|
| 1 | +package org.apache.sysds.hops.rewriter; |
| 2 | + |
| 3 | +import org.apache.commons.lang3.NotImplementedException; |
| 4 | + |
| 5 | +import java.util.HashMap; |
| 6 | +import java.util.HashSet; |
| 7 | +import java.util.Map; |
| 8 | +import java.util.Set; |
| 9 | +import java.util.function.BiFunction; |
| 10 | +import java.util.function.Function; |
| 11 | + |
| 12 | +public class DMLCodeGenerator { |
| 13 | + public static final double EPS = 1e-10; |
| 14 | + |
| 15 | + |
| 16 | + private static final HashSet<String> printAsBinary = new HashSet<>(); |
| 17 | + private static final HashMap<String, BiFunction<RewriterStatement, StringBuilder, Boolean>> customEncoders = new HashMap<>(); |
| 18 | + private static final RuleContext ctx = RewriterUtils.buildDefaultContext(); |
| 19 | + |
| 20 | + static { |
| 21 | + printAsBinary.add("+"); |
| 22 | + printAsBinary.add("-"); |
| 23 | + printAsBinary.add("*"); |
| 24 | + printAsBinary.add("/"); |
| 25 | + printAsBinary.add("^"); |
| 26 | + printAsBinary.add("=="); |
| 27 | + printAsBinary.add("!="); |
| 28 | + printAsBinary.add(">"); |
| 29 | + printAsBinary.add(">="); |
| 30 | + printAsBinary.add("<"); |
| 31 | + printAsBinary.add("<="); |
| 32 | + |
| 33 | + customEncoders.put("[]", (stmt, sb) -> { |
| 34 | + if (stmt.getOperands().size() == 3) { |
| 35 | + sb.append('('); |
| 36 | + appendExpression(stmt.getChild(0), sb); |
| 37 | + sb.append(")["); |
| 38 | + appendExpression(stmt.getChild(1), sb); |
| 39 | + sb.append(", "); |
| 40 | + appendExpression(stmt.getChild(2), sb); |
| 41 | + sb.append(']'); |
| 42 | + return true; |
| 43 | + } else if (stmt.getOperands().size() == 5) { |
| 44 | + sb.append('('); |
| 45 | + appendExpression(stmt.getChild(0), sb); |
| 46 | + sb.append(")["); |
| 47 | + appendExpression(stmt.getChild(1), sb); |
| 48 | + sb.append(" : "); |
| 49 | + appendExpression(stmt.getChild(2), sb); |
| 50 | + sb.append(", "); |
| 51 | + appendExpression(stmt.getChild(3), sb); |
| 52 | + sb.append(" : "); |
| 53 | + appendExpression(stmt.getChild(4), sb); |
| 54 | + sb.append(']'); |
| 55 | + return true; |
| 56 | + } |
| 57 | + |
| 58 | + return false; |
| 59 | + }); |
| 60 | + } |
| 61 | + |
| 62 | + public static String generateRuleValidationDML(RewriterRule rule, double eps, String sessionId) { |
| 63 | + RewriterStatement stmtFrom = rule.getStmt1(); |
| 64 | + RewriterStatement stmtTo = rule.getStmt2(); |
| 65 | + |
| 66 | + Set<RewriterStatement> vars = new HashSet<>(); |
| 67 | + |
| 68 | + stmtFrom.forEachPostOrder((stmt, pred) -> { |
| 69 | + if (!stmt.isInstruction() && !stmt.isLiteral()) |
| 70 | + vars.add(stmt); |
| 71 | + }, false); |
| 72 | + |
| 73 | + stmtTo.forEachPostOrder((stmt, pred) -> { |
| 74 | + if (!stmt.isInstruction() && !stmt.isLiteral()) |
| 75 | + vars.add(stmt); |
| 76 | + }, false); |
| 77 | + |
| 78 | + StringBuilder sb = new StringBuilder(); |
| 79 | + |
| 80 | + for (RewriterStatement var : vars) { |
| 81 | + switch (var.getResultingDataType(ctx)) { |
| 82 | + case "MATRIX": |
| 83 | + sb.append(var.getId() + " = rand(rows=1000, cols=1000, min=0.0, max=1.0)\n"); |
| 84 | + break; |
| 85 | + case "FLOAT": |
| 86 | + sb.append(var.getId() + " = as.scalar(rand())\n"); |
| 87 | + break; |
| 88 | + case "INT": |
| 89 | + sb.append(var.getId() + " = as.integer(as.scalar(rand(min=0.0, max=10000.0)))\n"); |
| 90 | + break; |
| 91 | + case "BOOL": |
| 92 | + sb.append(var.getId() + " = as.scalar(rand()) < 0.5\n"); |
| 93 | + break; |
| 94 | + default: |
| 95 | + throw new NotImplementedException(var.getResultingDataType(ctx)); |
| 96 | + } |
| 97 | + } |
| 98 | + |
| 99 | + sb.append('\n'); |
| 100 | + sb.append("R1 = "); |
| 101 | + sb.append(generateDML(stmtFrom)); |
| 102 | + sb.append('\n'); |
| 103 | + sb.append("R2 = "); |
| 104 | + sb.append(generateDML(stmtTo)); |
| 105 | + sb.append('\n'); |
| 106 | + sb.append("print(\""); |
| 107 | + sb.append(sessionId); |
| 108 | + sb.append(" valid: \" + ("); |
| 109 | + sb.append(generateEqualityCheck("R1", "R2", stmtFrom.getResultingDataType(ctx), eps)); |
| 110 | + sb.append("))"); |
| 111 | + |
| 112 | + return sb.toString(); |
| 113 | + } |
| 114 | + |
| 115 | + public static String generateEqualityCheck(String stmt1Var, String stmt2Var, String dataType, double eps) { |
| 116 | + switch (dataType) { |
| 117 | + case "MATRIX": |
| 118 | + return "sum(abs(" + stmt1Var + " - " + stmt2Var + ") < " + eps + ") == length(" + stmt1Var + ")"; |
| 119 | + case "INT": |
| 120 | + case "BOOL": |
| 121 | + return stmt1Var + " == " + stmt2Var; |
| 122 | + case "FLOAT": |
| 123 | + return "abs(" + stmt1Var + " - " + stmt2Var + ") < " + eps; |
| 124 | + } |
| 125 | + |
| 126 | + throw new NotImplementedException(); |
| 127 | + } |
| 128 | + |
| 129 | + public static String generateDMLDefs(Map<String, RewriterStatement> defs) { |
| 130 | + StringBuilder sb = new StringBuilder(); |
| 131 | + |
| 132 | + defs.forEach((k, v) -> { |
| 133 | + sb.append(k); |
| 134 | + sb.append(" = "); |
| 135 | + sb.append(generateDML(v)); |
| 136 | + sb.append('\n'); |
| 137 | + }); |
| 138 | + |
| 139 | + return sb.toString(); |
| 140 | + } |
| 141 | + |
| 142 | + public static String generateDML(RewriterStatement root) { |
| 143 | + StringBuilder sb = new StringBuilder(); |
| 144 | + appendExpression(root, sb); |
| 145 | + |
| 146 | + return sb.toString(); |
| 147 | + } |
| 148 | + |
| 149 | + private static void appendExpression(RewriterStatement cur, StringBuilder sb) { |
| 150 | + if (cur.isInstruction()) { |
| 151 | + resolveExpression((RewriterInstruction) cur, sb); |
| 152 | + } else { |
| 153 | + if (cur.isLiteral()) |
| 154 | + sb.append(cur.getLiteral()); |
| 155 | + else |
| 156 | + sb.append(cur.getId()); |
| 157 | + } |
| 158 | + } |
| 159 | + |
| 160 | + private static void resolveExpression(RewriterInstruction expr, StringBuilder sb) { |
| 161 | + String typedInstr = expr.trueTypedInstruction(ctx); |
| 162 | + String unTypedInstr = expr.trueInstruction(); |
| 163 | + |
| 164 | + if (expr.getOperands().size() == 2 && (printAsBinary.contains(typedInstr) || printAsBinary.contains(unTypedInstr))) { |
| 165 | + sb.append('('); |
| 166 | + appendExpression(expr.getChild(0), sb); |
| 167 | + sb.append(") "); |
| 168 | + sb.append(unTypedInstr); |
| 169 | + sb.append(" ("); |
| 170 | + appendExpression(expr.getChild(1), sb); |
| 171 | + sb.append(')'); |
| 172 | + return; |
| 173 | + } |
| 174 | + |
| 175 | + BiFunction<RewriterStatement, StringBuilder, Boolean> customEncoder = customEncoders.get(typedInstr); |
| 176 | + |
| 177 | + if (customEncoder == null) |
| 178 | + customEncoder = customEncoders.get(unTypedInstr); |
| 179 | + |
| 180 | + if (customEncoder == null) { |
| 181 | + sb.append(unTypedInstr); |
| 182 | + sb.append('('); |
| 183 | + |
| 184 | + for (int i = 0; i < expr.getOperands().size(); i++) { |
| 185 | + if (i != 0) |
| 186 | + sb.append(", "); |
| 187 | + |
| 188 | + appendExpression(expr.getChild(i), sb); |
| 189 | + } |
| 190 | + |
| 191 | + sb.append(')'); |
| 192 | + } else { |
| 193 | + customEncoder.apply(expr, sb); |
| 194 | + } |
| 195 | + } |
| 196 | +} |
0 commit comments