[ATO-1315] Flows yaml schema

changelog/12901.improvement.md

@@ -0,0 +1 @@
+Added Schema file and schema validation for flows.

rasa/dialogue_understanding/patterns/default_flows_for_patterns.yml

@@ -125,15 +125,15 @@ flows:
       - action: validate_{{context.collect}}
         next: 
         - if: "{{context.collect}} is not null"
-          then: "done"
+          then: "END"
         - else: "ask_collect"
       - id: "ask_collect"
         action: "{{context.utter}}"
         next: "listen"
       - id: "listen"
         action: action_listen
         next: "start"
-      - id: "done"
+
 
   pattern_code_change:
     description: flow used to clean the stack after a bot update

rasa/shared/core/flows/flows_yaml_schema.json

@@ -0,0 +1,286 @@
+{
+  "type": "object",
+  "required": [
+    "flows"
+  ],
+  "properties": {
+    "version": {
+      "type": "string"
+    },
+    "flows": {
+      "type": "object",
+      "patternProperties": {
+        "^[A-Za-z_][A-Za-z0-9_]*$": {
+          "$ref": "#$defs/flow"
+        }
+      }
+    }
+  },
+  "$defs": {
+    "steps": {
+      "type": "array",
+      "minContains": 1,
+      "items": {
+        "type": "object",
+        "oneOf": [
+          {
+            "required": [
+              "action"
+            ],
+            "additionalProperties": false,
+            "properties": {
+              "id": {
+                "type": "string"
+              },
+              "action": {
+                "type": "string"
+              },
+              "next": {
+                "$ref": "#$defs/next"
+              }
+            }
+          },
+          {
+            "required": [
+              "collect"
+            ],
+            "additionalProperties": false,
+            "properties": {
+              "id": {
+                "type": "string"
+              },
+              "description":{
+                "type": "string"
+              },
+              "collect": {
+                "type": "string"
+              },
+              "ask_before_filling": {
+                "type": "boolean"
+              },
+              "reset_after_flow_ends": {
+                "type": "boolean"
+              },
+              "utter": {
+                "type": "string"
+              },
+              "rejections": {
+                "type": "array",
+                "items": {
+                  "type": "object",
+                  "required": [
+                    "if",
+                    "utter"
+                  ],
+                  "properties": {
+                    "if": {
+                      "type": "string"
+                    },
+                    "utter": {
+                      "type": "string"
+                    }
+                  }
+                }
+              },
+              "next": {
+                "$ref": "#$defs/next"
+              }
+            }
+          },
+          {
+            "required": [
+              "link"
+            ],
+            "additionalProperties": false,
+            "properties": {
+              "id": {
+                "type": "string"
+              },
+              "link": {
+                "type": "string"
+              },
+              "next": {
+                "type": "null"
+              }
+            }
+          },
+          {
+            "required": [
+              "set_slots"
+            ],
+            "additionalProperties": false,
+            "properties": {
+              "id": {
+                "type": "string"
+              },
+              "set_slots": {
+                "$ref": "#$defs/set_slots"
+              },
+              "next": {
+                "$ref": "#$defs/next"
+              }
+            }
+          },
+          {
+            "required": [
+              "next"
+            ],
+            "additionalProperties": false,
+            "properties": {
+              "next": {
+                "$ref": "#$defs/next"
+              },
+              "id": {
+                "type": "string"
+              }
+            }
+          },
+          {
+            "required": [
+              "generation_prompt"
+            ],
+            "additionalProperties": false,
+            "properties": {
+              "generation_prompt": {
+                "type": "string"
+              },
+              "id": {
+                "type": "string"
+              },
+              "next": {
+                "$ref": "#$defs/next"
+              }
+            }
+          },
+          {
+            "required": [
+              "entry_prompt"
+            ],
+            "additionalProperties": false,
+            "properties": {
+              "id": {
+                "type": "string"
+              },
+              "entry_prompt": {
+                "type": "string"
+              },
+              "next": {
+                "$ref": "#$defs/next"
+              }
+            }
+          }
+        ]
+      }
+    },
+    "flow": {
+      "required": [
+        "steps"
+      ],
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "description": {
+          "type": "string"
+        },
+        "if": {
+          "type": "string"
+        },
+        "name": {
+          "type": "string"
+        },
+        "nlu_trigger": {
+          "type": "array",
+          "items": {
+            "required": [
+              "intent"
+            ],
+            "type": "object",
+            "additionalProperties": false, 
+            "properties": {
+              "intent": {
+                "type": "object",
+                "properties": {
+                  "confidence_threshold": {
+                    "type": "number"
+                  },
+                  "name": {
+                    "type": "string"
+                  }
+                },
+                "required": [
+                  "name"
+                ]
+              }
+            }
+          }
+        },
+        "steps": {
+          "$ref": "#$defs/steps"
+        }
+      }
+    },
+    "next": {
+      "anyOf": [
+        {
+          "type": "array",
+          "minContains": 1,
+          "items": {
+            "type": "object",
+            "oneOf": [
+              {
+                "required": [
+                  "if",
+                  "then"
+                ]
+              },
+              {
+                "required": [
+                  "else"
+                ]
+              }
+            ],
+            "properties": {
+              "else": {
+                "oneOf": [
+                  {
+                    "type": "string"
+                  },
+                  {
+                    "$ref": "#$defs/steps"
+                  }
+                ]
+              },
+              "if": {
+                "type": "string"
+              },
+              "then": {
+                "oneOf": [
+                  {
+                    "$ref": "#$defs/steps"
+                  },
+                  {
+                    "type": "string"
+                  }
+                ]
+              }
+            }
+          }
+        },
+        {
+          "type": "string"
+        }
+      ]
+    },
+    "set_slots": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "patternProperties": {
+          "^[A-Za-z_][A-Za-z0-9_]*$": {
+            "type": ["string", "null", "boolean", "number"]
+          }
+        }
+      }
+    }
+  }
+}

rasa/shared/core/flows/yaml_flows_io.py

@@ -1,6 +1,7 @@
 import textwrap
 from pathlib import Path
 from typing import List, Text, Union
+
 from rasa.shared.core.flows.utils import KEY_FLOWS
 
 import rasa.shared.utils.io
@@ -9,7 +10,7 @@
 
 from rasa.shared.core.flows.flow import Flow, FlowsList
 
-FLOWS_SCHEMA_FILE = "/shared/core/flows/flows_yaml_schema.yml"
+FLOWS_SCHEMA_FILE = "shared/core/flows/flows_yaml_schema.json"
 
 
 class YAMLFlowsReader:
@@ -53,7 +54,9 @@ def read_from_string(cls, string: Text, skip_validation: bool = False) -> FlowsL
             `Flow`s read from `string`.
         """
         if not skip_validation:
-            rasa.shared.utils.validation.validate_yaml_schema(string, FLOWS_SCHEMA_FILE)
+            rasa.shared.utils.validation.validate_yaml_with_jsonschema(
+                string, FLOWS_SCHEMA_FILE
+            )
 
         yaml_content = rasa.shared.utils.io.read_yaml(string)
 

rasa/shared/utils/validation.py

@@ -289,3 +289,44 @@ def validate_training_data_format_version(
         docs=DOCS_URL_TRAINING_DATA,
     )
     return False
+
+
+def validate_yaml_with_jsonschema(
+    yaml_file_content: Text, schema_path: Text, package_name: Text = PACKAGE_NAME
+) -> None:
+    """Validate data format.
+
+    Args:
+        yaml_file_content: the content of the yaml file to be validated
+        schema_path: the schema of the yaml file
+        package_name: the name of the package the schema is located in. defaults
+            to `rasa`.
+
+    Raises:
+        YamlSyntaxException: if the yaml file is not valid.
+        SchemaValidationError: if validation fails.
+    """
+    from jsonschema import validate, ValidationError
+    from ruamel.yaml import YAMLError
+    import pkg_resources
+
+    schema_file = pkg_resources.resource_filename(package_name, schema_path)
+    schema_content = rasa.shared.utils.io.read_json_file(schema_file)
+
+    try:
+        # we need "rt" since
+        # it will add meta information to the parsed output. this meta information
+        # will include e.g. at which line an object was parsed. this is very
+        # helpful when we validate files later on and want to point the user to the
+        # right line
+        source_data = rasa.shared.utils.io.read_yaml(
+            yaml_file_content, reader_type=["safe", "rt"]
+        )
+    except (YAMLError, DuplicateKeyError) as e:
+        raise YamlSyntaxException(underlying_yaml_exception=e)
+
+    try:
+        validate(source_data, schema_content)
+    except ValidationError as error:
+        error.message += ". Failed to validate data, make sure your data is valid."
+        raise SchemaValidationError.create_from(error) from error