Initial implementation of converter for training data files

changelog/6404.feature.md

@@ -0,0 +1 @@
+User can use ``rasa data convert {nlu|core} -f yaml`` command to convert training data from Markdown format to YAML format.

rasa/cli/arguments/convert.py

@@ -0,0 +1,22 @@
+import argparse
+
+from rasa.cli.arguments.default_arguments import add_model_param, add_domain_param
+
+
+def set_convert_arguments(parser: argparse.ArgumentParser):
+    parser.add_argument(
+        "--training_data",
+        required=True,
+        help="Paths to the source NLU and Core data files in a Markdown format",
+    )
+
+    parser.add_argument(
+        "--output",
+        type=str,
+        required=True,
+        help=(
+            "Path to the output directory where all the converted training data files "
+            "will be written to. Converted files will have the same name as the "
+            "original ones with a '_converted' suffix."
+        ),
+    )

rasa/cli/arguments/data.py

@@ -1,4 +1,5 @@
 import argparse
+from typing import Text
 
 from rasa.cli.arguments.default_arguments import (
     add_nlu_data_param,
@@ -8,8 +9,8 @@
 )
 
 
-def set_convert_arguments(parser: argparse.ArgumentParser):
-    add_data_param(parser, required=True, default=None, data_type="Rasa NLU ")
+def set_convert_arguments(parser: argparse.ArgumentParser, data_type: Text):
+    add_data_param(parser, required=True, default=None, data_type=data_type)
 
     add_out_param(
         parser,
@@ -24,8 +25,10 @@ def set_convert_arguments(parser: argparse.ArgumentParser):
         "-f",
         "--format",
         required=True,
-        choices=["json", "md"],
-        help="Output format the training data should be converted into.",
+        choices=["json", "md", "yaml"],
+        help="Output format the training data should be converted into. "
+        "Note: currently training data can be converted to 'yaml' format "
+        "only from 'markdown' format",
     )
 
 

rasa/cli/data.py

@@ -1,16 +1,31 @@
 import logging
 import argparse
 import asyncio
+import os
+from pathlib import Path
 from typing import List
 
 from rasa import data
 from rasa.cli.arguments import data as arguments
 import rasa.cli.utils
 from rasa.constants import DEFAULT_DATA_PATH
+from rasa.core.interpreter import RegexInterpreter
+from rasa.core.training.story_reader.markdown_story_reader import MarkdownStoryReader
+from rasa.core.training.story_writer.yaml_story_writer import YAMLStoryWriter
+from rasa.nlu.convert import convert_training_data
+from rasa.nlu.training_data.formats import MarkdownReader
+from rasa.nlu.training_data.formats.rasa_yaml import RasaYAMLWriter
 from rasa.validator import Validator
 from rasa.importers.rasa import RasaFileImporter
+from rasa.cli.utils import (
+    print_success,
+    print_error_and_exit,
+    print_info,
+    print_warning,
+)
 
 logger = logging.getLogger(__name__)
+CONVERTED_FILE_SUFFIX = "_converted.yml"
 
 
 # noinspection PyProtectedMember
@@ -51,11 +66,21 @@ def _add_data_convert_parsers(
         "nlu",
         formatter_class=argparse.ArgumentDefaultsHelpFormatter,
         parents=parents,
-        help="Converts NLU data between Markdown and json formats.",
+        help="Converts NLU data between formats.",
     )
-    convert_nlu_parser.set_defaults(func=convert.main)
+    convert_nlu_parser.set_defaults(func=_convert_nlu_data)
 
-    arguments.set_convert_arguments(convert_nlu_parser)
+    arguments.set_convert_arguments(convert_nlu_parser, data_type="Rasa NLU")
+
+    convert_core_parser = convert_subparsers.add_parser(
+        "core",
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+        parents=parents,
+        help="Converts Core data between formats.",
+    )
+    convert_core_parser.set_defaults(func=_convert_core_data)
+
+    arguments.set_convert_arguments(convert_core_parser, data_type="Rasa Core")
 
 
 def _add_data_split_parsers(
@@ -182,3 +207,86 @@ def _validate_story_structure(validator: Validator, args: argparse.Namespace) ->
     return validator.verify_story_structure(
         not args.fail_on_warnings, max_history=args.max_history
     )
+
+
+def _convert_nlu_data(args: argparse.Namespace):
+    if args.format in ["json", "md"]:
+        convert_training_data(args.data, args.out, args.format, args.language)
+    elif args.format == "yaml":
+        convert_to_yaml(args, True)
+    else:
+        print_error_and_exit(
+            "Did not recognize output format. Supported output formats: 'json', "
+            "'md', 'yaml'. Specify the desired output format with '--format'."
+        )
+
+
+def _convert_core_data(args: argparse.Namespace):
+    if args.format == "yaml":
+        convert_to_yaml(args, False)
+    else:
+        print_error_and_exit(
+            "Did not recognize output format. Supported output formats: "
+            "'yaml'. Specify the desired output format with '--format'."
+        )
+
+
+def convert_to_yaml(args: argparse.Namespace, is_nlu: bool) -> None:
+
+    output = Path(args.out)
+    if not os.path.exists(output):
+        print_error_and_exit(
+            f"The output path '{output}' doesn't exist. Please make sure to specify "
+            f"an existing directory and try again."
+        )
+
+    training_data = Path(args.data)
+    if not os.path.exists(training_data):
+        print_error_and_exit(
+            f"The training data path {training_data} doesn't exist "
+            f"and will be skipped."
+        )
+
+    num_of_files_converted = 0
+    for file in os.listdir(training_data):
+        source_path = training_data / file
+        output_path = output / f"{source_path.stem}{CONVERTED_FILE_SUFFIX}"
+
+        if MarkdownReader.is_markdown_nlu_file(source_path):
+            if not is_nlu:
+                continue
+            _write_nlu_yaml(source_path, output_path, source_path)
+            num_of_files_converted += 1
+        elif not is_nlu and MarkdownStoryReader.is_markdown_story_file(source_path):
+            _write_core_yaml(source_path, output_path, source_path)
+            num_of_files_converted += 1
+        else:
+            print_warning(f"Skipped file '{source_path}'")
+
+    print_info(f"Converted {num_of_files_converted} file(s), saved in '{output}'.")
+
+
+def _write_nlu_yaml(
+    training_data_path: Path, output_path: Path, source_path: Path
+) -> None:
+    reader = MarkdownReader()
+    writer = RasaYAMLWriter()
+
+    training_data = reader.read(training_data_path)
+    writer.dump(output_path, training_data)
+
+    print_success(f"Converted NLU file: '{source_path}' >> '{output_path}'.")
+
+
+def _write_core_yaml(
+    training_data_path: Path, output_path: Path, source_path: Path
+) -> None:
+    reader = MarkdownStoryReader(RegexInterpreter())
+    writer = YAMLStoryWriter()
+
+    loop = asyncio.get_event_loop()
+    steps = loop.run_until_complete(reader.read_from_file(training_data_path))
+
+    writer.dump(output_path, steps)
+
+    print_success(f"Converted Core file: '{source_path}' >> '{output_path}'.")

rasa/core/training/story_reader/markdown_story_reader.py

@@ -3,8 +3,8 @@
 import logging
 import os
 import re
-from pathlib import PurePath
-from typing import Dict, Text, List, Any
+from pathlib import PurePath, Path
+from typing import Dict, Text, List, Any, Union
 
 import rasa.utils.io as io_utils
 from rasa.constants import DOCS_URL_DOMAINS, DOCS_URL_STORIES
@@ -27,7 +27,7 @@ class MarkdownStoryReader(StoryReader):
 
     """
 
-    async def read_from_file(self, filename: Text) -> List[StoryStep]:
+    async def read_from_file(self, filename: Union[Text, Path]) -> List[StoryStep]:
         """Given a md file reads the contained stories."""
 
         try:

rasa/nlu/training_data/formats/markdown.py

@@ -3,8 +3,9 @@
 import typing
 from collections import OrderedDict
 from json import JSONDecodeError
-from typing import Any, Text, Optional, Tuple, Dict, Match
+from typing import Any, Text, Optional, Tuple, Dict
 
+import rasa.utils.io as io_utils
 from rasa.constants import DOCS_URL_TRAINING_DATA_NLU
 from rasa.nlu.training_data.formats.readerwriter import (
     TrainingDataReader,
@@ -201,6 +202,11 @@ def _set_current_section(self, section: Text, title: Text) -> None:
         self.current_section = section
         self.current_title = title
 
+    @staticmethod
+    def is_markdown_nlu_file(filename: Text) -> bool:
+        content = io_utils.read_file(filename)
+        return any(marker in content for marker in MARKDOWN_SECTION_MARKERS)
+
 
 class MarkdownWriter(TrainingDataWriter):
     def dumps(self, training_data: "TrainingData") -> Text:

rasa/nlu/training_data/formats/readerwriter.py

@@ -1,5 +1,6 @@
 import json
 from collections import OrderedDict
+from pathlib import Path
 
 from rasa.core.constants import INTENT_MESSAGE_PREFIX
 
@@ -16,7 +17,7 @@
 import rasa.utils.io as io_utils
 import typing
 from rasa.nlu import utils
-from typing import Text, Dict, Any
+from typing import Text, Dict, Any, Union
 
 if typing.TYPE_CHECKING:
     from rasa.nlu.training_data import TrainingData
@@ -26,7 +27,7 @@ class TrainingDataReader:
     def __init__(self):
         self.filename: Text = ""
 
-    def read(self, filename: Text, **kwargs: Any) -> "TrainingData":
+    def read(self, filename: Union[Text, Path], **kwargs: Any) -> "TrainingData":
         """Reads TrainingData from a file."""
         self.filename = filename
         return self.reads(io_utils.read_file(filename), **kwargs)

tests/cli/test_rasa_data.py

@@ -57,7 +57,7 @@ def test_data_convert_help(run: Callable[..., RunResult]):
     output = run("data", "convert", "nlu", "--help")
 
     help_text = """usage: rasa data convert nlu [-h] [-v] [-vv] [--quiet] --data DATA --out OUT
-                             [-l LANGUAGE] -f {json,md}"""
+                             [-l LANGUAGE] -f {json,md,yaml}"""
 
     lines = help_text.split("\n")
 
@@ -112,3 +112,53 @@ def test_validate_files_exit_early():
 
     assert pytest_e.type == SystemExit
     assert pytest_e.value.code == 1
+
+
+def test_rasa_data_convert_to_yaml(
+    run_in_simple_project: Callable[..., RunResult], run: Callable[..., RunResult]
+):
+    converted_data_folder = "converted_data"
+    os.mkdir(converted_data_folder)
+
+    simple_nlu_md = """
+    ## intent:greet
+    - hey
+    - hello
+    """
+
+    with open("data/nlu.md", "w") as f:
+        f.write(simple_nlu_md)
+
+    simple_story_md = """
+    ## happy path
+    * greet
+        - utter_greet
+    """
+
+    with open("data/stories.md", "w") as f:
+        f.write(simple_story_md)
+
+    run_in_simple_project(
+        "data",
+        "convert",
+        "nlu",
+        "-f",
+        "yaml",
+        "--data",
+        "data",
+        "--out",
+        converted_data_folder,
+    )
+    run_in_simple_project(
+        "data",
+        "convert",
+        "core",
+        "-f",
+        "yaml",
+        "--data",
+        "data",
+        "--out",
+        converted_data_folder,
+    )
+
+    assert len(os.listdir(converted_data_folder)) == 2