ibex 0.1.0 → 0.2.0

data/schema/automaton-ir-v1.schema.json

@@ -0,0 +1,401 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://raw.githubusercontent.com/ydah/ibex/main/schema/automaton-ir-v1.schema.json",
+  "title": "Ibex Automaton IR version 1",
+  "description": "The self-contained parser automaton exchanged between Ibex analysis and code generation stages.",
+  "type": "object",
+  "required": [
+    "ibex_ir",
+    "schema_version",
+    "algorithm",
+    "grammar_digest",
+    "grammar",
+    "states",
+    "conflict_summary"
+  ],
+  "properties": {
+    "ibex_ir": {
+      "const": "automaton"
+    },
+    "schema_version": {
+      "const": 1
+    },
+    "algorithm": {
+      "enum": [
+        "slr",
+        "lalr1",
+        "ielr1",
+        "lr1"
+      ]
+    },
+    "grammar_digest": {
+      "type": "string",
+      "pattern": "^sha256:[0-9a-f]{64}$"
+    },
+    "grammar": {
+      "$ref": "grammar-ir-v1.schema.json"
+    },
+    "states": {
+      "type": "array",
+      "minItems": 1,
+      "items": {
+        "$ref": "#/$defs/state"
+      }
+    },
+    "conflict_summary": {
+      "$ref": "#/$defs/conflict_summary"
+    }
+  },
+  "additionalProperties": false,
+  "$defs": {
+    "item": {
+      "type": "object",
+      "required": [
+        "production",
+        "dot",
+        "lookaheads"
+      ],
+      "properties": {
+        "production": {
+          "type": "integer",
+          "minimum": -1
+        },
+        "dot": {
+          "type": "integer",
+          "minimum": 0
+        },
+        "lookaheads": {
+          "type": "array",
+          "items": {
+            "type": "string",
+            "minLength": 1
+          }
+        }
+      },
+      "additionalProperties": false
+    },
+    "shift_action": {
+      "type": "object",
+      "required": [
+        "type",
+        "state"
+      ],
+      "properties": {
+        "type": {
+          "const": "shift"
+        },
+        "state": {
+          "type": "integer",
+          "minimum": 0
+        }
+      },
+      "additionalProperties": false
+    },
+    "reduce_action": {
+      "type": "object",
+      "required": [
+        "type",
+        "production"
+      ],
+      "properties": {
+        "type": {
+          "const": "reduce"
+        },
+        "production": {
+          "type": "integer",
+          "minimum": 0
+        }
+      },
+      "additionalProperties": false
+    },
+    "accept_action": {
+      "type": "object",
+      "required": [
+        "type"
+      ],
+      "properties": {
+        "type": {
+          "const": "accept"
+        }
+      },
+      "additionalProperties": false
+    },
+    "error_action": {
+      "type": "object",
+      "required": [
+        "type"
+      ],
+      "properties": {
+        "type": {
+          "const": "error"
+        }
+      },
+      "additionalProperties": false
+    },
+    "parser_action": {
+      "oneOf": [
+        {
+          "$ref": "#/$defs/shift_action"
+        },
+        {
+          "$ref": "#/$defs/reduce_action"
+        },
+        {
+          "$ref": "#/$defs/accept_action"
+        },
+        {
+          "$ref": "#/$defs/error_action"
+        }
+      ]
+    },
+    "nullable_parser_action": {
+      "oneOf": [
+        {
+          "$ref": "#/$defs/parser_action"
+        },
+        {
+          "type": "null"
+        }
+      ]
+    },
+    "resolution": {
+      "type": "object",
+      "required": [
+        "by",
+        "chose"
+      ],
+      "properties": {
+        "by": {
+          "enum": [
+            "definition_order",
+            "default_shift",
+            "precedence",
+            "associativity"
+          ]
+        },
+        "chose": {
+          "oneOf": [
+            {
+              "enum": [
+                "shift",
+                "reduce",
+                "error"
+              ]
+            },
+            {
+              "type": "integer",
+              "minimum": 0
+            }
+          ]
+        },
+        "associativity": {
+          "enum": [
+            "left",
+            "right",
+            "nonassoc"
+          ]
+        }
+      },
+      "additionalProperties": false
+    },
+    "shift_reduce_conflict": {
+      "type": "object",
+      "required": [
+        "type",
+        "symbol",
+        "shift_to",
+        "reduce",
+        "resolution"
+      ],
+      "properties": {
+        "type": {
+          "const": "shift_reduce"
+        },
+        "symbol": {
+          "type": "string",
+          "minLength": 1
+        },
+        "shift_to": {
+          "type": "integer",
+          "minimum": 0
+        },
+        "reduce": {
+          "type": "integer",
+          "minimum": 0
+        },
+        "resolution": {
+          "$ref": "#/$defs/resolution"
+        },
+        "midrule_origins": {
+          "type": "array",
+          "minItems": 1,
+          "items": {
+            "$ref": "grammar-ir-v1.schema.json#/$defs/location"
+          }
+        },
+        "entries": {
+          "type": "array",
+          "minItems": 1,
+          "uniqueItems": true,
+          "items": {
+            "type": "string",
+            "minLength": 1
+          }
+        },
+        "composite": {
+          "type": "boolean"
+        }
+      },
+      "additionalProperties": false
+    },
+    "reduce_reduce_conflict": {
+      "type": "object",
+      "required": [
+        "type",
+        "symbol",
+        "reductions",
+        "resolution"
+      ],
+      "properties": {
+        "type": {
+          "const": "reduce_reduce"
+        },
+        "symbol": {
+          "type": "string",
+          "minLength": 1
+        },
+        "reductions": {
+          "type": "array",
+          "minItems": 2,
+          "uniqueItems": true,
+          "items": {
+            "type": "integer",
+            "minimum": 0
+          }
+        },
+        "resolution": {
+          "$ref": "#/$defs/resolution"
+        },
+        "midrule_origins": {
+          "type": "array",
+          "minItems": 1,
+          "items": {
+            "$ref": "grammar-ir-v1.schema.json#/$defs/location"
+          }
+        },
+        "entries": {
+          "type": "array",
+          "minItems": 1,
+          "uniqueItems": true,
+          "items": {
+            "type": "string",
+            "minLength": 1
+          }
+        },
+        "composite": {
+          "type": "boolean"
+        }
+      },
+      "additionalProperties": false
+    },
+    "state": {
+      "type": "object",
+      "required": [
+        "id",
+        "items",
+        "transitions",
+        "actions",
+        "gotos",
+        "default_action",
+        "conflicts"
+      ],
+      "properties": {
+        "id": {
+          "type": "integer",
+          "minimum": 0
+        },
+        "items": {
+          "type": "array",
+          "items": {
+            "$ref": "#/$defs/item"
+          }
+        },
+        "transitions": {
+          "$ref": "#/$defs/state_map"
+        },
+        "actions": {
+          "type": "object",
+          "additionalProperties": {
+            "$ref": "#/$defs/parser_action"
+          }
+        },
+        "gotos": {
+          "$ref": "#/$defs/state_map"
+        },
+        "default_action": {
+          "$ref": "#/$defs/nullable_parser_action"
+        },
+        "conflicts": {
+          "type": "array",
+          "items": {
+            "oneOf": [
+              {
+                "$ref": "#/$defs/shift_reduce_conflict"
+              },
+              {
+                "$ref": "#/$defs/reduce_reduce_conflict"
+              }
+            ]
+          }
+        }
+      },
+      "additionalProperties": false
+    },
+    "state_map": {
+      "type": "object",
+      "additionalProperties": {
+        "type": "integer",
+        "minimum": 0
+      }
+    },
+    "conflict_summary": {
+      "type": "object",
+      "required": [
+        "sr",
+        "resolved_sr",
+        "rr",
+        "expected_sr",
+        "expectation_met"
+      ],
+      "properties": {
+        "sr": {
+          "type": "integer",
+          "minimum": 0
+        },
+        "resolved_sr": {
+          "type": "integer",
+          "minimum": 0
+        },
+        "rr": {
+          "type": "integer",
+          "minimum": 0
+        },
+        "expected_sr": {
+          "type": "integer",
+          "minimum": 0
+        },
+        "expectation_met": {
+          "type": "boolean"
+        },
+        "expected_rr": {
+          "type": "integer",
+          "minimum": 0
+        },
+        "rr_expectation_met": {
+          "type": "boolean"
+        }
+      },
+      "additionalProperties": false
+    }
+  }
+}

data/schema/automaton-ir-v2.schema.json

@@ -0,0 +1,58 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://raw.githubusercontent.com/ydah/ibex/main/schema/automaton-ir-v2.schema.json",
+  "title": "Ibex Automaton IR version 2",
+  "description": "A self-contained parser automaton embedding Grammar IR version 2.",
+  "type": "object",
+  "required": [
+    "ibex_ir",
+    "schema_version",
+    "algorithm",
+    "grammar_digest",
+    "grammar",
+    "states",
+    "conflict_summary"
+  ],
+  "properties": {
+    "ibex_ir": {
+      "const": "automaton"
+    },
+    "schema_version": {
+      "const": 2
+    },
+    "algorithm": {
+      "enum": [
+        "slr",
+        "lalr1",
+        "ielr1",
+        "lr1"
+      ]
+    },
+    "grammar_digest": {
+      "type": "string",
+      "pattern": "^sha256:[0-9a-f]{64}$"
+    },
+    "grammar": {
+      "$ref": "grammar-ir-v2.schema.json"
+    },
+    "entry_states": {
+      "type": "object",
+      "minProperties": 1,
+      "additionalProperties": {
+        "type": "integer",
+        "minimum": 0
+      }
+    },
+    "states": {
+      "type": "array",
+      "minItems": 1,
+      "items": {
+        "$ref": "automaton-ir-v1.schema.json#/$defs/state"
+      }
+    },
+    "conflict_summary": {
+      "$ref": "automaton-ir-v1.schema.json#/$defs/conflict_summary"
+    }
+  },
+  "additionalProperties": false
+}

data/schema/benchmark-v1.schema.json

@@ -0,0 +1,212 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://raw.githubusercontent.com/ydah/ibex/main/schema/benchmark-v1.schema.json",
+  "title": "Ibex benchmark artifact v1",
+  "type": "object",
+  "additionalProperties": false,
+  "required": [
+    "artifact",
+    "schema_version",
+    "recorded_at",
+    "environment",
+    "configuration",
+    "measurements",
+    "structure",
+    "digests"
+  ],
+  "properties": {
+    "artifact": { "const": "ibex_benchmark" },
+    "schema_version": { "const": 1 },
+    "recorded_at": { "type": "string", "format": "date-time" },
+    "environment": { "$ref": "#/$defs/environment" },
+    "configuration": { "$ref": "#/$defs/configuration" },
+    "measurements": { "$ref": "#/$defs/measurements" },
+    "structure": { "$ref": "#/$defs/structure" },
+    "digests": { "$ref": "#/$defs/digests" }
+  },
+  "$defs": {
+    "nonNegativeNumber": {
+      "type": "number",
+      "minimum": 0
+    },
+    "nonNegativeInteger": {
+      "type": "integer",
+      "minimum": 0
+    },
+    "sha256": {
+      "type": "string",
+      "pattern": "^[0-9a-f]{64}$"
+    },
+    "environment": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": [
+        "ruby_engine",
+        "ruby_version",
+        "ruby_platform",
+        "ibex_version",
+        "host_os",
+        "host_cpu",
+        "processors"
+      ],
+      "properties": {
+        "ruby_engine": { "type": "string", "minLength": 1 },
+        "ruby_version": { "type": "string", "minLength": 1 },
+        "ruby_platform": { "type": "string", "minLength": 1 },
+        "ibex_version": { "type": "string", "minLength": 1 },
+        "host_os": { "type": "string", "minLength": 1 },
+        "host_cpu": { "type": "string", "minLength": 1 },
+        "processors": { "type": "integer", "minimum": 1 }
+      }
+    },
+    "configuration": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": [
+        "grammar",
+        "input",
+        "seed",
+        "build_iterations",
+        "runtime_iterations"
+      ],
+      "properties": {
+        "grammar": { "type": "string", "minLength": 1 },
+        "input": { "type": "string", "minLength": 1 },
+        "seed": { "type": "integer" },
+        "build_iterations": { "type": "integer", "minimum": 1 },
+        "runtime_iterations": { "type": "integer", "minimum": 1 }
+      }
+    },
+    "measurements": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": [
+        "generation_ms",
+        "runtime_parse_ms",
+        "peak_rss_bytes",
+        "stage_ms"
+      ],
+      "properties": {
+        "generation_ms": { "$ref": "#/$defs/nonNegativeNumber" },
+        "runtime_parse_ms": {
+          "type": "object",
+          "additionalProperties": false,
+          "required": ["plain", "compact"],
+          "properties": {
+            "plain": { "$ref": "#/$defs/nonNegativeNumber" },
+            "compact": { "$ref": "#/$defs/nonNegativeNumber" }
+          }
+        },
+        "peak_rss_bytes": {
+          "oneOf": [
+            { "$ref": "#/$defs/nonNegativeInteger" },
+            { "type": "null" }
+          ]
+        },
+        "stage_ms": {
+          "type": "object",
+          "additionalProperties": false,
+          "required": [
+            "parse",
+            "normalize",
+            "automaton",
+            "table_plain",
+            "table_compact",
+            "codegen_plain",
+            "codegen_compact"
+          ],
+          "properties": {
+            "parse": { "$ref": "#/$defs/nonNegativeNumber" },
+            "normalize": { "$ref": "#/$defs/nonNegativeNumber" },
+            "automaton": { "$ref": "#/$defs/nonNegativeNumber" },
+            "table_plain": { "$ref": "#/$defs/nonNegativeNumber" },
+            "table_compact": { "$ref": "#/$defs/nonNegativeNumber" },
+            "codegen_plain": { "$ref": "#/$defs/nonNegativeNumber" },
+            "codegen_compact": { "$ref": "#/$defs/nonNegativeNumber" }
+          }
+        }
+      }
+    },
+    "table": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": [
+        "action_cells",
+        "goto_cells",
+        "default_cells",
+        "total_cells",
+        "bytes"
+      ],
+      "properties": {
+        "action_cells": { "$ref": "#/$defs/nonNegativeInteger" },
+        "goto_cells": { "$ref": "#/$defs/nonNegativeInteger" },
+        "default_cells": { "$ref": "#/$defs/nonNegativeInteger" },
+        "total_cells": { "$ref": "#/$defs/nonNegativeInteger" },
+        "bytes": { "$ref": "#/$defs/nonNegativeInteger" }
+      }
+    },
+    "variantSizes": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": ["plain", "compact"],
+      "properties": {
+        "plain": { "$ref": "#/$defs/nonNegativeInteger" },
+        "compact": { "$ref": "#/$defs/nonNegativeInteger" }
+      }
+    },
+    "structure": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": [
+        "productions",
+        "canonical_intermediate_states",
+        "final_states",
+        "tables",
+        "generated_output_bytes"
+      ],
+      "properties": {
+        "productions": { "$ref": "#/$defs/nonNegativeInteger" },
+        "canonical_intermediate_states": { "$ref": "#/$defs/nonNegativeInteger" },
+        "final_states": { "$ref": "#/$defs/nonNegativeInteger" },
+        "tables": {
+          "type": "object",
+          "additionalProperties": false,
+          "required": ["plain", "compact"],
+          "properties": {
+            "plain": { "$ref": "#/$defs/table" },
+            "compact": { "$ref": "#/$defs/table" }
+          }
+        },
+        "generated_output_bytes": { "$ref": "#/$defs/variantSizes" }
+      }
+    },
+    "digests": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": [
+        "grammar_source_sha256",
+        "runtime_input_sha256",
+        "grammar_ir_sha256",
+        "automaton_ir_sha256",
+        "plain_tables_sha256",
+        "compact_tables_sha256",
+        "plain_output_sha256",
+        "compact_output_sha256",
+        "runtime_result_sha256",
+        "artifact_sha256"
+      ],
+      "properties": {
+        "grammar_source_sha256": { "$ref": "#/$defs/sha256" },
+        "runtime_input_sha256": { "$ref": "#/$defs/sha256" },
+        "grammar_ir_sha256": { "$ref": "#/$defs/sha256" },
+        "automaton_ir_sha256": { "$ref": "#/$defs/sha256" },
+        "plain_tables_sha256": { "$ref": "#/$defs/sha256" },
+        "compact_tables_sha256": { "$ref": "#/$defs/sha256" },
+        "plain_output_sha256": { "$ref": "#/$defs/sha256" },
+        "compact_output_sha256": { "$ref": "#/$defs/sha256" },
+        "runtime_result_sha256": { "$ref": "#/$defs/sha256" },
+        "artifact_sha256": { "$ref": "#/$defs/sha256" }
+      }
+    }
+  }
+}