Specifies which fingerprint keys decide whether this sub-model is retrained.
Container for the results of a fingerprint comparison.
Trueif the NLU model should be retrained.
Trueif the Core model should be retrained.
Trueif the responses in the domain should be updated.
Trueif a training of all parts is forced.
Check if anything has to be retrained.
Check if the Core model has to be updated.
Check if the responses have to be updated.
Check if the NLU model has to be updated.
Get a model and unpack it. Raises a
ModelNotFound exception if
no model could be found at the provided path.
model_path- Path to the zipped model. If it's a directory, the latest trained model is returned.
Path to the unpacked model.
Get the latest model from a path.
model_path- Path to a directory containing zipped models.
Path to latest model in the given directory.
Unpack a zipped Rasa model.
model_file- Path to zipped model.
working_directory- Location where the model should be unpacked to. If
Nonea temporary directory will be created.
Path to unpacked Rasa model.
Return paths for Core and NLU model directories, if they exist.
If neither directories exist, a
ModelNotFound exception is raised.
unpacked_model_path- Path to unpacked Rasa model.
Tuple (path to Core subdirectory if it exists or
path to NLU subdirectory if it exists or
Create a zipped Rasa model from trained model files.
training_directory- Path to the directory which contains the trained model files.
output_filename- Name of the zipped model file to be created.
fingerprint- A unique fingerprint to identify the model version.
Path to zipped model.
Create a hash for the project in the current working directory.
Create a model fingerprint from its used configuration and training data.
file_importer- File importer which provides the training data and model config.
Load a persisted fingerprint.
model_path- Path to directory containing the fingerprint.
The fingerprint or an empty dict if no fingerprint was found.
Persist a model fingerprint.
output_path- Directory in which the fingerprint should be saved.
fingerprint- The fingerprint to be persisted.
Check whether the fingerprint of a section has changed.
Move two model directories.
source- The original folder which should be merged in another.
target- The destination folder where it should be moved to.
True if the merge was successful, else
Check which components of a model should be retrained.
new_fingerprint- The fingerprint of the new model to be trained.
old_model- Path to the old zipped model file.
train_path- Path to the directory in which the new model will be trained.
has_e2e_examples- Whether the new training data contains e2e examples.
force_training- Indicates if the model needs to be retrained even if the data has not changed.
A FingerprintComparisonResult object indicating whether Rasa Core and/or Rasa NLU needs to be retrained or not.
Checks if components of a model can be finetuned with incremental training.
last_fingerprint- The fingerprint of the old model to potentially be fine-tuned.
new_fingerprint- The fingerprint of the new model.
core- Check sections for finetuning a core model.
nlu- Check sections for finetuning an nlu model.
True if the old model can be finetuned,
Compress a trained model.
fingerprint- fingerprint of the model
output_directory- path to the directory in which the model should be stored
train_path- path to uncompressed model
fixed_model_name- name of the compressed model file
model_prefix- prefix of the compressed model file
Returns- path to 'tar.gz' model file
Overwrites the domain of an unpacked model with a new domain.
importer- Importer which provides the new domain.
unpacked_model_path- Path to the unpacked model.
Gets validated path for model to finetune.
previous_model_file- Path to model file which should be used for finetuning or a directory in case the latest trained model should be used.
Path to model archive.
None if there is no model.