diff --git a/examples/json_schema_to_grammar.py b/examples/json_schema_to_grammar.py index 7d889c3fe..b1b7a974d 100755 --- a/examples/json_schema_to_grammar.py +++ b/examples/json_schema_to_grammar.py @@ -4,7 +4,7 @@ import itertools import json import re import sys -from typing import Any, Dict, List, Set, Tuple, Union +from typing import Any, Callable, Dict, List, Optional, Set, Tuple, Union def _build_repetition(item_rule, min_items, max_items, separator_rule=None): @@ -23,6 +23,169 @@ def _build_repetition(item_rule, min_items, max_items, separator_rule=None): result = item_rule + ' ' + _build_repetition(f'({separator_rule} {item_rule})', min_items - 1 if min_items > 0 else 0, max_items - 1 if max_items is not None else None) return f'({result})?' if min_items == 0 else result +def _generate_min_max_int(min_value: Optional[int], max_value: Optional[int], out: list, decimals_left: int = 16, top_level: bool = True): + has_min = min_value != None + has_max = max_value != None + + def digit_range(from_char: str, to_char: str): + out.append("[") + if from_char == to_char: + out.append(from_char) + else: + out.append(from_char) + out.append("-") + out.append(to_char) + out.append("]") + + def more_digits(min_digits: int, max_digits: int): + out.append("[0-9]") + if min_digits == max_digits and min_digits == 1: + return + out.append("{") + out.append(str(min_digits)) + if max_digits != min_digits: + out.append(",") + if max_digits != sys.maxsize: + out.append(str(max_digits)) + out.append("}") + + def uniform_range(from_str: str, to_str: str): + i = 0 + while i < len(from_str) and from_str[i] == to_str[i]: + i += 1 + if i > 0: + out.append("\"") + out.append(from_str[:i]) + out.append("\"") + if i < len(from_str): + if i > 0: + out.append(" ") + sub_len = len(from_str) - i - 1 + if sub_len > 0: + from_sub = from_str[i+1:] + to_sub = to_str[i+1:] + sub_zeros = "0" * sub_len + sub_nines = "9" * sub_len + + to_reached = False + out.append("(") + if from_sub == sub_zeros: + digit_range(from_str[i], chr(ord(to_str[i]) - 1)) + out.append(" ") + more_digits(sub_len, sub_len) + else: + out.append("[") + out.append(from_str[i]) + out.append("] ") + out.append("(") + uniform_range(from_sub, sub_nines) + out.append(")") + if ord(from_str[i]) < ord(to_str[i]) - 1: + out.append(" | ") + if to_sub == sub_nines: + digit_range(chr(ord(from_str[i]) + 1), to_str[i]) + to_reached = True + else: + digit_range(chr(ord(from_str[i]) + 1), chr(ord(to_str[i]) - 1)) + out.append(" ") + more_digits(sub_len, sub_len) + if not to_reached: + out.append(" | ") + digit_range(to_str[i], to_str[i]) + out.append(" ") + uniform_range(sub_zeros, to_sub) + out.append(")") + else: + out.append("[") + out.append(from_str[i]) + out.append("-") + out.append(to_str[i]) + out.append("]") + + if has_min and has_max: + if min_value < 0 and max_value < 0: + out.append("\"-\" (") + _generate_min_max_int(-max_value, -min_value, out, decimals_left, top_level=True) + out.append(")") + return + + if min_value < 0: + out.append("\"-\" (") + _generate_min_max_int(0, -min_value, out, decimals_left, top_level=True) + out.append(") | ") + min_value = 0 + + min_s = str(min_value) + max_s = str(max_value) + min_digits = len(min_s) + max_digits = len(max_s) + + for digits in range(min_digits, max_digits): + uniform_range(min_s, "9" * digits) + min_s = "1" + "0" * digits + out.append(" | ") + uniform_range(min_s, max_s) + return + + less_decimals = max(decimals_left - 1, 1) + + if has_min: + if min_value < 0: + out.append("\"-\" (") + _generate_min_max_int(None, -min_value, out, decimals_left, top_level=False) + out.append(") | [0] | [1-9] ") + more_digits(0, decimals_left - 1) + elif min_value == 0: + if top_level: + out.append("[0] | [1-9] ") + more_digits(0, less_decimals) + else: + more_digits(1, decimals_left) + elif min_value <= 9: + c = str(min_value) + if min_value > (1 if top_level else 0): + digit_range("0", chr(ord(c) - 1)) + out.append(" ") + more_digits(1, less_decimals) + out.append(" | ") + digit_range(c, "9") + out.append(" ") + more_digits(0, less_decimals) + else: + min_s = str(min_value) + length = len(min_s) + c = min_s[0] + + if c > "1": + digit_range("1" if top_level else "0", chr(ord(c) - 1)) + out.append(" ") + more_digits(length, less_decimals) + out.append(" | ") + digit_range(c, c) + out.append(" (") + _generate_min_max_int(int(min_s[1:]), None, out, less_decimals, top_level=False) + out.append(")") + if c < "9": + out.append(" | ") + digit_range(chr(ord(c) + 1), "9") + out.append(" ") + more_digits(length - 1, less_decimals) + return + + if has_max: + if max_value >= 0: + if top_level: + out.append("\"-\" [1-9] ") + more_digits(0, less_decimals) + out.append(" | ") + _generate_min_max_int(0, max_value, out, decimals_left, top_level=True) + else: + out.append("\"-\" (") + _generate_min_max_int(-max_value, None, out, decimals_left, top_level=False) + out.append(")") + return + + raise RuntimeError("At least one of min_value or max_value must be set") class BuiltinRule: def __init__(self, content: str, deps: list = None): @@ -433,6 +596,24 @@ class SchemaConverter: return self._add_rule(rule_name, r'"\"" ' + _build_repetition(char_rule, min_len, max_len) + r' "\"" space') + elif schema_type in (None, 'integer') and \ + ('minimum' in schema or 'exclusiveMinimum' in schema or 'maximum' in schema or 'exclusiveMaximum' in schema): + min_value = None + max_value = None + if 'minimum' in schema: + min_value = schema['minimum'] + elif 'exclusiveMinimum' in schema: + min_value = schema['exclusiveMinimum'] + 1 + if 'maximum' in schema: + max_value = schema['maximum'] + elif 'exclusiveMaximum' in schema: + max_value = schema['exclusiveMaximum'] - 1 + + out = ["("] + _generate_min_max_int(min_value, max_value, out) + out.append(") space") + return self._add_rule(rule_name, ''.join(out)) + elif (schema_type == 'object') or (len(schema) == 0): return self._add_rule(rule_name, self._add_primitive('object', PRIMITIVE_RULES['object'])) diff --git a/examples/server/public/json-schema-to-grammar.mjs b/examples/server/public/json-schema-to-grammar.mjs index cef11eab8..7c0f9c7a2 100644 --- a/examples/server/public/json-schema-to-grammar.mjs +++ b/examples/server/public/json-schema-to-grammar.mjs @@ -24,6 +24,200 @@ function _buildRepetition(itemRule, minItems, maxItems, opts={}) { return minItems === 0 ? `(${result})?` : result; } +function _generateMinMaxInt(minValue, maxValue, out, decimalsLeft = 16, topLevel = true) { + const hasMin = minValue !== null; + const hasMax = maxValue !== null; + + function digitRange(fromChar, toChar) { + out.push("["); + if (fromChar === toChar) { + out.push(fromChar); + } else { + out.push(fromChar); + out.push("-"); + out.push(toChar); + } + out.push("]"); + } + + function moreDigits(minDigits, maxDigits) { + out.push("[0-9]"); + if (minDigits === maxDigits && minDigits === 1) { + return; + } + out.push("{"); + out.push(minDigits.toString()); + if (maxDigits !== minDigits) { + out.push(","); + if (maxDigits !== Number.MAX_SAFE_INTEGER) { + out.push(maxDigits.toString()); + } + } + out.push("}"); + } + + function uniformRange(fromStr, toStr) { + let i = 0; + while (i < fromStr.length && fromStr[i] === toStr[i]) { + i++; + } + if (i > 0) { + out.push("\""); + out.push(fromStr.slice(0, i)); + out.push("\""); + } + if (i < fromStr.length) { + if (i > 0) { + out.push(" "); + } + const subLen = fromStr.length - i - 1; + if (subLen > 0) { + const fromSub = fromStr.slice(i + 1); + const toSub = toStr.slice(i + 1); + const subZeros = "0".repeat(subLen); + const subNines = "9".repeat(subLen); + + let toReached = false; + out.push("("); + if (fromSub === subZeros) { + digitRange(fromStr[i], String.fromCharCode(toStr.charCodeAt(i) - 1)); + out.push(" "); + moreDigits(subLen, subLen); + } else { + out.push("["); + out.push(fromStr[i]); + out.push("] "); + out.push("("); + uniformRange(fromSub, subNines); + out.push(")"); + if (fromStr.charCodeAt(i) < toStr.charCodeAt(i) - 1) { + out.push(" | "); + if (toSub === subNines) { + digitRange(String.fromCharCode(fromStr.charCodeAt(i) + 1), toStr[i]); + toReached = true; + } else { + digitRange(String.fromCharCode(fromStr.charCodeAt(i) + 1), String.fromCharCode(toStr.charCodeAt(i) - 1)); + } + out.push(" "); + moreDigits(subLen, subLen); + } + } + if (!toReached) { + out.push(" | "); + digitRange(toStr[i], toStr[i]); + out.push(" "); + uniformRange(subZeros, toSub); + } + out.push(")"); + } else { + out.push("["); + out.push(fromStr[i]); + out.push("-"); + out.push(toStr[i]); + out.push("]"); + } + } + } + + if (hasMin && hasMax) { + if (minValue < 0 && maxValue < 0) { + out.push("\"-\" ("); + _generateMinMaxInt(-maxValue, -minValue, out, decimalsLeft, true); + out.push(")"); + return; + } + + if (minValue < 0) { + out.push("\"-\" ("); + _generateMinMaxInt(0, -minValue, out, decimalsLeft, true); + out.push(") | "); + minValue = 0; + } + + let minS = minValue.toString(); + const maxS = maxValue.toString(); + const minDigits = minS.length; + const maxDigits = maxS.length; + + for (let digits = minDigits; digits < maxDigits; digits++) { + uniformRange(minS, "9".repeat(digits)); + minS = "1" + "0".repeat(digits); + out.push(" | "); + } + uniformRange(minS, maxS); + return; + } + + const lessDecimals = Math.max(decimalsLeft - 1, 1); + + if (hasMin) { + if (minValue < 0) { + out.push("\"-\" ("); + _generateMinMaxInt(null, -minValue, out, decimalsLeft, false); + out.push(") | [0] | [1-9] "); + moreDigits(0, decimalsLeft - 1); + } else if (minValue === 0) { + if (topLevel) { + out.push("[0] | [1-9] "); + moreDigits(0, lessDecimals); + } else { + moreDigits(1, decimalsLeft); + } + } else if (minValue <= 9) { + const c = minValue.toString(); + if (minValue > (topLevel ? 1 : 0)) { + digitRange("0", String.fromCharCode(c.charCodeAt(0) - 1)); + out.push(" "); + moreDigits(1, lessDecimals); + out.push(" | "); + } + digitRange(c, "9"); + out.push(" "); + moreDigits(0, lessDecimals); + } else { + const minS = minValue.toString(); + const length = minS.length; + const c = minS[0]; + + if (c > "1") { + digitRange(topLevel ? "1" : "0", String.fromCharCode(c.charCodeAt(0) - 1)); + out.push(" "); + moreDigits(length, lessDecimals); + out.push(" | "); + } + digitRange(c, c); + out.push(" ("); + _generateMinMaxInt(parseInt(minS.slice(1)), null, out, lessDecimals, false); + out.push(")"); + if (c < "9") { + out.push(" | "); + digitRange(String.fromCharCode(c.charCodeAt(0) + 1), "9"); + out.push(" "); + moreDigits(length - 1, lessDecimals); + } + } + return; + } + + if (hasMax) { + if (maxValue >= 0) { + if (topLevel) { + out.push("\"-\" [1-9] "); + moreDigits(0, lessDecimals); + out.push(" | "); + } + _generateMinMaxInt(0, maxValue, out, decimalsLeft, true); + } else { + out.push("\"-\" ("); + _generateMinMaxInt(-maxValue, null, out, decimalsLeft, false); + out.push(")"); + } + return; + } + + throw new Error("At least one of minValue or maxValue must be set"); +} + class BuiltinRule { constructor(content, deps) { this.content = content; @@ -435,6 +629,24 @@ export class SchemaConverter { const minLen = schema.minLength || 0; const maxLen = schema.maxLength; return this._addRule(ruleName, '"\\\"" ' + _buildRepetition(charRuleName, minLen, maxLen) + ' "\\\"" space'); + } else if (schemaType === 'integer' && ('minimum' in schema || 'exclusiveMinimum' in schema || 'maximum' in schema || 'exclusiveMaximum' in schema)) { + let minValue = null; + let maxValue = null; + if ('minimum' in schema) { + minValue = schema.minimum; + } else if ('exclusiveMinimum' in schema) { + minValue = schema.exclusiveMinimum + 1; + } + if ('maximum' in schema) { + maxValue = schema.maximum; + } else if ('exclusiveMaximum' in schema) { + maxValue = schema.exclusiveMaximum - 1; + } + + const out = ["("]; + _generateMinMaxInt(minValue, maxValue, out); + out.push(") space"); + return this._addRule(ruleName, out.join('')); } else if ((schemaType === 'object') || (Object.keys(schema).length === 0)) { return this._addRule(ruleName, this._addPrimitive('object', PRIMITIVE_RULES['object'])); } else {