ibex 0.2.0 → 0.4.0

data/schema/grammar-ir.schema.json

@@ -0,0 +1,271 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://raw.githubusercontent.com/ydah/ibex/main/schema/grammar-ir.schema.json",
+  "title": "Ibex current Grammar IR",
+  "description": "The normalized grammar with one root-owned parser contract.",
+  "type": "object",
+  "required": [
+    "ibex_ir",
+    "schema_version",
+    "class_name",
+    "superclass",
+    "start",
+    "expect",
+    "options",
+    "symbols",
+    "productions",
+    "user_code",
+    "conversions",
+    "warnings",
+    "source_provenance",
+    "parser_contract"
+  ],
+  "properties": {
+    "ibex_ir": {
+      "const": "grammar"
+    },
+    "schema_version": {
+      "const": 1
+    },
+    "class_name": {
+      "type": "string",
+      "minLength": 1
+    },
+    "superclass": {
+      "type": [
+        "string",
+        "null"
+      ]
+    },
+    "start": {
+      "type": "string",
+      "minLength": 1
+    },
+    "mode": {
+      "enum": [
+        "default",
+        "extended"
+      ]
+    },
+    "starts": {
+      "type": "array",
+      "minItems": 2,
+      "uniqueItems": true,
+      "items": {
+        "type": "string",
+        "minLength": 1
+      }
+    },
+    "expect": {
+      "type": "integer",
+      "minimum": 0
+    },
+    "expect_rr": {
+      "type": "integer",
+      "minimum": 0
+    },
+    "options": {
+      "$ref": "grammar-ir-extensions.schema.json#/$defs/options"
+    },
+    "params": {
+      "type": "array",
+      "items": {
+        "$ref": "grammar-ir-extensions.schema.json#/$defs/parser_parameter"
+      }
+    },
+    "printers": {
+      "type": "array",
+      "items": {
+        "$ref": "grammar-ir-extensions.schema.json#/$defs/value_printer"
+      }
+    },
+    "tests": {
+      "type": "array",
+      "items": {
+        "$ref": "grammar-ir-extensions.schema.json#/$defs/grammar_test"
+      }
+    },
+    "recovery": {
+      "$ref": "grammar-ir-extensions.schema.json#/$defs/recovery_policy"
+    },
+    "lexer": {
+      "$ref": "lexer-ir-v1.schema.json"
+    },
+    "symbols": {
+      "type": "array",
+      "minItems": 2,
+      "items": {
+        "$ref": "grammar-ir-extensions.schema.json#/$defs/symbol"
+      }
+    },
+    "productions": {
+      "type": "array",
+      "items": {
+        "$ref": "grammar-ir-extensions.schema.json#/$defs/production"
+      }
+    },
+    "user_code": {
+      "$ref": "grammar-ir-foundation.schema.json#/$defs/string_map"
+    },
+    "conversions": {
+      "$ref": "grammar-ir-foundation.schema.json#/$defs/string_map"
+    },
+    "warnings": {
+      "type": "array",
+      "items": {
+        "$ref": "grammar-ir-foundation.schema.json#/$defs/warning"
+      }
+    },
+    "user_code_chunks": {
+      "type": "object",
+      "additionalProperties": {
+        "type": "array",
+        "items": {
+          "$ref": "grammar-ir-foundation.schema.json#/$defs/user_code_chunk"
+        }
+      }
+    },
+    "source_provenance": {
+      "$ref": "grammar-ir-extensions.schema.json#/$defs/nullable_source_provenance"
+    },
+    "parser_contract": {
+      "$ref": "#/$defs/parser_contract"
+    }
+  },
+  "additionalProperties": false,
+  "$defs": {
+    "parser_contract": {
+      "type": "object",
+      "required": [
+        "algorithm",
+        "entries",
+        "cst_trivia"
+      ],
+      "properties": {
+        "algorithm": {
+          "$ref": "#/$defs/algorithm_contract"
+        },
+        "entries": {
+          "$ref": "#/$defs/entries_contract"
+        },
+        "cst_trivia": {
+          "$ref": "#/$defs/cst_trivia_contract"
+        }
+      },
+      "additionalProperties": false
+    },
+    "unspecified_contract": {
+      "type": "object",
+      "required": [
+        "value",
+        "explicit",
+        "loc"
+      ],
+      "properties": {
+        "value": {
+          "type": "null"
+        },
+        "explicit": {
+          "const": false
+        },
+        "loc": {
+          "type": "null"
+        }
+      },
+      "additionalProperties": false
+    },
+    "algorithm_contract": {
+      "oneOf": [
+        {
+          "$ref": "#/$defs/unspecified_contract"
+        },
+        {
+          "type": "object",
+          "required": [
+            "value",
+            "explicit",
+            "loc"
+          ],
+          "properties": {
+            "value": {
+              "enum": [
+                "slr",
+                "lalr",
+                "ielr",
+                "lr1"
+              ]
+            },
+            "explicit": {
+              "const": true
+            },
+            "loc": {
+              "$ref": "grammar-ir-foundation.schema.json#/$defs/location"
+            }
+          },
+          "additionalProperties": false
+        }
+      ]
+    },
+    "entries_contract": {
+      "oneOf": [
+        {
+          "$ref": "#/$defs/unspecified_contract"
+        },
+        {
+          "type": "object",
+          "required": [
+            "value",
+            "explicit",
+            "loc"
+          ],
+          "properties": {
+            "value": {
+              "enum": [
+                "shared",
+                "isolated"
+              ]
+            },
+            "explicit": {
+              "const": true
+            },
+            "loc": {
+              "$ref": "grammar-ir-foundation.schema.json#/$defs/location"
+            }
+          },
+          "additionalProperties": false
+        }
+      ]
+    },
+    "cst_trivia_contract": {
+      "oneOf": [
+        {
+          "$ref": "#/$defs/unspecified_contract"
+        },
+        {
+          "type": "object",
+          "required": [
+            "value",
+            "explicit",
+            "loc"
+          ],
+          "properties": {
+            "value": {
+              "enum": [
+                "leading",
+                "balanced",
+                "drop"
+              ]
+            },
+            "explicit": {
+              "const": true
+            },
+            "loc": {
+              "$ref": "grammar-ir-foundation.schema.json#/$defs/location"
+            }
+          },
+          "additionalProperties": false
+        }
+      ]
+    }
+  }
+}

data/schema/ielr-benchmark-v1.schema.json

@@ -0,0 +1,85 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://raw.githubusercontent.com/ydah/ibex/main/schema/ielr-benchmark-v1.schema.json",
+  "title": "Ibex direct IELR construction benchmark v1",
+  "type": "object",
+  "required": ["ibex_report", "schema_version", "trust", "environment", "configuration", "workloads", "limitations"],
+  "properties": {
+    "ibex_report": { "const": "ielr_benchmark" },
+    "schema_version": { "const": 1 },
+    "trust": { "const": "internal_local_observation" },
+    "environment": { "$ref": "#/$defs/environment" },
+    "configuration": {
+      "type": "object",
+      "required": ["wall_seconds_per_run", "workload_count", "workload_selection"],
+      "properties": {
+        "wall_seconds_per_run": { "type": "number", "exclusiveMinimum": 0 },
+        "workload_count": { "type": "integer", "minimum": 1 },
+        "workload_selection": { "type": "array", "items": { "type": "string", "minLength": 1 }, "minItems": 1 }
+      },
+      "additionalProperties": false
+    },
+    "workloads": {
+      "type": "array",
+      "items": { "$ref": "#/$defs/workload" },
+      "minItems": 2
+    },
+    "limitations": { "type": "array", "items": { "type": "string", "minLength": 1 }, "minItems": 1, "uniqueItems": true }
+  },
+  "additionalProperties": false,
+  "$defs": {
+    "environment": {
+      "type": "object",
+      "required": ["ruby_engine", "ruby_version", "ruby_platform", "host_os", "host_cpu"],
+      "properties": {
+        "ruby_engine": { "type": "string", "minLength": 1 },
+        "ruby_version": { "type": "string", "minLength": 1 },
+        "ruby_platform": { "type": "string", "minLength": 1 },
+        "host_os": { "type": "string", "minLength": 1 },
+        "host_cpu": { "type": "string", "minLength": 1 }
+      },
+      "additionalProperties": false
+    },
+    "status_value": {
+      "type": "object",
+      "required": ["status"],
+      "properties": {
+        "status": { "type": "string", "minLength": 1 },
+        "value": {},
+        "reason": { "type": "string", "minLength": 1 }
+      },
+      "additionalProperties": false
+    },
+    "workload": {
+      "type": "object",
+      "required": ["id", "path", "source_sha256", "strategy", "status", "structural", "conflicts", "observations"],
+      "properties": {
+        "id": { "type": "string", "minLength": 1 },
+        "path": { "type": "string", "minLength": 1 },
+        "source_sha256": { "type": "string", "pattern": "^[0-9a-f]{64}$" },
+        "strategy": { "enum": ["direct", "partition"] },
+        "status": { "enum": ["completed", "limit_exceeded", "resource_exhausted", "failed"] },
+        "structural": { "type": "object", "additionalProperties": { "$ref": "#/$defs/status_value" } },
+        "conflicts": { "type": ["object", "null"], "additionalProperties": { "type": "integer", "minimum": 0 } },
+        "observations": {
+          "type": "object",
+          "required": ["elapsed_seconds"],
+          "properties": {
+            "elapsed_seconds": {
+              "type": "object",
+              "required": ["status", "value", "release_gate"],
+              "properties": {
+                "status": { "const": "observation" },
+                "value": { "type": "number", "minimum": 0 },
+                "release_gate": { "const": false }
+              },
+              "additionalProperties": false
+            }
+          },
+          "additionalProperties": false
+        }
+      },
+      "additionalProperties": false
+    }
+  }
+}

data/schema/impact-v1.schema.json

@@ -0,0 +1,208 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://raw.githubusercontent.com/ydah/ibex/main/schema/impact-v1.schema.json",
+  "title": "Ibex grammar impact report v1",
+  "type": "object",
+  "required": ["ibex_report", "schema_version", "mode", "algorithm", "grammar_digest", "seeds", "symbols", "automaton", "actions", "coverage", "warnings", "totals"],
+  "properties": {
+    "ibex_report": { "const": "impact" },
+    "schema_version": { "const": 1 },
+    "mode": { "enum": ["potential", "diff"] },
+    "algorithm": { "enum": ["slr", "lalr1", "ielr1", "lr1"] },
+    "grammar_digest": {
+      "type": "object",
+      "required": ["before", "after"],
+      "properties": {
+        "before": { "type": ["string", "null"], "pattern": "^sha256:[0-9a-f]{64}$" },
+        "after": { "type": ["string", "null"], "pattern": "^sha256:[0-9a-f]{64}$" }
+      },
+      "additionalProperties": false
+    },
+    "seeds": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "required": ["symbol", "origin", "nullable_boundary"],
+        "properties": {
+          "symbol": { "type": "string" },
+          "origin": { "enum": ["symbol", "diff"] },
+          "nullable_boundary": { "type": "boolean" }
+        },
+        "additionalProperties": false
+      },
+      "maxItems": 100000
+    },
+    "symbols": {
+      "type": "array",
+      "items": { "$ref": "#/$defs/symbol" },
+      "maxItems": 100000
+    },
+    "automaton": {
+      "type": "object",
+      "required": ["states", "affected_states", "conflicts", "unreachable", "unreachable_nonterminals"],
+      "properties": {
+        "states": {
+          "type": "object",
+          "required": ["before", "after", "delta"],
+          "properties": {
+            "before": { "type": ["integer", "null"], "minimum": 0 },
+            "after": { "type": "integer", "minimum": 0 },
+            "delta": { "type": ["integer", "null"] }
+          },
+          "additionalProperties": false
+        },
+        "affected_states": { "type": "array", "items": { "type": "integer", "minimum": 0 }, "uniqueItems": true, "maxItems": 100000 },
+        "conflicts": { "$ref": "#/$defs/classification" },
+        "unreachable": { "type": "array", "items": { "type": "integer", "minimum": 0 }, "uniqueItems": true, "maxItems": 100000 },
+        "unreachable_nonterminals": {
+          "type": "array",
+          "items": { "type": "string" },
+          "uniqueItems": true,
+          "maxItems": 100000
+        }
+      },
+      "additionalProperties": false
+    },
+    "actions": {
+      "type": "array",
+      "items": { "$ref": "#/$defs/action" },
+      "maxItems": 100000
+    },
+    "coverage": {
+      "type": "object",
+      "required": ["status", "reports"],
+      "properties": {
+        "status": { "enum": ["not_requested", "matched", "unmatched"] },
+        "reports": {
+          "type": "array",
+          "items": {
+            "type": "object",
+            "required": ["path", "productions"],
+            "properties": {
+              "path": { "type": "string" },
+              "productions": { "type": "array", "items": { "type": "integer", "minimum": 0 }, "uniqueItems": true, "maxItems": 100000 }
+            },
+            "additionalProperties": false
+          },
+          "maxItems": 100000
+        }
+      },
+      "additionalProperties": false
+    },
+    "warnings": { "type": "array", "items": { "type": "string" }, "uniqueItems": true, "maxItems": 100000 },
+    "totals": {
+      "type": "object",
+      "required": ["info", "low", "medium", "high", "critical"],
+      "properties": {
+        "info": { "type": "integer", "minimum": 0 },
+        "low": { "type": "integer", "minimum": 0 },
+        "medium": { "type": "integer", "minimum": 0 },
+        "high": { "type": "integer", "minimum": 0 },
+        "critical": { "type": "integer", "minimum": 0 }
+      },
+      "additionalProperties": false
+    }
+  },
+  "additionalProperties": false,
+  "$defs": {
+    "classification": {
+      "type": "object",
+      "required": ["added", "removed", "changed"],
+      "properties": {
+        "added": { "type": "array", "items": { "$ref": "#/$defs/classification_record" }, "maxItems": 100000 },
+        "removed": { "type": "array", "items": { "$ref": "#/$defs/classification_record" }, "maxItems": 100000 },
+        "changed": { "type": "array", "items": { "$ref": "#/$defs/classification_record" }, "maxItems": 100000 }
+      },
+      "additionalProperties": false
+    },
+    "classification_record": {
+      "oneOf": [
+        {
+          "type": "object",
+          "required": ["id", "value"],
+          "properties": { "id": { "type": "string" }, "value": {} },
+          "additionalProperties": false
+        },
+        {
+          "type": "object",
+          "required": ["id", "before", "after"],
+          "properties": { "id": { "type": "string" }, "before": {}, "after": {} },
+          "additionalProperties": false
+        }
+      ]
+    },
+    "change": {
+      "type": "object",
+      "required": ["added", "removed"],
+      "properties": {
+        "added": { "type": "array", "items": { "type": "string" }, "maxItems": 100000 },
+        "removed": { "type": "array", "items": { "type": "string" }, "maxItems": 100000 }
+      },
+      "additionalProperties": false
+    },
+    "symbol": {
+      "type": "object",
+      "required": ["symbol", "severity", "kinds", "distance", "component", "sets", "witness"],
+      "properties": {
+        "symbol": { "type": "string" },
+        "severity": { "enum": ["info", "low", "medium", "high", "critical"] },
+        "kinds": { "type": "array", "items": { "type": "string" }, "uniqueItems": true, "maxItems": 100000 },
+        "distance": { "type": "integer", "minimum": 0 },
+        "component": { "type": "array", "items": { "type": "string" }, "uniqueItems": true, "maxItems": 100000 },
+        "sets": {
+          "type": "object",
+          "required": ["first", "follow", "nullable"],
+          "properties": {
+            "first": { "$ref": "#/$defs/change" },
+            "follow": { "$ref": "#/$defs/change" },
+            "nullable": {
+              "type": ["object", "null"],
+              "required": ["before", "after"],
+              "properties": { "before": { "type": "boolean" }, "after": { "type": "boolean" } },
+              "additionalProperties": false
+            }
+          },
+          "additionalProperties": false
+        },
+        "witness": { "type": "array", "items": { "$ref": "#/$defs/witness" }, "maxItems": 100000 }
+      },
+      "additionalProperties": false
+    },
+    "witness": {
+      "type": "object",
+      "required": ["from", "to", "production", "loc"],
+      "properties": {
+        "from": { "type": "string" },
+        "to": { "type": "string" },
+        "production": { "type": ["string", "null"] },
+        "loc": {
+          "type": ["object", "null"],
+          "properties": {
+            "file": { "type": "string" },
+            "line": { "type": "integer", "minimum": 1 },
+            "column": { "type": "integer", "minimum": 1 }
+          },
+          "additionalProperties": false
+        }
+      },
+      "additionalProperties": false
+    },
+    "action": {
+      "type": "object",
+      "required": ["production", "severity", "reason", "context_length", "loc"],
+      "properties": {
+        "production": { "type": "string" },
+        "severity": { "enum": ["medium", "high"] },
+        "reason": { "type": "string" },
+        "context_length": {
+          "type": "object",
+          "required": ["before", "after"],
+          "properties": { "before": { "type": "integer", "minimum": 0 }, "after": { "type": "integer", "minimum": 0 } },
+          "additionalProperties": false
+        },
+        "loc": { "$ref": "#/$defs/witness/properties/loc" }
+      },
+      "additionalProperties": false
+    }
+  }
+}