INTEGER  Number of trials to determine optimal thresholds\n                          [default: 11]\n  -c, --code PATH         Path to Python file with additional code (registered\n                          functions) to be imported\n  -g, --gpu-id INTEGER    GPU ID or -1 for CPU  [default: -1]\n  -G, --gold-preproc      Use gold preprocessing\n  -V, -VV, --verbose      Display more information for debugging purposes\n  --help                  Show this message and exit.\n",
    "info": "Usage: python -m spacy info [OPTIONS] [MODEL]\n\n  Print info about spaCy installation. If a pipeline is specified as an\n  argument, print its meta information. Flag --markdown prints details in\n  Markdown for easy copy-pasting to GitHub issues.\n\n  Flag --url prints only the download URL of the most recent compatible version\n  of the pipeline.\n\n  DOCS: https://spacy.io/api/cli#info\n\nArguments:\n  [MODEL]  Optional loadable spaCy pipeline\n\nOptions:\n  -md, --markdown     Generate Markdown for GitHub issues\n  -s, -S, --silent    Don't print anything (just return)\n  -e, --exclude TEXT  Comma-separated keys to exclude from the print-out\n                      [default: labels]\n  -u, --url           Print the URL to download the most recent compatible\n                      version of the pipeline\n  --help              Show this message and exit.\n",
    "init config": "Usage: python -m spacy init config [OPTIONS] OUTPUT_FILE\n\n  Generate a starter config file for training. Based on your requirements\n  specified via the CLI arguments, this command generates a config with the\n  optimal settings for your use case. This includes the choice of architecture,\n  pretrained weights and related hyperparameters.\n\n  DOCS: https://spacy.io/api/cli#init-config\n\nArguments:\n  OUTPUT_FILE  File to save the config to or - for stdout (will only output\n               config and no additional logging info)  [required]\n\nOptions:\n  -l, --lang TEXT                 Two-letter code of the language to use\n                                  [default: en]\n  -p, --pipeline TEXT             Comma-separated names of trainable pipeline\n                                  components to include (without 'tok2vec' or\n                                  'transformer')  [default: tagger,parser,ner]\n  -o, --optimize [efficiency|accuracy]\n                                  Whether to optimize for efficiency (faster\n                                  inference, smaller model, lower memory\n                                  consumption) or higher accuracy (potentially\n                                  larger and slower model). This will impact the\n                                  choice of architecture, pretrained weights and\n                                  related hyperparameters.  [default:\n                                  efficiency]\n  -G, --gpu                       Whether the model can run on GPU. This will\n                                  impact the choice of architecture, pretrained\n                                  weights and related hyperparameters.\n  -pt, --pretraining              Include config for pretraining (with 'spacy\n                                  pretrain')\n  -F, --force                     Force overwriting the output file\n  --help                          Show this message and exit.\n",
    "init fill-config": "Usage: python -m spacy init fill-config [OPTIONS] BASE_PATH [OUTPUT_FILE]\n\n  Fill partial config file with default values. Will add all missing settings\n  from the default config and will create all objects, check the registered\n  functions for their default values and update the base config. This command\n  can be used with a config generated via the training quickstart widget:\n  https://spacy.io/usage/training#quickstart\n\n  DOCS: https://spacy.io/api/cli#init-fill-config\n\nArguments:\n  BASE_PATH      Path to base config to fill  [required]\n  [OUTPUT_FILE]  Path to output .cfg file (or - for stdout)  [default: -]\n\nOptions:\n  -pt, --pretraining            Include config for pretraining (with 'spacy\n                                pretrain')\n  -D, --diff                    Print a visual diff highlighting the changes\n  -c, --code-path, --code PATH  Path to Python file with additional code\n                                (registered functions) to be imported\n  --help                        Show this message and exit.\n",
    "init labels": "Usage: python -m spacy init labels [OPTIONS] CONFIG_PATH OUTPUT_PATH\n\n  Generate JSON files for the labels in the data. This helps speed up the\n  training process, since spaCy won't have to preprocess the data to extract the\n  labels.\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n  OUTPUT_PATH  Output directory for the labels  [required]\n\nOptions:\n  -c, --code PATH       Path to Python file with additional code (registered\n                        functions) to be imported\n  -V, -VV, --verbose    Display more information for debugging purposes\n  -g, --gpu-id INTEGER  GPU ID or -1 for CPU  [default: -1]\n  --help                Show this message and exit.\n",
    "init nlp": "Usage: python -m spacy init nlp [OPTIONS] CONFIG_PATH OUTPUT_PATH\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n  OUTPUT_PATH  Output directory for the prepared data  [required]\n\nOptions:\n  -c, --code PATH       Path to Python file with additional code (registered\n                        functions) to be imported\n  -V, -VV, --verbose    Display more information for debugging purposes\n  -g, --gpu-id INTEGER  GPU ID or -1 for CPU  [default: -1]\n  --help                Show this message and exit.\n",
    "init vectors": "Usage: python -m spacy init vectors [OPTIONS] LANG VECTORS_LOC OUTPUT_DIR\n\n  Convert word vectors for use with spaCy. Will export an nlp object that you\n  can use in the [initialize] block of your config to initialize a model with\n  vectors.\n\nArguments:\n  LANG         The language of the nlp object to create  [required]\n  VECTORS_LOC  Vectors file in Word2Vec format  [required]\n  OUTPUT_DIR   Pipeline output directory  [required]\n\nOptions:\n  -p, --prune INTEGER     Optional number of vectors to prune to  [default: -1]\n  -t, --truncate INTEGER  Optional number of vectors to truncate to when reading\n                          in vectors file  [default: 0]\n  -m, --mode TEXT         Vectors mode: default or floret  [default: default]\n  -n, --name TEXT         Optional name for the word vectors, e.g.\n                          en_core_web_lg.vectors\n  -V, -VV, --verbose      Display more information for debugging purposes\n  -a, --attr TEXT         Optional token attribute to use for vectors, e.g.\n                          LOWER or NORM  [default: ORTH]\n  --help                  Show this message and exit.\n",
    "link": "Usage: python -m spacy link [OPTIONS] ARGS KWARGS\n\n  As of spaCy v3.0, symlinks like \"en\" are not supported anymore. You can load\n  trained pipeline packages using their full names or from a directory path.\n  (DEPRECATED)\n\nArguments:\n  ARGS    [required]\n  KWARGS  [required]\n\nOptions:\n  --help  Show this message and exit.\n",
    "package": "Usage: python -m spacy package [OPTIONS] INPUT_DIR OUTPUT_DIR\n\n  Generate an installable Python package for a pipeline. Includes binary data,\n  meta and required installation files. A new directory will be created in the\n  specified output directory, and the data will be copied over. If --create-meta\n  is set and a meta.json already exists in the output directory, the existing\n  values will be used as the defaults in the command-line prompt. After\n  packaging, \"python -m build --sdist\" is run in the package directory, which\n  will create a .tar.gz archive that can be installed via \"pip install\".\n\n  If additional code files are provided (e.g. Python files containing custom\n  registered functions like pipeline components), they are copied into the\n  package and imported in the __init__.py.\n\n  DOCS: https://spacy.io/api/cli#package\n\nArguments:\n  INPUT_DIR   Directory with pipeline data  [required]\n  OUTPUT_DIR  Output parent directory  [required]\n\nOptions:\n  -c, --code TEXT                 Comma-separated paths to Python file with\n                                  additional code (registered functions) to be\n                                  included in the package\n  -m, --meta-path, --meta FILE    Path to meta.json\n  -C, --create-meta               Create meta.json, even if one exists\n  -n, --name TEXT                 Package name to override meta\n  -v, --version TEXT              Package version to override meta\n  -b, --build TEXT                Comma-separated formats to build: sdist and/or\n                                  wheel, or none.  [default: sdist]\n  -f, -F, --force                 Force overwriting existing data in output\n                                  directory\n  -R, -R, --require-parent / --no-require-parent\n                                  Include the parent package (e.g. spacy) in the\n                                  requirements  [default: require-parent]\n  --help                          Show this message and exit.\n",
    "pretrain": "Usage: python -m spacy pretrain [OPTIONS] CONFIG_PATH OUTPUT_DIR\n\n  Pre-train the 'token-to-vector' (tok2vec) layer of pipeline components, using\n  an approximate language-modelling objective. Two objective types are\n  available, vector-based and character-based.\n\n  In the vector-based objective, we load word vectors that have been trained\n  using a word2vec-style distributional similarity algorithm, and train a\n  component like a CNN, BiLSTM, etc to predict vectors which match the\n  pretrained ones. The weights are saved to a directory after each epoch. You\n  can then pass a path to one of these pretrained weights files to the 'spacy\n  train' command.\n\n  This technique may be especially helpful if you have little labelled data.\n  However, it's still quite experimental, so your mileage may vary.\n\n  To load the weights back in during 'spacy train', you need to ensure all\n  settings are the same between pretraining and training. Ideally, this is done\n  by using the same config file for both commands.\n\n  DOCS: https://spacy.io/api/cli#pretrain\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n  OUTPUT_DIR   Directory to write weights to on each epoch  [required]\n\nOptions:\n  -c, --code PATH              Path to Python file with additional code\n                               (registered functions) to be imported\n  -r, --resume-path PATH       Path to pretrained weights from which to resume\n                               pretraining\n  -er, --epoch-resume INTEGER  The epoch to resume counting from when using\n                               --resume-path. Prevents unintended overwriting of\n                               existing weight files.\n  -g, --gpu-id INTEGER         GPU ID or -1 for CPU  [default: -1]\n  -L, --skip-last              Skip saving model-last.bin\n  --help                       Show this message and exit.\n",
    "profile": "Usage: python -m spacy profile [OPTIONS] MODEL [INPUTS]\n\n  Profile which functions take the most time in a spaCy pipeline. Input should\n  be formatted as one JSON object per line with a key \"text\". It can either be\n  provided as a JSONL file, or be read from sys.sytdin. If no input file is\n  specified, the IMDB dataset is loaded via Thinc.\n\n  DOCS: https://spacy.io/api/cli#debug-profile\n\nArguments:\n  MODEL     Trained pipeline to load  [required]\n  [INPUTS]  Location of input file. '-' for stdin.\n\nOptions:\n  -n, --n-texts INTEGER  Maximum number of texts to use if available  [default:\n                         10000]\n  --help                 Show this message and exit.\n",
    "project assets": "Usage: python -m spacy project assets [OPTIONS] [PROJECT_DIR]\n\n  Fetch project assets like datasets and pretrained weights. Assets are defined\n  in the \"assets\" section of the project.yml. If a checksum is provided in the\n  project.yml, the file is only downloaded if no local file with the same\n  checksum exists.\n\n  DOCS: https://github.com/explosion/weasel/tree/main/docs/tutorial/directory-\n  and-assets.md\n\nArguments:\n  [PROJECT_DIR]  Path to cloned project. Defaults to current working directory.\n                 [default: /Users/matt/repos/spacy-monorepo/spacy]\n\nOptions:\n  -S, --sparse  Use sparse checkout for assets provided via Git, to only check\n                out and clone the files needed. Requires Git v22.2+.\n  -e, --extra   Download all assets, including those marked as 'extra'.\n  --help        Show this message and exit.\n",
    "project clone": "Usage: python -m spacy project clone [OPTIONS] NAME [DEST]\n\n  Clone a project template from a repository. Calls into \"git\" and will only\n  download the files from the given subdirectory. The GitHub repo defaults to\n  the official Weasel template repo, but can be customized (including using a\n  private repo).\n\n  DOCS: https://github.com/explosion/weasel/tree/main/docs/cli.md#clipboard-\n  clone\n\nArguments:\n  NAME    The name of the template to clone  [required]\n  [DEST]  Where to clone the project. Defaults to current working directory\n\nOptions:\n  -r, --repo TEXT    The repository to clone from  [default:\n                     https://github.com/explosion/projects]\n  -b, --branch TEXT  The branch to clone from. If not provided, will attempt\n                     main, master\n  -S, --sparse       Use sparse Git checkout to only check out and clone the\n                     files needed. Requires Git v22.2+.\n  --help             Show this message and exit.\n",
    "project document": "Usage: python -m spacy project document [OPTIONS] [PROJECT_DIR]\n\n  Auto-generate a README.md for a project. If the content is saved to a file,\n  hidden markers are added so you can add custom content before or after the\n  auto-generated section and only the auto-generated docs will be replaced when\n  you re-run the command.\n\n  DOCS: https://github.com/explosion/weasel/tree/main/docs/cli.md#closed_book-\n  document\n\nArguments:\n  [PROJECT_DIR]  Path to cloned project. Defaults to current working directory.\n                 [default: /Users/matt/repos/spacy-monorepo/spacy]\n\nOptions:\n  -o, --output PATH  Path to output Markdown file for output. Defaults to - for\n                     standard output  [default: -]\n  -NE, --no-emoji    Don't use emoji\n  --help             Show this message and exit.\n",
    "project dvc": "Usage: python -m spacy project dvc [OPTIONS] [PROJECT_DIR] [WORKFLOW]\n\n  Auto-generate Data Version Control (DVC) config. A DVC project can only define\n  one pipeline, so you need to specify one workflow defined in the project.yml.\n  If no workflow is specified, the first defined workflow is used. The DVC\n  config will only be updated if the project.yml changed.\n\n  DOCS: https://github.com/explosion/weasel/tree/main/docs/cli.md#repeat-dvc\n\nArguments:\n  [PROJECT_DIR]  Location of project directory. Defaults to current working\n                 directory.  [default: /Users/matt/repos/spacy-monorepo/spacy]\n  [WORKFLOW]     Name of workflow defined in project.yml. Defaults to first\n                 workflow if not set.\n\nOptions:\n  -V, --verbose  Print more info\n  -q, --quiet    Print less info\n  -F, --force    Force update DVC config\n  --help         Show this message and exit.\n",
    "project pull": "Usage: python -m spacy project pull [OPTIONS] [REMOTE] [PROJECT_DIR]\n\n  Retrieve available precomputed outputs from a remote storage. You can alias\n  remotes in your project.yml by mapping them to storage paths. A storage can be\n  anything that the smart_open library can upload to, e.g. AWS, Google Cloud\n  Storage, SSH, local directories etc.\n\n  DOCS: https://github.com/explosion/weasel/tree/main/docs/cli.md#arrow_down-\n  push\n\nArguments:\n  [REMOTE]       Name or path of remote storage  [default: default]\n  [PROJECT_DIR]  Location of project directory. Defaults to current working\n                 directory.  [default: /Users/matt/repos/spacy-monorepo/spacy]\n\nOptions:\n  --help  Show this message and exit.\n",
    "project push": "Usage: python -m spacy project push [OPTIONS] [REMOTE] [PROJECT_DIR]\n\n  Persist outputs to a remote storage. You can alias remotes in your project.yml\n  by mapping them to storage paths. A storage can be anything that the\n  smart_open library can upload to, e.g. AWS, Google Cloud Storage, SSH, local\n  directories etc.\n\n  DOCS: https://github.com/explosion/weasel/tree/main/docs/cli.md#arrow_up-push\n\nArguments:\n  [REMOTE]       Name or path of remote storage  [default: default]\n  [PROJECT_DIR]  Location of project directory. Defaults to current working\n                 directory.  [default: /Users/matt/repos/spacy-monorepo/spacy]\n\nOptions:\n  --help  Show this message and exit.\n",
    "train": "Usage: python -m spacy train [OPTIONS] CONFIG_PATH\n\n  Train or update a spaCy pipeline. Requires data in spaCy's binary format. To\n  convert data from other formats, use the `spacy convert` command. The config\n  file includes all settings and hyperparameters used during training. To\n  override settings in the config, e.g. settings that point to local paths or\n  that you want to experiment with, you can override them as command line\n  options. For instance, --training.batch_size 128 overrides the value of\n  \"batch_size\" in the block \"[training]\". The --code argument lets you pass in a\n  Python file that's imported before training. It can be used to register custom\n  functions and architectures that can then be referenced in the config.\n\n  DOCS: https://spacy.io/api/cli#train\n\nArguments:\n  CONFIG_PATH  Path to config file  [required]\n\nOptions:\n  -o, --output, --output-path PATH\n                                  Output directory to store trained pipeline in\n  -c, --code PATH                 Path to Python file with additional code\n                                  (registered functions) to be imported\n  -V, -VV, --verbose              Display more information for debugging\n                                  purposes\n  -g, --gpu-id INTEGER            GPU ID or -1 for CPU  [default: -1]\n  --help                          Show this message and exit.\n",
    "validate": "Usage: python -m spacy validate [OPTIONS]\n\n  Validate the currently installed pipeline packages and spaCy version. Checks\n  if the installed packages are compatible and shows upgrade instructions if\n  available. Should be run after `pip install -U spacy`.\n\n  DOCS: https://spacy.io/api/cli#validate\n\nOptions:\n  --help  Show this message and exit.\n"
  },
  "errors": {
    "missing_command": "Usage: python -m spacy [OPTIONS] COMMAND [ARGS]...\nTry 'python -m spacy --help' for help.\n\nError: Missing command.\n",
    "unknown_command": "Usage: python -m spacy [OPTIONS] COMMAND [ARGS]...\nTry 'python -m spacy --help' for help.\n\nError: No such command '__SPACY_UNKNOWN_COMMAND__'.\n",
    "unknown_subcommand": {
      "benchmark": "Usage: python -m spacy benchmark [OPTIONS] COMMAND [ARGS]...\nTry 'python -m spacy benchmark --help' for help.\n\nError: No such command '__SPACY_UNKNOWN_SUBCOMMAND__'.\n",
      "debug": "Usage: python -m spacy debug [OPTIONS] COMMAND [ARGS]...\nTry 'python -m spacy debug --help' for help.\n\nError: No such command '__SPACY_UNKNOWN_SUBCOMMAND__'.\n",
      "init": "Usage: python -m spacy init [OPTIONS] COMMAND [ARGS]...\nTry 'python -m spacy init --help' for help.\n\nError: No such command '__SPACY_UNKNOWN_SUBCOMMAND__'.\n",
      "project": "Usage: python -m spacy project [OPTIONS] COMMAND [ARGS]...\nTry 'python -m spacy project --help' for help.\n\nError: No such command '__SPACY_UNKNOWN_SUBCOMMAND__'.\n"
    }
  },
  "group_help": {
    "benchmark": "Usage: python -m spacy benchmark [OPTIONS] COMMAND [ARGS]...\n\n  Commands for benchmarking pipelines.\n\nOptions:\n  --help  Show this message and exit.\n\nCommands:\n  accuracy  Evaluate a trained pipeline.\n  speed     Benchmark a pipeline.\n",
    "debug": "Usage: python -m spacy debug [OPTIONS] COMMAND [ARGS]...\n\n  Suite of helpful commands for debugging and profiling. Includes commands to\n  check and validate your config files, training and evaluation data, and custom\n  model implementations.\n\nOptions:\n  --help  Show this message and exit.\n\nCommands:\n  data         Analyze, debug and validate your training and development data.\n  profile      Profile which functions take the most time in a spaCy pipeline.\n  config       Debug a config file and show validation errors.\n  diff-config  Show a diff of a config file with respect to spaCy's...\n  model        Analyze a Thinc model implementation.\n",
    "init": "Usage: python -m spacy init [OPTIONS] COMMAND [ARGS]...\n\n  Commands for initializing configs and pipeline packages.\n\nOptions:\n  --help  Show this message and exit.\n\nCommands:\n  config       Generate a starter config file for training.\n  fill-config  Fill partial config file with default values.\n  vectors      Convert word vectors for use with spaCy.\n  labels       Generate JSON files for the labels in the data.\n",
    "project": "Usage: python -m spacy project [OPTIONS] COMMAND [ARGS]...\n\n  Command-line interface for spaCy projects and templates. You'd typically start\n  by cloning a project template to a local directory and fetching its assets\n  like datasets etc. See the project's project.yml for the available commands.\n\nOptions:\n  --help  Show this message and exit.\n\nCommands:\n  assets    Fetch project assets like datasets and pretrained weights.\n  clone     Clone a project template from a repository.\n  document  Auto-generate a README.md for a project.\n  dvc       Auto-generate Data Version Control (DVC) config.\n  run       Run a named command or workflow defined in the project.yml.\n  pull      Retrieve available precomputed outputs from a remote storage.\n  push      Persist outputs to a remote storage.\n"
  },
  "hidden_group_commands": {
    "benchmark": [],
    "debug": [],
    "init": [
      "nlp"
    ],
    "project": []
  },
  "hidden_top_level": [
    "link",
    "debug-data",
    "profile"
  ],
  "known_groups": {
    "benchmark": [
      "accuracy",
      "speed"
    ],
    "debug": [
      "config",
      "data",
      "diff-config",
      "model",
      "profile"
    ],
    "init": [
      "config",
      "fill-config",
      "labels",
      "nlp",
      "vectors"
    ],
    "project": [
      "assets",
      "clone",
      "document",
      "dvc",
      "pull",
      "push",
      "run"
    ]
  },
  "known_top_level": [
    "apply",
    "assemble",
    "benchmark",
    "convert",
    "debug",
    "debug-data",
    "download",
    "evaluate",
    "find-function",
    "find-threshold",
    "info",
    "init",
    "link",
    "package",
    "pretrain",
    "profile",
    "project",
    "train",
    "validate"
  ],
  "root_help": "Usage: python -m spacy [OPTIONS] COMMAND [ARGS]...\n\n  spaCy Command-line Interface\n\n  DOCS: https://spacy.io/api/cli\n\nOptions:\n  --install-completion  Install completion for the current shell.\n  --show-completion     Show completion for the current shell, to copy it or\n                        customize the installation.\n  --help                Show this message and exit.\n\nCommands:\n  download        Download compatible trained pipeline from the default...\n  info            Print info about spaCy installation.\n  apply           Apply a trained pipeline to documents to get predictions.\n  assemble        Assemble a spaCy pipeline from a config file.\n  convert         Convert files into json or DocBin format for training.\n  evaluate        Evaluate a trained pipeline.\n  find-function   Find the module, path and line number to the file the...\n  find-threshold  Runs prediction trials for a trained model with varying...\n  package         Generate an installable Python package for a pipeline.\n  pretrain        Pre-train the 'token-to-vector' (tok2vec) layer of...\n  train           Train or update a spaCy pipeline.\n  validate        Validate the currently installed pipeline packages and...\n  debug           Suite of helpful commands for debugging and profiling.\n  benchmark       Commands for benchmarking pipelines.\n  init            Commands for initializing configs and pipeline packages.\n  project         Command-line interface for spaCy projects and templates.\n"
}