datalex-cli 0.1.1__py3-none-any.whl

datalex_core/_schemas/datalex/entity.schema.json

@@ -0,0 +1,158 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://duckcode.dev/schemas/datalex/entity.schema.json",
+  "title": "DataLex entity",
+  "type": "object",
+  "required": ["kind", "name", "columns"],
+  "additionalProperties": false,
+  "properties": {
+    "kind": { "const": "entity" },
+    "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "previous_name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "layer": {
+      "type": "string",
+      "enum": ["conceptual", "logical", "physical"],
+      "description": "Required. Conceptual: business view. Logical: attributes/types. Physical: DBMS-specific DDL target."
+    },
+    "logical": {
+      "type": "string",
+      "description": "At physical layer: back-reference to the logical entity name this physical implements.",
+      "pattern": "^[a-z][a-z0-9_]*$"
+    },
+    "dialect": {
+      "type": "string",
+      "pattern": "^[a-z][a-z0-9_]*$",
+      "description": "Required at physical layer. Names the dialect plugin that owns emission."
+    },
+    "database": { "type": "string" },
+    "schema": { "type": "string" },
+    "physical_name": { "type": "string" },
+    "description": { "type": "string" },
+    "owner": { "type": "string" },
+    "domain": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "subject_area": { "type": "string" },
+    "tags": {
+      "type": "array",
+      "items": { "type": "string", "pattern": "^[a-z][a-z0-9-]*$" }
+    },
+    "terms": {
+      "type": "array",
+      "items": { "type": "string", "pattern": "^(term:)?[a-z][a-z0-9_]*$" }
+    },
+    "raw_ddl": { "type": "string" },
+    "columns": {
+      "type": "array",
+      "minItems": 1,
+      "items": { "$ref": "#/$defs/column" }
+    },
+    "indexes": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "required": ["name", "columns"],
+        "additionalProperties": false,
+        "properties": {
+          "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+          "previous_name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+          "columns": {
+            "type": "array",
+            "minItems": 1,
+            "items": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" }
+          },
+          "unique": { "type": "boolean" },
+          "type": { "type": "string", "enum": ["btree", "hash", "gin", "gist", "brin"] }
+        }
+      }
+    },
+    "partition_by": {
+      "type": "array",
+      "items": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" }
+    },
+    "cluster_by": {
+      "type": "array",
+      "items": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" }
+    },
+    "meta": {
+      "type": "object",
+      "additionalProperties": true
+    }
+  },
+  "$defs": {
+    "column": {
+      "type": "object",
+      "required": ["name", "type"],
+      "additionalProperties": false,
+      "properties": {
+        "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+        "previous_name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+        "type": { "type": "string", "minLength": 1 },
+        "physical": {
+          "type": "object",
+          "additionalProperties": {
+            "type": "object",
+            "additionalProperties": false,
+            "properties": {
+              "type": { "type": "string" },
+              "raw_ddl": { "type": "string" }
+            }
+          }
+        },
+        "logical": { "type": "string" },
+        "description": { "type": "string" },
+        "nullable": { "type": "boolean" },
+        "primary_key": { "type": "boolean" },
+        "unique": { "type": "boolean" },
+        "default": { "type": ["string", "number", "boolean", "null"] },
+        "references": {
+          "type": "object",
+          "required": ["entity", "column"],
+          "additionalProperties": false,
+          "properties": {
+            "entity": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+            "column": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+            "on_delete": { "type": "string", "enum": ["restrict", "cascade", "set_null", "no_action"] },
+            "on_update": { "type": "string", "enum": ["restrict", "cascade", "set_null", "no_action"] },
+            "relationship": { "type": "string", "enum": ["many_to_one", "one_to_one", "many_to_many"] }
+          }
+        },
+        "constraints": {
+          "type": "array",
+          "items": {
+            "type": "object",
+            "required": ["type"],
+            "additionalProperties": true,
+            "properties": {
+              "type": { "type": "string", "enum": ["check", "not_null", "unique", "primary_key", "foreign_key"] },
+              "expression": { "type": "string" }
+            }
+          }
+        },
+        "sensitivity": { "type": "string" },
+        "terms": {
+          "type": "array",
+          "items": { "type": "string", "pattern": "^(term:)?[a-z][a-z0-9_]*$" }
+        },
+        "tags": {
+          "type": "array",
+          "items": { "type": "string", "pattern": "^[a-z][a-z0-9-]*$" }
+        },
+        "use": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+        "examples": {
+          "type": "array",
+          "items": { "type": ["string", "number", "boolean", "null"] }
+        },
+        "deprecated": { "type": "boolean" },
+        "meta": {
+          "type": "object",
+          "additionalProperties": true
+        }
+      }
+    }
+  },
+  "allOf": [
+    {
+      "if": { "properties": { "layer": { "const": "physical" } }, "required": ["layer"] },
+      "then": { "required": ["dialect"] }
+    }
+  ]
+}

datalex_core/_schemas/datalex/model.schema.json

@@ -0,0 +1,141 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://duckcode.dev/schemas/datalex/model.schema.json",
+  "title": "DataLex model (dbt-style transformation / derived table)",
+  "type": "object",
+  "required": ["kind", "name"],
+  "additionalProperties": false,
+  "properties": {
+    "kind": { "const": "model" },
+    "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "previous_name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "description": { "type": "string" },
+    "materialization": {
+      "type": "string",
+      "enum": ["view", "table", "incremental", "ephemeral", "materialized_view", "snapshot"]
+    },
+    "database": { "type": "string" },
+    "schema": { "type": "string" },
+    "tags": {
+      "type": "array",
+      "items": { "type": "string", "pattern": "^[a-z][a-z0-9-]*$" }
+    },
+    "owner": { "type": "string" },
+    "domain": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "derived_sql": {
+      "type": "string",
+      "description": "Inline SQL for the transformation. Used for column-level lineage extraction."
+    },
+    "sql_path": {
+      "type": "string",
+      "description": "Path to a .sql file (relative to the YAML file) containing the transformation."
+    },
+    "depends_on": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "maxProperties": 1,
+        "minProperties": 1,
+        "additionalProperties": false,
+        "properties": {
+          "ref": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+          "source": {
+            "type": "object",
+            "required": ["source", "name"],
+            "additionalProperties": false,
+            "properties": {
+              "source": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+              "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" }
+            }
+          }
+        }
+      }
+    },
+    "contract": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "enforced": { "type": "boolean" }
+      }
+    },
+    "columns": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "required": ["name"],
+        "additionalProperties": false,
+        "properties": {
+          "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+          "previous_name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+          "type": { "type": "string" },
+          "description": { "type": "string" },
+          "tests": {
+            "type": "array",
+            "items": {
+              "oneOf": [
+                { "type": "string", "enum": ["unique", "not_null"] },
+                {
+                  "type": "object",
+                  "maxProperties": 1,
+                  "minProperties": 1,
+                  "additionalProperties": true
+                }
+              ]
+            }
+          },
+          "constraints": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "required": ["type"],
+              "additionalProperties": true,
+              "properties": {
+                "type": { "type": "string", "enum": ["primary_key", "not_null", "unique", "foreign_key", "check"] }
+              }
+            }
+          },
+          "contract": {
+            "type": "object",
+            "additionalProperties": false,
+            "properties": { "enforced": { "type": "boolean" } }
+          },
+          "sensitivity": { "type": "string" },
+          "terms": {
+            "type": "array",
+            "items": { "type": "string", "pattern": "^(term:)?[a-z][a-z0-9_]*$" }
+          },
+          "tags": {
+            "type": "array",
+            "items": { "type": "string", "pattern": "^[a-z][a-z0-9-]*$" }
+          },
+          "meta": { "type": "object", "additionalProperties": true }
+        }
+      }
+    },
+    "freshness": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "warn_after": {
+          "type": "object",
+          "required": ["count", "period"],
+          "additionalProperties": false,
+          "properties": {
+            "count": { "type": "integer", "minimum": 1 },
+            "period": { "type": "string", "enum": ["minute", "hour", "day"] }
+          }
+        },
+        "error_after": {
+          "type": "object",
+          "required": ["count", "period"],
+          "additionalProperties": false,
+          "properties": {
+            "count": { "type": "integer", "minimum": 1 },
+            "period": { "type": "string", "enum": ["minute", "hour", "day"] }
+          }
+        }
+      }
+    },
+    "meta": { "type": "object", "additionalProperties": true }
+  }
+}

datalex_core/_schemas/datalex/policy.schema.json

@@ -0,0 +1,70 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://duckcode.dev/schemas/datalex/policy.schema.json",
+  "title": "DataLex policy (access / masking / governance rule)",
+  "type": "object",
+  "required": ["kind", "name", "rule"],
+  "additionalProperties": false,
+  "properties": {
+    "kind": { "const": "policy" },
+    "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "description": { "type": "string" },
+    "rule": {
+      "type": "string",
+      "enum": [
+        "require_description",
+        "require_owner",
+        "require_primary_key",
+        "require_indexes",
+        "require_sla",
+        "naming_convention",
+        "deprecation_check",
+        "column_sensitivity_required",
+        "contract_required",
+        "custom_expression"
+      ]
+    },
+    "scope": {
+      "type": "string",
+      "enum": ["entity", "column", "source", "model", "project"],
+      "default": "entity"
+    },
+    "severity": {
+      "type": "string",
+      "enum": ["info", "warn", "error"],
+      "default": "warn"
+    },
+    "applies_to": {
+      "type": "object",
+      "description": "Filter which objects the policy applies to.",
+      "additionalProperties": true
+    },
+    "expression": { "type": "string" },
+    "grants": {
+      "type": "array",
+      "description": "For access policies: principals and the privileges they receive.",
+      "items": {
+        "type": "object",
+        "required": ["principal", "privileges"],
+        "additionalProperties": false,
+        "properties": {
+          "principal": { "type": "string" },
+          "privileges": {
+            "type": "array",
+            "minItems": 1,
+            "items": { "type": "string", "enum": ["SELECT", "INSERT", "UPDATE", "DELETE", "TRUNCATE", "REFERENCES", "TRIGGER", "USAGE", "ALL"] }
+          },
+          "on_entities": {
+            "type": "array",
+            "items": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" }
+          }
+        }
+      }
+    },
+    "extends": {
+      "type": "string",
+      "description": "Policy-pack inheritance."
+    },
+    "meta": { "type": "object", "additionalProperties": true }
+  }
+}

datalex_core/_schemas/datalex/project.schema.json

@@ -0,0 +1,82 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://duckcode.dev/schemas/datalex/project.schema.json",
+  "title": "DataLex project manifest",
+  "type": "object",
+  "required": ["kind", "name", "version"],
+  "additionalProperties": false,
+  "properties": {
+    "kind": { "const": "project" },
+    "name": { "type": "string", "minLength": 1 },
+    "version": { "type": ["string", "integer"] },
+    "description": { "type": "string" },
+    "owner": { "type": "string" },
+    "dialects": {
+      "type": "array",
+      "items": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+      "description": "Enabled dialect plugins. Must correspond to modules under datalex_core/dialects/."
+    },
+    "default_dialect": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "glossary": {
+      "type": "string",
+      "description": "Glob for glossary files, e.g. 'glossary/**/*.yaml'."
+    },
+    "models": {
+      "type": "string",
+      "description": "Glob for entity/source/model files, e.g. 'models/**/*.yaml'."
+    },
+    "sources": {
+      "type": "string",
+      "description": "Glob for kind:source files, e.g. 'sources/**/*.yaml'. Optional."
+    },
+    "snippets": {
+      "type": "string",
+      "description": "Glob for snippet files, default '.datalex/snippets/**/*.yaml'."
+    },
+    "policies": {
+      "type": "string",
+      "description": "Glob for policy files."
+    },
+    "imports": {
+      "type": "array",
+      "description": "Cross-repo package imports. Resolver fetches packages into ~/.datalex/packages/ and pins them in .datalex/lock.yaml.",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["package"],
+        "properties": {
+          "package": {
+            "type": "string",
+            "description": "Package identifier. Forms: 'org/name', 'org/name@version', or any Git-URL-like string when paired with git:."
+          },
+          "alias": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+          "path": {
+            "type": "string",
+            "description": "Local directory (relative to project root or absolute). Mutually exclusive with git:."
+          },
+          "git": {
+            "type": "string",
+            "description": "Git URL (https or ssh) to clone for resolution."
+          },
+          "ref": {
+            "type": "string",
+            "description": "Git ref — tag, branch, or commit SHA."
+          },
+          "version": {
+            "type": "string",
+            "description": "Version override (falls back to parsing 'org/name@version' from package:)."
+          },
+          "entities": {
+            "type": "array",
+            "items": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+            "description": "Optional allowlist of entity names to import (default: all)."
+          }
+        }
+      }
+    },
+    "meta": {
+      "type": "object",
+      "additionalProperties": true
+    }
+  }
+}

datalex_core/_schemas/datalex/snippet.schema.json

@@ -0,0 +1,24 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://duckcode.dev/schemas/datalex/snippet.schema.json",
+  "title": "DataLex snippet (reusable YAML fragment)",
+  "type": "object",
+  "required": ["kind", "name", "apply"],
+  "additionalProperties": false,
+  "properties": {
+    "kind": { "const": "snippet" },
+    "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "description": { "type": "string" },
+    "targets": {
+      "type": "array",
+      "description": "Which object kinds this snippet is valid on. Default: column.",
+      "items": { "type": "string", "enum": ["entity", "column", "source_column", "model_column"] }
+    },
+    "apply": {
+      "type": "object",
+      "description": "Keys merged into the target. Target keys win on conflict unless override=true is set on a key.",
+      "additionalProperties": true
+    },
+    "meta": { "type": "object", "additionalProperties": true }
+  }
+}

datalex_core/_schemas/datalex/source.schema.json

@@ -0,0 +1,104 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://duckcode.dev/schemas/datalex/source.schema.json",
+  "title": "DataLex source (dbt-style external data source)",
+  "type": "object",
+  "required": ["kind", "name", "tables"],
+  "additionalProperties": false,
+  "properties": {
+    "kind": { "const": "source" },
+    "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "previous_name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "description": { "type": "string" },
+    "database": { "type": "string" },
+    "schema": { "type": "string" },
+    "loader": { "type": "string" },
+    "loaded_at_field": { "type": "string" },
+    "freshness": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "warn_after": { "$ref": "#/$defs/duration" },
+        "error_after": { "$ref": "#/$defs/duration" },
+        "filter": { "type": "string" }
+      }
+    },
+    "owner": { "type": "string" },
+    "tags": {
+      "type": "array",
+      "items": { "type": "string", "pattern": "^[a-z][a-z0-9-]*$" }
+    },
+    "tables": {
+      "type": "array",
+      "minItems": 1,
+      "items": {
+        "type": "object",
+        "required": ["name"],
+        "additionalProperties": false,
+        "properties": {
+          "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+          "previous_name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+          "description": { "type": "string" },
+          "identifier": { "type": "string" },
+          "loaded_at_field": { "type": "string" },
+          "freshness": {
+            "type": "object",
+            "additionalProperties": false,
+            "properties": {
+              "warn_after": { "$ref": "#/$defs/duration" },
+              "error_after": { "$ref": "#/$defs/duration" },
+              "filter": { "type": "string" }
+            }
+          },
+          "columns": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "required": ["name"],
+              "additionalProperties": false,
+              "properties": {
+                "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+                "type": { "type": "string" },
+                "description": { "type": "string" },
+                "tests": { "$ref": "#/$defs/tests" },
+                "tags": {
+                  "type": "array",
+                  "items": { "type": "string", "pattern": "^[a-z][a-z0-9-]*$" }
+                },
+                "sensitivity": { "type": "string" },
+                "meta": { "type": "object", "additionalProperties": true }
+              }
+            }
+          },
+          "meta": { "type": "object", "additionalProperties": true }
+        }
+      }
+    },
+    "meta": { "type": "object", "additionalProperties": true }
+  },
+  "$defs": {
+    "duration": {
+      "type": "object",
+      "required": ["count", "period"],
+      "additionalProperties": false,
+      "properties": {
+        "count": { "type": "integer", "minimum": 1 },
+        "period": { "type": "string", "enum": ["minute", "hour", "day"] }
+      }
+    },
+    "tests": {
+      "type": "array",
+      "items": {
+        "oneOf": [
+          { "type": "string", "enum": ["unique", "not_null"] },
+          {
+            "type": "object",
+            "maxProperties": 1,
+            "minProperties": 1,
+            "additionalProperties": true
+          }
+        ]
+      }
+    }
+  }
+}

datalex_core/_schemas/datalex/term.schema.json

@@ -0,0 +1,30 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://duckcode.dev/schemas/datalex/term.schema.json",
+  "title": "DataLex glossary term",
+  "type": "object",
+  "required": ["kind", "name", "definition"],
+  "additionalProperties": false,
+  "properties": {
+    "kind": { "const": "term" },
+    "name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "previous_name": { "type": "string", "pattern": "^[a-z][a-z0-9_]*$" },
+    "definition": { "type": "string", "minLength": 1 },
+    "abbreviation": { "type": "string" },
+    "synonyms": {
+      "type": "array",
+      "items": { "type": "string", "minLength": 1 }
+    },
+    "steward": { "type": "string" },
+    "owner": { "type": "string" },
+    "tags": {
+      "type": "array",
+      "items": { "type": "string", "pattern": "^[a-z][a-z0-9-]*$" }
+    },
+    "related_terms": {
+      "type": "array",
+      "items": { "type": "string", "pattern": "^(term:)?[a-z][a-z0-9_]*$" }
+    },
+    "meta": { "type": "object", "additionalProperties": true }
+  }
+}