fix: fix JSON to YAML conversion

changelog/10675.bugfix.md

@@ -0,0 +1 @@
+Fix broken conversion from Rasa JSON NLU data to Rasa YAML NLU data.

rasa/cli/data.py

@@ -224,7 +224,7 @@ def _validate_story_structure(validator: "Validator", args: argparse.Namespace)
 def _convert_nlu_data(args: argparse.Namespace) -> None:
     import rasa.nlu.convert
 
-    if args.format == "json":
+    if args.format in ["json", "yaml"]:
         rasa.nlu.convert.convert_training_data(
             args.data, args.out, args.format, args.language
         )

rasa/nlu/convert.py

@@ -1,7 +1,7 @@
-import argparse
 import os
 from typing import Text, Union
 
+from rasa.shared.nlu.training_data.formats.rasa_yaml import RasaYAMLWriter
 from rasa.shared.utils.cli import print_error
 import rasa.shared.nlu.training_data.loading
 from rasa.nlu.utils import write_to_file
@@ -31,18 +31,10 @@ def convert_training_data(
         )
         return
 
+    td = rasa.shared.nlu.training_data.loading.load_data(data_file, language)
     if output_format == "json":
-        td = rasa.shared.nlu.training_data.loading.load_data(data_file, language)
         output = td.nlu_as_json(indent=2)
     else:
-        print_error(
-            "Did not recognize output format. Supported output formats: 'json' and "
-            "'md'. Specify the desired output format with '--format'."
-        )
-        return
+        output = RasaYAMLWriter().dumps(td)
 
     write_to_file(out_file, output)
-
-
-def main(args: argparse.Namespace) -> None:
-    convert_training_data(args.data, args.out, args.format, args.language)

tests/cli/test_rasa_data.py

@@ -6,10 +6,12 @@
 from collections import namedtuple
 from typing import Callable, Text
 
+from _pytest.fixtures import FixtureRequest
 from _pytest.monkeypatch import MonkeyPatch
 from _pytest.pytester import RunResult
 from rasa.cli import data
 from rasa.shared.importers.importer import TrainingDataImporter
+from rasa.shared.nlu.training_data.formats import RasaYAMLReader
 from rasa.validator import Validator
 import rasa.shared.utils.io
 
@@ -53,7 +55,7 @@ def test_data_split_nlu(run_in_simple_project: Callable[..., RunResult]):
         assert yml_file.exists(), f"{yml_file} file does not exist"
 
 
-def test_data_convert_nlu(run_in_simple_project: Callable[..., RunResult]):
+def test_data_convert_nlu_json(run_in_simple_project: Callable[..., RunResult]):
     run_in_simple_project(
         "data",
         "convert",
@@ -69,6 +71,41 @@ def test_data_convert_nlu(run_in_simple_project: Callable[..., RunResult]):
     assert os.path.exists("out_nlu_data.json")
 
 
+def test_data_convert_nlu_yml(
+    run: Callable[..., RunResult], tmp_path: Path, request: FixtureRequest
+):
+
+    target_file = tmp_path / "out.yml"
+
+    # The request rootdir is required as the `testdir` fixture in `run` changes the
+    # working directory
+    test_data_dir = Path(request.config.rootdir, "data", "examples", "rasa")
+    source_file = (test_data_dir / "demo-rasa.json").absolute()
+    result = run(
+        "data",
+        "convert",
+        "nlu",
+        "--data",
+        str(source_file),
+        "--out",
+        str(target_file),
+        "-f",
+        "yaml",
+    )
+
+    assert result.ret == 0
+    assert target_file.exists()
+
+    actual_data = RasaYAMLReader().read(target_file)
+    expected = RasaYAMLReader().read(test_data_dir / "demo-rasa.yml")
+
+    assert len(actual_data.training_examples) == len(expected.training_examples)
+    assert len(actual_data.entity_synonyms) == len(expected.entity_synonyms)
+    assert len(actual_data.regex_features) == len(expected.regex_features)
+    assert len(actual_data.lookup_tables) == len(expected.lookup_tables)
+    assert actual_data.entities == expected.entities
+
+
 def test_data_split_help(run: Callable[..., RunResult]):
     output = run("data", "split", "nlu", "--help")