def _train_nlu_with_validated_data(
config: Text,
nlu_data_directory: Text,
output: Text,
train_path: Optional[Text] = None,
fixed_model_name: Optional[Text] = None,
) -> Optional[Text]:
"""Train NLU with validated training and config data."""
import rasa.nlu.train
_train_path = train_path or tempfile.mkdtemp()
print_color("Training NLU model...", color=bcolors.OKBLUE)
_, nlu_model, _ = rasa.nlu.train(
config, nlu_data_directory, _train_path, fixed_model_name="nlu"
)
print_color("NLU model training completed.", color=bcolors.OKBLUE)
if train_path is None:
# Only NLU was trained
new_fingerprint = model.model_fingerprint(config, nlu_data=nlu_data_directory)
return _package_model(
new_fingerprint=new_fingerprint,
output_path=output,
train_path=_train_path,
fixed_model_name=fixed_model_name,
model_prefix="nlu-",
)
return _train_path
def trained_nlu_model(request):
cfg = RasaNLUModelConfig({"pipeline": "keyword"})
trainer = Trainer(cfg)
td = training_data.load_data(DEFAULT_DATA_PATH)
trainer.train(td)
model_path = trainer.persist(NLU_MODEL_PATH)
nlu_data = data.get_nlu_directory(DEFAULT_DATA_PATH)
output_path = os.path.join(NLU_MODEL_PATH, NLU_MODEL_NAME)
new_fingerprint = model.model_fingerprint(NLU_DEFAULT_CONFIG_PATH,
nlu_data=nlu_data)
model.create_package_rasa(model_path, output_path, new_fingerprint)
def fin():
if os.path.exists(NLU_MODEL_PATH):
shutil.rmtree(NLU_MODEL_PATH)
if os.path.exists(output_path):
shutil.rmtree(output_path)
request.addfinalizer(fin)
return output_path
def _core_model_for_finetuning(
model_to_finetune: Text,
file_importer: TrainingDataImporter,
finetuning_epoch_fraction: float = 1.0,
) -> Optional[Agent]:
path_to_archive = model.get_model_for_finetuning(model_to_finetune)
if not path_to_archive:
return None
rasa.shared.utils.cli.print_info(
f"Loading Core model from {path_to_archive} for finetuning...", )
with model.unpack_model(path_to_archive) as unpacked:
new_fingerprint = model.model_fingerprint(file_importer)
old_fingerprint = model.fingerprint_from_path(unpacked)
if not model.can_finetune(old_fingerprint, new_fingerprint, core=True):
rasa.shared.utils.cli.print_error_and_exit(
"Core model can not be finetuned.")
config = file_importer.get_config()
agent = Agent.load(
unpacked,
new_config=config,
finetuning_epoch_fraction=finetuning_epoch_fraction,
)
# Agent might be empty if no underlying Core model was found.
if agent.domain is not None and agent.policy_ensemble is not None:
return agent
return None
def _nlu_model_for_finetuning(
model_to_finetune: Text,
file_importer: TrainingDataImporter,
finetuning_epoch_fraction: float = 1.0,
called_from_combined_training: bool = False,
) -> Optional[Interpreter]:
path_to_archive = model.get_model_for_finetuning(model_to_finetune)
if not path_to_archive:
return None
rasa.shared.utils.cli.print_info(
f"Loading NLU model from {path_to_archive} for finetuning...", )
with model.unpack_model(path_to_archive) as unpacked:
_, old_nlu = model.get_model_subdirectories(unpacked)
new_fingerprint = model.model_fingerprint(file_importer)
old_fingerprint = model.fingerprint_from_path(unpacked)
if not model.can_finetune(
old_fingerprint,
new_fingerprint,
nlu=True,
core=called_from_combined_training,
):
rasa.shared.utils.cli.print_error_and_exit(
"NLU model can not be finetuned.")
config = file_importer.get_config()
loaded_model_to_finetune = Interpreter.load(
old_nlu,
new_config=config,
finetuning_epoch_fraction=finetuning_epoch_fraction,
)
if not loaded_model_to_finetune:
return None
return loaded_model_to_finetune
def train_nlu(config: Text, nlu_data: Text, output: Text,
train_path: Optional[Text]) -> Optional["Interpreter"]:
"""Trains a NLU model.
Args:
config: Path to the config file for NLU.
nlu_data: Path to the NLU training data.
output: Output path.
train_path: If `None` the model will be trained in a temporary
directory, otherwise in the provided directory.
Returns:
If `train_path` is given it returns the path to the model archive,
otherwise the path to the directory with the trained model files.
"""
import rasa_nlu
_train_path = train_path or tempfile.mkdtemp()
_, nlu_model, _ = rasa_nlu.train(config, nlu_data, _train_path,
project="",
fixed_model_name="nlu")
if not train_path:
nlu_data = data.get_nlu_directory(nlu_data)
output_path = create_output_path(output, prefix="nlu-")
new_fingerprint = model.model_fingerprint(config, nlu_data=nlu_data)
model.create_package_rasa(_train_path, output_path, new_fingerprint)
print_success("Your Rasa NLU model is trained and saved at '{}'."
"".format(output_path))
return nlu_model
async def train(request):
# if set will not generate a model name but use the passed one
model_name = request.args.get("model", None)
try:
model_config, data_dict = extract_data_and_config(request)
except Exception as e:
logger.debug(traceback.format_exc())
raise ErrorResponse(
500,
"ServerError",
"An unexpected error occurred.",
details={"error": str(e)},
)
data_file = dump_to_data_file(data_dict)
config_file = dump_to_data_file(model_config, "_config")
try:
path_to_model = await data_router.start_train_process(
data_file, RasaNLUModelConfig(model_config), model_name)
# store trained model as tar.gz file
output_path = create_model_path(model_name, path_to_model)
nlu_data = data.get_nlu_directory(data_file)
new_fingerprint = model.model_fingerprint(config_file,
nlu_data=nlu_data)
model.create_package_rasa(path_to_model, output_path,
new_fingerprint)
logger.info("Rasa NLU model trained and persisted to '{}'.".format(
output_path))
await data_router.load_model(output_path)
return await response.file(output_path)
except MaxWorkerProcessError as e:
raise ErrorResponse(
403,
"NoFreeProcess",
"No process available for training.",
details={"error": str(e)},
)
except InvalidModelError as e:
raise ErrorResponse(
404,
"ModelNotFound",
"Model '{}' not found.".format(model_name),
details={"error": str(e)},
)
except TrainingException as e:
logger.debug(traceback.format_exc())
raise ErrorResponse(
500,
"ServerError",
"An unexpected error occurred.",
details={"error": str(e)},
)
def test_can_finetune_min_version(
project: Text,
monkeypatch: MonkeyPatch,
old_model_version: Text,
min_compatible_version: Text,
can_tune: bool,
):
importer = _project_files(project)
monkeypatch.setattr(rasa.constants, "MINIMUM_COMPATIBLE_VERSION",
min_compatible_version)
monkeypatch.setattr(rasa, "__version__", old_model_version)
old_fingerprint = model_fingerprint(importer)
new_fingerprint = model_fingerprint(importer)
with mock.patch("rasa.model.MINIMUM_COMPATIBLE_VERSION",
min_compatible_version):
assert can_finetune(old_fingerprint, new_fingerprint) == can_tune
def test_fingerprinting_additional_action(project: Text):
importer = _project_files(project)
old_fingerprint = model_fingerprint(importer)
old_domain = importer.get_domain()
domain_with_new_action = old_domain.as_dict()
domain_with_new_action[KEY_RESPONSES]["utter_new"] = [{"text": "hi"}]
domain_with_new_action = Domain.from_dict(domain_with_new_action)
importer.get_domain = lambda: domain_with_new_action
new_fingerprint = model_fingerprint(importer)
assert (old_fingerprint[FINGERPRINT_DOMAIN_WITHOUT_NLG_KEY] !=
new_fingerprint[FINGERPRINT_DOMAIN_WITHOUT_NLG_KEY])
assert old_fingerprint[FINGERPRINT_NLG_KEY] != new_fingerprint[
FINGERPRINT_NLG_KEY]
def _train_nlu_with_validated_data(
config: Dict[Text, Text],
nlu_data_directory: Text,
output: Text,
train_path: Optional[Text] = None,
fixed_model_name: Optional[Text] = None,
retrain_nlu: Union[bool, List[Text]] = True
) -> Optional[Text]:
"""Train NLU with validated training and config data."""
import rasa.nlu.train
import re
with ExitStack() as stack:
models = {}
from rasa.nlu import config as cfg_loader
if train_path:
# If the train path was provided, do nothing on exit.
_train_path = train_path
else:
# Otherwise, create a temp train path and clean it up on exit.
_train_path = stack.enter_context(TempDirectoryPath(tempfile.mkdtemp()))
pattern = r'(\w\w)*(?=\.)'
for file in os.listdir(nlu_data_directory):
lang = re.search(pattern, file).groups()[0]
if isinstance(retrain_nlu, bool) and retrain_nlu or lang in retrain_nlu:
nlu_file_path = os.path.join(nlu_data_directory, file)
print_color("Start training {} NLU model ...".format(lang), color=bcolors.OKBLUE)
nlu_config = cfg_loader.load(config[lang])
nlu_config.language = lang
_, models[lang], _ = rasa.nlu.train(
nlu_config, nlu_file_path, _train_path, fixed_model_name="nlu-{}".format(lang)
)
else:
print_color("{} NLU data didn't change, skipping training...".format(lang), color=bcolors.OKBLUE)
print_color("NLU model training completed.", color=bcolors.OKBLUE)
if train_path is None:
# Only NLU was trained
new_fingerprint = model.model_fingerprint(
config, nlu_data=nlu_data_directory
)
return _package_model(
new_fingerprint=new_fingerprint,
output_path=output,
train_path=_train_path,
fixed_model_name=fixed_model_name,
model_prefix="nlu-",
)
return _train_path
async def train_comparison_models(
story_file: Text,
domain: Text,
output_path: Text = "",
exclusion_percentages: Optional[List] = None,
policy_configs: Optional[List] = None,
runs: int = 1,
dump_stories: bool = False,
additional_arguments: Optional[Dict] = None,
):
"""Train multiple models for comparison of policies"""
from rasa import model
from rasa.importers.importer import TrainingDataImporter
exclusion_percentages = exclusion_percentages or []
policy_configs = policy_configs or []
for r in range(runs):
logging.info("Starting run {}/{}".format(r + 1, runs))
for current_run, percentage in enumerate(exclusion_percentages, 1):
for policy_config in policy_configs:
file_importer = TrainingDataImporter.load_core_importer_from_config(
policy_config, domain, [story_file])
config_name = os.path.splitext(
os.path.basename(policy_config))[0]
logging.info("Starting to train {} round {}/{}"
" with {}% exclusion"
"".format(config_name, current_run,
len(exclusion_percentages), percentage))
with TempDirectoryPath(tempfile.mkdtemp()) as train_path:
_, new_fingerprint = await asyncio.gather(
train(
domain,
file_importer,
train_path,
policy_config=policy_config,
exclusion_percentage=percentage,
additional_arguments=additional_arguments,
dump_stories=dump_stories,
),
model.model_fingerprint(file_importer),
)
output_dir = os.path.join(output_path, "run_" + str(r + 1))
model_name = config_name + PERCENTAGE_KEY + str(percentage)
model.package_model(
fingerprint=new_fingerprint,
output_directory=output_dir,
train_path=train_path,
fixed_model_name=model_name,
)
def train_nlu(config: Text, nlu_data: Text, output: Text,
train_path: Optional[Text]) -> Optional[Text]:
"""Trains a NLU model.
Args:
config: Path to the config file for NLU.
nlu_data: Path to the NLU training data.
output: Output path.
train_path: If `None` the model will be trained in a temporary
directory, otherwise in the provided directory.
Returns:
If `train_path` is given it returns the path to the model archive,
otherwise the path to the directory with the trained model files.
"""
import rasa.nlu.train
config = get_valid_config(config, CONFIG_MANDATORY_KEYS_NLU)
if not train_path:
# training NLU only hence the training files still have to be selected
skill_imports = SkillSelector.load(config)
nlu_data_directory = data.get_nlu_directory(nlu_data, skill_imports)
else:
nlu_data_directory = nlu_data
if not os.listdir(nlu_data_directory):
print_error(
"No NLU data given. Please provide NLU data in order to train "
"a Rasa NLU model.")
return
_train_path = train_path or tempfile.mkdtemp()
print_color("Start training NLU model ...", color=bcolors.OKBLUE)
_, nlu_model, _ = rasa.nlu.train(config,
nlu_data_directory,
_train_path,
fixed_model_name="nlu")
print_color("Done.", color=bcolors.OKBLUE)
if not train_path:
output_path = create_output_path(output, prefix="nlu-")
new_fingerprint = model.model_fingerprint(config,
nlu_data=nlu_data_directory)
model.create_package_rasa(_train_path, output_path, new_fingerprint)
print_success(
"Your Rasa NLU model is trained and saved at '{}'.".format(
output_path))
return output_path
return _train_path
def test_fingerprinting_changed_response_text(project: Text):
importer = _project_files(project)
old_fingerprint = model_fingerprint(importer)
old_domain = importer.get_domain()
# Change NLG content but keep actions the same
domain_with_changed_nlg = old_domain.as_dict()
domain_with_changed_nlg[KEY_RESPONSES]["utter_greet"].append(
{"text": "hi"})
domain_with_changed_nlg = Domain.from_dict(domain_with_changed_nlg)
importer.get_domain = lambda: domain_with_changed_nlg
new_fingerprint = model_fingerprint(importer)
assert (old_fingerprint[FINGERPRINT_DOMAIN_WITHOUT_NLG_KEY] ==
new_fingerprint[FINGERPRINT_DOMAIN_WITHOUT_NLG_KEY])
assert old_fingerprint[FINGERPRINT_NLG_KEY] != new_fingerprint[
FINGERPRINT_NLG_KEY]
async def _train_core_with_validated_data(
domain: Domain,
config: Text,
story_directory: Text,
output: Text,
train_path: Optional[Text] = None,
fixed_model_name: Optional[Text] = None,
uncompress: bool = False,
kwargs: Optional[Dict] = None,
) -> Optional[Text]:
"""Train Core with validated training and config data."""
import rasa.core.train
if not os.listdir(story_directory):
print_error(
"No dialogue data given. Please provide dialogue data in order to "
"train a Rasa Core model.")
return
_train_path = train_path or tempfile.mkdtemp()
# normal (not compare) training
print_color("Start training dialogue model ...", color=bcolors.OKBLUE)
await rasa.core.train(
domain_file=domain,
stories_file=story_directory,
output_path=os.path.join(_train_path, "core"),
policy_config=config,
kwargs=kwargs,
)
print_color("Done.", color=bcolors.OKBLUE)
if not train_path:
# Only Core was trained.
output_path = create_output_path(output,
prefix="core-",
fixed_name=fixed_model_name)
new_fingerprint = model.model_fingerprint(config,
domain,
stories=story_directory)
model.create_package_rasa(_train_path, output_path, new_fingerprint)
if uncompress:
output_path = decompress(output_path)
print_success(
"Your Rasa Core model is trained and saved at '{}'.".format(
output_path))
return output_path
return _train_path
def test_create_fingerprint_from_invalid_paths(project, project_files):
project_files = _project_files(project, *project_files)
expected = _fingerprint([], [], rasa_version=rasa.__version__, stories=[], nlu=[])
actual = model_fingerprint(**project_files)
assert actual[FINGERPRINT_TRAINED_AT_KEY] is not None
del actual[FINGERPRINT_TRAINED_AT_KEY]
del expected[FINGERPRINT_TRAINED_AT_KEY]
assert actual == expected
async def train_core_async(
domain: Text,
config: Text,
stories: Text,
output: Text,
train_path: Optional[Text] = None,
kwargs: Optional[Dict] = None,
) -> Optional[Text]:
"""Trains a Core model.
Args:
domain: Path to the domain file.
config: Path to the config file for Core.
stories: Path to the Core training data.
output: Output path.
train_path: If `None` the model will be trained in a temporary
directory, otherwise in the provided directory.
kwargs: Additional training parameters.
Returns:
If `train_path` is given it returns the path to the model archive,
otherwise the path to the directory with the trained model files.
"""
import rasa.core.train
config = get_valid_config(config, CONFIG_MANDATORY_KEYS_CORE)
_train_path = train_path or tempfile.mkdtemp()
# normal (not compare) training
core_model = await rasa.core.train(
domain_file=domain,
stories_file=data.get_core_directory(stories),
output_path=os.path.join(_train_path, "core"),
policy_config=config,
kwargs=kwargs,
)
if not train_path:
# Only Core was trained.
stories = data.get_core_directory(stories)
output_path = create_output_path(output, prefix="core-")
new_fingerprint = model.model_fingerprint(config,
domain,
stories=stories)
model.create_package_rasa(_train_path, output_path, new_fingerprint)
print_success(
"Your Rasa Core model is trained and saved at '{}'.".format(
output_path))
return core_model
async def _train_core_with_validated_data(
domain: Domain,
config: Text,
story_directory: Text,
output: Text,
train_path: Optional[Text] = None,
fixed_model_name: Optional[Text] = None,
kwargs: Optional[Dict] = None,
) -> Optional[Text]:
"""Train Core with validated training and config data."""
import rasa.core.train
with ExitStack() as stack:
if train_path:
# If the train path was provided, do nothing on exit.
_train_path = train_path
else:
# Otherwise, create a temp train path and clean it up on exit.
_train_path = stack.enter_context(
TempDirectoryPath(tempfile.mkdtemp()))
# normal (not compare) training
print_color("Training Core model...", color=bcolors.OKBLUE)
await rasa.core.train(
domain_file=domain,
stories_file=story_directory,
output_path=os.path.join(_train_path, "core"),
policy_config=config,
kwargs=kwargs,
)
print_color("Core model training completed.", color=bcolors.OKBLUE)
if train_path is None:
# Only Core was trained.
new_fingerprint = model.model_fingerprint(config,
domain,
stories=story_directory)
return _package_model(
new_fingerprint=new_fingerprint,
output_path=output,
train_path=_train_path,
fixed_model_name=fixed_model_name,
model_prefix="core-",
)
return _train_path
def _train_nlu_with_validated_data(
config: Text,
nlu_data_directory: Text,
output: Text,
train_path: Optional[Text] = None,
fixed_model_name: Optional[Text] = None,
uncompress: bool = False,
) -> Optional[Text]:
"""Train NLU with validated training and config data."""
import rasa.nlu.train
if not os.listdir(nlu_data_directory):
print_error(
"No NLU data given. Please provide NLU data in order to train "
"a Rasa NLU model.")
return
_train_path = train_path or tempfile.mkdtemp()
print_color("Start training NLU model ...", color=bcolors.OKBLUE)
_, nlu_model, _ = rasa.nlu.train(config,
nlu_data_directory,
_train_path,
fixed_model_name="nlu")
print_color("Done.", color=bcolors.OKBLUE)
if not train_path:
# Only NLU was trained
output_path = create_output_path(output,
prefix="nlu-",
fixed_name=fixed_model_name)
new_fingerprint = model.model_fingerprint(config,
nlu_data=nlu_data_directory)
model.create_package_rasa(_train_path, output_path, new_fingerprint)
if uncompress:
output_path = decompress(output_path)
print_success(
"Your Rasa NLU model is trained and saved at '{}'.".format(
output_path))
return output_path
return _train_path
async def _train_core_with_validated_data(
domain: Domain,
config: Text,
story_directory: Text,
output: Text,
train_path: Optional[Text] = None,
fixed_model_name: Optional[Text] = None,
uncompress: bool = False,
kwargs: Optional[Dict] = None,
) -> Optional[Text]:
"""Train Core with validated training and config data."""
import rasa.core.train
_train_path = train_path or tempfile.mkdtemp()
# normal (not compare) training
print_color("Start training dialogue model ...", color=bcolors.OKBLUE)
await rasa.core.train(
domain_file=domain,
stories_file=story_directory,
output_path=os.path.join(_train_path, "core"),
policy_config=config,
kwargs=kwargs,
)
print_color("Done.", color=bcolors.OKBLUE)
if train_path is None:
# Only Core was trained.
new_fingerprint = model.model_fingerprint(config,
domain,
stories=story_directory)
return _package_model(
new_fingerprint=new_fingerprint,
output_path=output,
train_path=_train_path,
fixed_model_name=fixed_model_name,
model_prefix="core-",
uncompress=uncompress,
)
return _train_path
def _train_nlu_with_validated_data(
config: Text,
nlu_data_directory: Text,
output: Text,
train_path: Optional[Text] = None,
fixed_model_name: Optional[Text] = None,
) -> Optional[Text]:
"""Train NLU with validated training and config data."""
import rasa.nlu.train
with ExitStack() as stack:
if train_path:
# If the train path was provided, do nothing on exit.
_train_path = train_path
else:
# Otherwise, create a temp train path and clean it up on exit.
_train_path = stack.enter_context(
TempDirectoryPath(tempfile.mkdtemp()))
print_color("Training NLU model...", color=bcolors.OKBLUE)
_, nlu_model, _ = rasa.nlu.train(config,
nlu_data_directory,
_train_path,
fixed_model_name="nlu")
print_color("NLU model training completed.", color=bcolors.OKBLUE)
if train_path is None:
# Only NLU was trained
new_fingerprint = model.model_fingerprint(
config, nlu_data=nlu_data_directory)
return _package_model(
new_fingerprint=new_fingerprint,
output_path=output,
train_path=_train_path,
fixed_model_name=fixed_model_name,
model_prefix="nlu-",
)
return _train_path
def test_should_not_retrain_core(domain_path: Text, tmp_path: Path,
stack_config_path: Text):
# Don't use `stories_path` as checkpoints currently break fingerprinting
story_file = tmp_path / "simple_story.yml"
story_file.write_text("""
stories:
- story: test_story
steps:
- intent: greet
- action: utter_greet
""")
trained_model = train_core(domain_path, stack_config_path, str(story_file),
str(tmp_path))
importer = TrainingDataImporter.load_from_config(
stack_config_path, domain_path, training_data_paths=[str(story_file)])
new_fingerprint = model.model_fingerprint(importer)
result = model.should_retrain(new_fingerprint, trained_model, tmp_path)
assert not result.should_retrain_core()
def test_rasa_packaging(trained_model, project, use_fingerprint):
unpacked_model_path = get_model(trained_model)
os.remove(os.path.join(unpacked_model_path, FINGERPRINT_FILE_PATH))
if use_fingerprint:
fingerprint = model_fingerprint(**_project_files(project))
else:
fingerprint = None
tempdir = tempfile.mkdtemp()
output_path = os.path.join(tempdir, "test.tar.gz")
create_package_rasa(unpacked_model_path, output_path, fingerprint)
unpacked = get_model(output_path)
assert (os.path.exists(os.path.join(
unpacked, FINGERPRINT_FILE_PATH)) == use_fingerprint)
assert os.path.exists(os.path.join(unpacked, "core"))
assert os.path.exists(os.path.join(unpacked, "nlu"))
assert not os.path.exists(unpacked_model_path)
async def train_core_async(domain: Text, config: Text, stories: Text,
output: Text,
train_path: Optional[Text]) -> Optional[Text]:
"""Trains a Core model.
Args:
domain: Path to the domain file.
config: Path to the config file for Core.
stories: Path to the Core training data.
output: Output path.
train_path: If `None` the model will be trained in a temporary
directory, otherwise in the provided directory.
Returns:
If `train_path` is given it returns the path to the model archive,
otherwise the path to the directory with the trained model files.
"""
import rasa_core.train
# normal (not compare) training
core_model = await rasa_core.train(domain_file=domain,
stories_file=stories,
output_path=os.path.join(
train_path, "core"),
policy_config=config)
if not train_path:
# Only Core was trained.
stories = data.get_core_directory(stories)
output_path = create_output_path(output, prefix="core-")
new_fingerprint = model.model_fingerprint(config,
domain,
stories=stories)
model.create_package_rasa(train_path, output_path, new_fingerprint)
print_success("Your Rasa Core model is trained and saved at '{}'."
"".format(output_path))
return core_model
def test_rasa_packaging(trained_rasa_model: Text, project: Text,
use_fingerprint: bool, tmp_path: Path):
unpacked_model_path = get_model(trained_rasa_model)
os.remove(os.path.join(unpacked_model_path, FINGERPRINT_FILE_PATH))
if use_fingerprint:
fingerprint = model_fingerprint(_project_files(project))
else:
fingerprint = None
output_path = str(tmp_path / "test.tar.gz")
create_package_rasa(unpacked_model_path, output_path, fingerprint)
unpacked = get_model(output_path)
assert (os.path.exists(os.path.join(
unpacked, FINGERPRINT_FILE_PATH)) == use_fingerprint)
assert os.path.exists(
os.path.join(unpacked, DEFAULT_CORE_SUBDIRECTORY_NAME))
assert os.path.exists(os.path.join(unpacked, "nlu"))
assert not os.path.exists(unpacked_model_path)
def test_create_fingerprint_from_paths(project):
project_files = _project_files(project)
assert model_fingerprint(**project_files)
async def train_async(domain: Text,
config: Text,
training_files: Union[Text, List[Text]],
output: Text = DEFAULT_MODELS_PATH,
force_training: bool = False) -> Optional[Text]:
"""Trains a Rasa model (Core and NLU).
Args:
domain: Path to the domain file.
config: Path to the config for Core and NLU.
training_files: Paths to the training data for Core and NLU.
output: Output path.
force_training: If `True` retrain model even if data has not changed.
Returns:
Path of the trained model archive.
"""
train_path = tempfile.mkdtemp()
old_model = model.get_latest_model(output)
retrain_core = True
retrain_nlu = True
story_directory, nlu_data_directory = data.get_core_nlu_directories(
training_files)
new_fingerprint = model.model_fingerprint(config, domain,
nlu_data_directory,
story_directory)
if not force_training and old_model:
unpacked = model.unpack_model(old_model)
old_core, old_nlu = model.get_model_subdirectories(unpacked)
last_fingerprint = model.fingerprint_from_path(unpacked)
if not model.core_fingerprint_changed(last_fingerprint,
new_fingerprint):
target_path = os.path.join(train_path, "core")
retrain_core = not model.merge_model(old_core, target_path)
if not model.nlu_fingerprint_changed(last_fingerprint, new_fingerprint):
target_path = os.path.join(train_path, "nlu")
retrain_nlu = not model.merge_model(old_nlu, target_path)
if force_training or retrain_core:
await train_core_async(domain, config, story_directory,
output, train_path)
else:
print("Dialogue data / configuration did not change. "
"No need to retrain dialogue model.")
if force_training or retrain_nlu:
train_nlu(config, nlu_data_directory, output, train_path)
else:
print("NLU data / configuration did not change. "
"No need to retrain NLU model.")
if retrain_core or retrain_nlu:
output = create_output_path(output)
model.create_package_rasa(train_path, output, new_fingerprint)
print("Train path: '{}'.".format(train_path))
print_success("Your bot is trained and ready to take for a spin!")
return output
else:
print("Nothing changed. You can use the old model stored at {}"
"".format(os.path.abspath(old_model)))
return old_model
async def _train_async_internal(
domain: Union[Domain, Text],
config: Text,
train_path: Text,
nlu_data_directory: Text,
story_directory: Text,
output_path: Text,
force_training: bool,
fixed_model_name: Optional[Text],
kwargs: Optional[Dict],
) -> Optional[Text]:
"""Trains a Rasa model (Core and NLU). Use only from `train_async`.
Args:
domain: Path to the domain file.
config: Path to the config for Core and NLU.
train_path: Directory in which to train the model.
nlu_data_directory: Path to NLU training files.
story_directory: Path to Core training files.
output_path: Output path.
force_training: If `True` retrain model even if data has not changed.
fixed_model_name: Name of model to be stored.
kwargs: Additional training parameters.
Returns:
Path of the trained model archive.
"""
new_fingerprint = model.model_fingerprint(config, domain,
nlu_data_directory,
story_directory)
dialogue_data_not_present = not os.listdir(story_directory)
nlu_data_not_present = not os.listdir(nlu_data_directory)
if dialogue_data_not_present and nlu_data_not_present:
print_error(
"No training data given. Please provide stories and NLU data in "
"order to train a Rasa model using the '--data' argument.")
return
if dialogue_data_not_present:
print_warning(
"No dialogue data present. Just a Rasa NLU model will be trained.")
return _train_nlu_with_validated_data(
config=config,
nlu_data_directory=nlu_data_directory,
output=output_path,
fixed_model_name=fixed_model_name,
)
if nlu_data_not_present:
print_warning(
"No NLU data present. Just a Rasa Core model will be trained.")
return await _train_core_with_validated_data(
domain=domain,
config=config,
story_directory=story_directory,
output=output_path,
fixed_model_name=fixed_model_name,
kwargs=kwargs,
)
old_model = model.get_latest_model(output_path)
retrain_core, retrain_nlu = should_retrain(new_fingerprint, old_model,
train_path)
if force_training or retrain_core or retrain_nlu:
await _do_training(
domain=domain,
config=config,
output_path=output_path,
train_path=train_path,
nlu_data_directory=nlu_data_directory,
story_directory=story_directory,
force_training=force_training,
retrain_core=retrain_core,
retrain_nlu=retrain_nlu,
fixed_model_name=fixed_model_name,
kwargs=kwargs,
)
return _package_model(
new_fingerprint=new_fingerprint,
output_path=output_path,
train_path=train_path,
fixed_model_name=fixed_model_name,
)
print_success("Nothing changed. You can use the old model stored at '{}'."
"".format(os.path.abspath(old_model)))
return old_model
async def train_async(
domain: Text,
config: Text,
training_files: Union[Text, List[Text]],
output: Text = DEFAULT_MODELS_PATH,
force_training: bool = False,
kwargs: Optional[Dict] = None,
) -> Optional[Text]:
"""Trains a Rasa model (Core and NLU).
Args:
domain: Path to the domain file.
config: Path to the config for Core and NLU.
training_files: Paths to the training data for Core and NLU.
output: Output path.
force_training: If `True` retrain model even if data has not changed.
kwargs: Additional training parameters.
Returns:
Path of the trained model archive.
"""
config = get_valid_config(config, CONFIG_MANDATORY_KEYS)
train_path = tempfile.mkdtemp()
old_model = model.get_latest_model(output)
retrain_core = True
retrain_nlu = True
story_directory, nlu_data_directory = data.get_core_nlu_directories(training_files)
new_fingerprint = model.model_fingerprint(
config, domain, nlu_data_directory, story_directory
)
dialogue_data_not_present = not os.listdir(story_directory)
nlu_data_not_present = not os.listdir(nlu_data_directory)
if dialogue_data_not_present and nlu_data_not_present:
print_error(
"No training data given. Please provide dialogue and NLU data in "
"order to train a Rasa model."
)
return
if dialogue_data_not_present:
print_warning(
"No dialogue data present. Just a Rasa NLU model will be trained."
)
return train_nlu(config, nlu_data_directory, output, None)
if nlu_data_not_present:
print_warning("No NLU data present. Just a Rasa Core model will be trained.")
return await train_core_async(
domain, config, story_directory, output, None, kwargs
)
if not force_training and old_model:
unpacked = model.unpack_model(old_model)
old_core, old_nlu = model.get_model_subdirectories(unpacked)
last_fingerprint = model.fingerprint_from_path(unpacked)
if not model.core_fingerprint_changed(last_fingerprint, new_fingerprint):
target_path = os.path.join(train_path, "core")
retrain_core = not model.merge_model(old_core, target_path)
if not model.nlu_fingerprint_changed(last_fingerprint, new_fingerprint):
target_path = os.path.join(train_path, "nlu")
retrain_nlu = not model.merge_model(old_nlu, target_path)
if force_training or retrain_core:
await train_core_async(
domain, config, story_directory, output, train_path, kwargs
)
else:
print (
"Dialogue data / configuration did not change. "
"No need to retrain dialogue model."
)
if force_training or retrain_nlu:
train_nlu(config, nlu_data_directory, output, train_path)
else:
print ("NLU data / configuration did not change. No need to retrain NLU model.")
if retrain_core or retrain_nlu:
output = create_output_path(output)
model.create_package_rasa(train_path, output, new_fingerprint)
print_success("Your bot is trained and ready to take for a spin!")
return output
else:
print_success(
"Nothing changed. You can use the old model stored at '{}'"
"".format(os.path.abspath(old_model))
)
return old_model
async def train_async(
domain: Optional,
config: Text,
training_files: Optional[Union[Text, List[Text]]],
output_path: Text = DEFAULT_MODELS_PATH,
force_training: bool = False,
fixed_model_name: Optional[Text] = None,
uncompress: bool = False,
kwargs: Optional[Dict] = None,
) -> Optional[Text]:
"""Trains a Rasa model (Core and NLU).
Args:
domain: Path to the domain file.
config: Path to the config for Core and NLU.
training_files: Paths to the training data for Core and NLU.
output_path: Output path.
force_training: If `True` retrain model even if data has not changed.
fixed_model_name: Name of model to be stored.
uncompress: If `True` the model will not be compressed.
kwargs: Additional training parameters.
Returns:
Path of the trained model archive.
"""
config = get_valid_config(config, CONFIG_MANDATORY_KEYS)
train_path = tempfile.mkdtemp()
old_model = model.get_latest_model(output_path)
retrain_core = True
retrain_nlu = True
skill_imports = SkillSelector.load(config)
try:
domain = Domain.load(domain, skill_imports)
except InvalidDomain as e:
print_error(e)
return None
story_directory, nlu_data_directory = data.get_core_nlu_directories(
training_files, skill_imports)
new_fingerprint = model.model_fingerprint(config, domain,
nlu_data_directory,
story_directory)
dialogue_data_not_present = not os.listdir(story_directory)
nlu_data_not_present = not os.listdir(nlu_data_directory)
if dialogue_data_not_present and nlu_data_not_present:
print_error(
"No training data given. Please provide dialogue and NLU data in "
"order to train a Rasa model.")
return
if dialogue_data_not_present:
print_warning(
"No dialogue data present. Just a Rasa NLU model will be trained.")
return _train_nlu_with_validated_data(
config=config,
nlu_data_directory=nlu_data_directory,
output=output_path,
fixed_model_name=fixed_model_name,
uncompress=uncompress,
)
if nlu_data_not_present:
print_warning(
"No NLU data present. Just a Rasa Core model will be trained.")
return await _train_core_with_validated_data(
domain=domain,
config=config,
story_directory=story_directory,
output=output_path,
fixed_model_name=fixed_model_name,
uncompress=uncompress,
kwargs=kwargs,
)
if not force_training and old_model:
unpacked = model.unpack_model(old_model)
old_core, old_nlu = model.get_model_subdirectories(unpacked)
last_fingerprint = model.fingerprint_from_path(unpacked)
if not model.core_fingerprint_changed(last_fingerprint,
new_fingerprint):
target_path = os.path.join(train_path, "core")
retrain_core = not model.merge_model(old_core, target_path)
if not model.nlu_fingerprint_changed(last_fingerprint,
new_fingerprint):
target_path = os.path.join(train_path, "nlu")
retrain_nlu = not model.merge_model(old_nlu, target_path)
if force_training or retrain_core:
await _train_core_with_validated_data(
domain=domain,
config=config,
story_directory=story_directory,
output=output_path,
train_path=train_path,
fixed_model_name=fixed_model_name,
uncompress=uncompress,
kwargs=kwargs,
)
else:
print("Dialogue data / configuration did not change. "
"No need to retrain dialogue model.")
if force_training or retrain_nlu:
_train_nlu_with_validated_data(
config=config,
nlu_data_directory=nlu_data_directory,
output=output_path,
train_path=train_path,
fixed_model_name=fixed_model_name,
uncompress=uncompress,
)
else:
print(
"NLU data / configuration did not change. No need to retrain NLU model."
)
if retrain_core or retrain_nlu:
output_path = create_output_path(output_path,
fixed_name=fixed_model_name)
model.create_package_rasa(train_path, output_path, new_fingerprint)
if uncompress:
output_path = decompress(output_path)
print_success("Your Rasa model is trained and saved at '{}'.".format(
output_path))
return output_path
else:
print_success(
"Nothing changed. You can use the old model stored at '{}'"
"".format(os.path.abspath(old_model)))
return old_model
async def train_async(
domain: Union[Domain, Text],
config: Text,
training_files: Optional[Union[Text, List[Text]]],
output_path: Text = DEFAULT_MODELS_PATH,
force_training: bool = False,
fixed_model_name: Optional[Text] = None,
uncompress: bool = False,
kwargs: Optional[Dict] = None,
) -> Optional[Text]:
"""Trains a Rasa model (Core and NLU).
Args:
domain: Path to the domain file.
config: Path to the config for Core and NLU.
training_files: Paths to the training data for Core and NLU.
output_path: Output path.
force_training: If `True` retrain model even if data has not changed.
fixed_model_name: Name of model to be stored.
uncompress: If `True` the model will not be compressed.
kwargs: Additional training parameters.
Returns:
Path of the trained model archive.
"""
config = _get_valid_config(config, CONFIG_MANDATORY_KEYS)
train_path = tempfile.mkdtemp()
skill_imports = SkillSelector.load(config)
try:
domain = Domain.load(domain, skill_imports)
except InvalidDomain as e:
print_error(
"Could not load domain due to: '{}'. To specify a valid domain path use "
"the '--domain' argument.".format(e))
return None
story_directory, nlu_data_directory = data.get_core_nlu_directories(
training_files, skill_imports)
new_fingerprint = model.model_fingerprint(config, domain,
nlu_data_directory,
story_directory)
dialogue_data_not_present = not os.listdir(story_directory)
nlu_data_not_present = not os.listdir(nlu_data_directory)
if dialogue_data_not_present and nlu_data_not_present:
print_error(
"No training data given. Please provide stories and NLU data in "
"order to train a Rasa model using the '--data' argument.")
return
if dialogue_data_not_present:
print_warning(
"No dialogue data present. Just a Rasa NLU model will be trained.")
return _train_nlu_with_validated_data(
config=config,
nlu_data_directory=nlu_data_directory,
output=output_path,
fixed_model_name=fixed_model_name,
uncompress=uncompress,
)
if nlu_data_not_present:
print_warning(
"No NLU data present. Just a Rasa Core model will be trained.")
return await _train_core_with_validated_data(
domain=domain,
config=config,
story_directory=story_directory,
output=output_path,
fixed_model_name=fixed_model_name,
uncompress=uncompress,
kwargs=kwargs,
)
old_model = model.get_latest_model(output_path)
retrain_core, retrain_nlu = should_retrain(new_fingerprint, old_model,
train_path)
if force_training or retrain_core or retrain_nlu:
await _do_training(
domain=domain,
config=config,
output_path=output_path,
train_path=train_path,
nlu_data_directory=nlu_data_directory,
story_directory=story_directory,
force_training=force_training,
retrain_core=retrain_core,
retrain_nlu=retrain_nlu,
fixed_model_name=fixed_model_name,
uncompress=uncompress,
kwargs=kwargs,
)
return _package_model(
new_fingerprint=new_fingerprint,
output_path=output_path,
train_path=train_path,
fixed_model_name=fixed_model_name,
uncompress=uncompress,
)
print_success("Nothing changed. You can use the old model stored at '{}'."
"".format(os.path.abspath(old_model)))
return old_model
async def train_core_async(
domain: Union[Domain, Text],
config: Text,
stories: Text,
output: Text,
train_path: Optional[Text] = None,
kwargs: Optional[Dict] = None,
) -> Optional[Text]:
"""Trains a Core model.
Args:
domain: Path to the domain file.
config: Path to the config file for Core.
stories: Path to the Core training data.
output: Output path.
train_path: If `None` the model will be trained in a temporary
directory, otherwise in the provided directory.
kwargs: Additional training parameters.
Returns:
If `train_path` is given it returns the path to the model archive,
otherwise the path to the directory with the trained model files.
"""
import rasa.core.train
config = get_valid_config(config, CONFIG_MANDATORY_KEYS_CORE)
_train_path = train_path or tempfile.mkdtemp()
if isinstance(Domain, str) or not train_path:
skill_imports = SkillSelector.load(config)
domain = Domain.load(domain, skill_imports)
story_directory = data.get_core_directory(stories, skill_imports)
else:
story_directory = stories
if not os.listdir(story_directory):
print_error(
"No dialogue data given. Please provide dialogue data in order to "
"train a Rasa Core model.")
return
# normal (not compare) training
print_color("Start training dialogue model ...", color=bcolors.OKBLUE)
await rasa.core.train(
domain_file=domain,
stories_file=story_directory,
output_path=os.path.join(_train_path, "core"),
policy_config=config,
kwargs=kwargs,
)
print_color("Done.", color=bcolors.OKBLUE)
if not train_path:
# Only Core was trained.
output_path = create_output_path(output, prefix="core-")
new_fingerprint = model.model_fingerprint(config,
domain,
stories=story_directory)
model.create_package_rasa(_train_path, output_path, new_fingerprint)
print_success(
"Your Rasa Core model is trained and saved at '{}'.".format(
output_path))
return output_path
return _train_path
