mlir/utils/generate-test-checks.py

   1 #!/usr/bin/env python3
   2 """A script to generate FileCheck statements for mlir unit tests.
   3
   4 This script is a utility to add FileCheck patterns to an mlir file.
   5
   6 NOTE: The input .mlir is expected to be the output from the parser, not a
   7 stripped down variant.
   8
   9 Example usage:
  10 $ generate-test-checks.py foo.mlir
  11 $ mlir-opt foo.mlir -transformation | generate-test-checks.py
  12 $ mlir-opt foo.mlir -transformation | generate-test-checks.py --source foo.mlir
  13 $ mlir-opt foo.mlir -transformation | generate-test-checks.py --source foo.mlir -i
  14 $ mlir-opt foo.mlir -transformation | generate-test-checks.py --source foo.mlir -i --source_delim_regex='gpu.func @'
  15
  16 The script will heuristically generate CHECK/CHECK-LABEL commands for each line
  17 within the file. By default this script will also try to insert string
  18 substitution blocks for all SSA value names. If --source file is specified, the
  19 script will attempt to insert the generated CHECKs to the source file by looking
  20 for line positions matched by --source_delim_regex.
  21
  22 The script is designed to make adding checks to a test case fast, it is *not*
  23 designed to be authoritative about what constitutes a good test!
  24 """
  25
  26 # Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
  27 # See https://llvm.org/LICENSE.txt for license information.
  28 # SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
  29
  30 import argparse
  31 import os  # Used to advertise this file's name ("autogenerated_note").
  32 import re
  33 import sys
  34
  35 ADVERT = '// NOTE: Assertions have been autogenerated by '
  36
  37 # Regex command to match an SSA identifier.
  38 SSA_RE_STR = '[0-9]+|[a-zA-Z$._-][a-zA-Z0-9$._-]*'
  39 SSA_RE = re.compile(SSA_RE_STR)
  40
  41
  42 # Class used to generate and manage string substitution blocks for SSA value
  43 # names.
  44 class SSAVariableNamer:
  45
  46   def __init__(self):
  47     self.scopes = []
  48     self.name_counter = 0
  49
  50   # Generate a substitution name for the given ssa value name.
  51   def generate_name(self, ssa_name):
  52     variable = 'VAL_' + str(self.name_counter)
  53     self.name_counter += 1
  54     self.scopes[-1][ssa_name] = variable
  55     return variable
  56
  57   # Push a new variable name scope.
  58   def push_name_scope(self):
  59     self.scopes.append({})
  60
  61   # Pop the last variable name scope.
  62   def pop_name_scope(self):
  63     self.scopes.pop()
  64
  65   # Return the level of nesting (number of pushed scopes).
  66   def num_scopes(self):
  67     return len(self.scopes)
  68
  69   # Reset the counter.
  70   def clear_counter(self):
  71     self.name_counter = 0
  72
  73
  74 # Process a line of input that has been split at each SSA identifier '%'.
  75 def process_line(line_chunks, variable_namer):
  76   output_line = ''
  77
  78   # Process the rest that contained an SSA value name.
  79   for chunk in line_chunks:
  80     m = SSA_RE.match(chunk)
  81     ssa_name = m.group(0)
  82
  83     # Check if an existing variable exists for this name.
  84     variable = None
  85     for scope in variable_namer.scopes:
  86       variable = scope.get(ssa_name)
  87       if variable is not None:
  88         break
  89
  90     # If one exists, then output the existing name.
  91     if variable is not None:
  92       output_line += '%[[' + variable + ']]'
  93     else:
  94       # Otherwise, generate a new variable.
  95       variable = variable_namer.generate_name(ssa_name)
  96       output_line += '%[[' + variable + ':.*]]'
  97
  98     # Append the non named group.
  99     output_line += chunk[len(ssa_name):]
 100
 101   return output_line.rstrip() + '\n'
 102
 103
 104 # Process the source file lines. The source file doesn't have to be .mlir.
 105 def process_source_lines(source_lines, note, args):
 106   source_split_re = re.compile(args.source_delim_regex)
 107
 108   source_segments = [[]]
 109   for line in source_lines:
 110     # Remove previous note.
 111     if line == note:
 112       continue
 113     # Remove previous CHECK lines.
 114     if line.find(args.check_prefix) != -1:
 115       continue
 116     # Segment the file based on --source_delim_regex.
 117     if source_split_re.search(line):
 118       source_segments.append([])
 119
 120     source_segments[-1].append(line + '\n')
 121   return source_segments
 122
 123
 124 # Pre-process a line of input to remove any character sequences that will be
 125 # problematic with FileCheck.
 126 def preprocess_line(line):
 127   # Replace any double brackets, '[[' with escaped replacements. '[['
 128   # corresponds to variable names in FileCheck.
 129   output_line = line.replace('[[', '{{\\[\\[}}')
 130
 131   # Replace any single brackets that are followed by an SSA identifier, the
 132   # identifier will be replace by a variable; Creating the same situation as
 133   # above.
 134   output_line = output_line.replace('[%', '{{\\[}}%')
 135
 136   return output_line
 137
 138
 139 def main():
 140   parser = argparse.ArgumentParser(
 141       description=__doc__, formatter_class=argparse.RawTextHelpFormatter)
 142   parser.add_argument(
 143       '--check-prefix', default='CHECK', help='Prefix to use from check file.')
 144   parser.add_argument(
 145       '-o',
 146       '--output',
 147       nargs='?',
 148       type=argparse.FileType('w'),
 149       default=None)
 150   parser.add_argument(
 151       'input',
 152       nargs='?',
 153       type=argparse.FileType('r'),
 154       default=sys.stdin)
 155   parser.add_argument(
 156       '--source', type=str,
 157       help='Print each CHECK chunk before each delimeter line in the source'
 158            'file, respectively. The delimeter lines are identified by '
 159            '--source_delim_regex.')
 160   parser.add_argument('--source_delim_regex', type=str, default='func @')
 161   parser.add_argument(
 162       '--starts_from_scope', type=int, default=1,
 163       help='Omit the top specified level of content. For example, by default '
 164            'it omits "module {"')
 165   parser.add_argument('-i', '--inplace', action='store_true', default=False)
 166
 167   args = parser.parse_args()
 168
 169   # Open the given input file.
 170   input_lines = [l.rstrip() for l in args.input]
 171   args.input.close()
 172
 173   # Generate a note used for the generated check file.
 174   script_name = os.path.basename(__file__)
 175   autogenerated_note = (ADVERT + 'utils/' + script_name)
 176
 177   source_segments = None
 178   if args.source:
 179     source_segments = process_source_lines(
 180         [l.rstrip() for l in open(args.source, 'r')],
 181         autogenerated_note,
 182         args
 183     )
 184
 185   if args.inplace:
 186     assert args.output is None
 187     output = open(args.source, 'w')
 188   elif args.output is None:
 189     output = sys.stdout
 190   else:
 191     output = args.output
 192
 193   output_segments = [[]]
 194   # A map containing data used for naming SSA value names.
 195   variable_namer = SSAVariableNamer()
 196   for input_line in input_lines:
 197     if not input_line:
 198       continue
 199     lstripped_input_line = input_line.lstrip()
 200
 201     # Lines with blocks begin with a ^. These lines have a trailing comment
 202     # that needs to be stripped.
 203     is_block = lstripped_input_line[0] == '^'
 204     if is_block:
 205       input_line = input_line.rsplit('//', 1)[0].rstrip()
 206
 207     cur_level = variable_namer.num_scopes()
 208
 209     # If the line starts with a '}', pop the last name scope.
 210     if lstripped_input_line[0] == '}':
 211       variable_namer.pop_name_scope()
 212       cur_level = variable_namer.num_scopes()
 213
 214     # If the line ends with a '{', push a new name scope.
 215     if input_line[-1] == '{':
 216       variable_namer.push_name_scope()
 217       if cur_level == args.starts_from_scope:
 218         output_segments.append([])
 219
 220     # Omit lines at the near top level e.g. "module {".
 221     if cur_level < args.starts_from_scope:
 222       continue
 223
 224     if len(output_segments[-1]) == 0:
 225       variable_namer.clear_counter()
 226
 227     # Preprocess the input to remove any sequences that may be problematic with
 228     # FileCheck.
 229     input_line = preprocess_line(input_line)
 230
 231     # Split the line at the each SSA value name.
 232     ssa_split = input_line.split('%')
 233
 234     # If this is a top-level operation use 'CHECK-LABEL', otherwise 'CHECK:'.
 235     if len(output_segments[-1]) != 0 or not ssa_split[0]:
 236       output_line = '// ' + args.check_prefix + ': '
 237       # Pad to align with the 'LABEL' statements.
 238       output_line += (' ' * len('-LABEL'))
 239
 240       # Output the first line chunk that does not contain an SSA name.
 241       output_line += ssa_split[0]
 242
 243       # Process the rest of the input line.
 244       output_line += process_line(ssa_split[1:], variable_namer)
 245
 246     else:
 247       # Output the first line chunk that does not contain an SSA name for the
 248       # label.
 249       output_line = '// ' + args.check_prefix + '-LABEL: ' + ssa_split[0] + '\n'
 250
 251       # Process the rest of the input line on separate check lines.
 252       for argument in ssa_split[1:]:
 253         output_line += '// ' + args.check_prefix + '-SAME:  '
 254
 255         # Pad to align with the original position in the line.
 256         output_line += ' ' * len(ssa_split[0])
 257
 258         # Process the rest of the line.
 259         output_line += process_line([argument], variable_namer)
 260
 261     # Append the output line.
 262     output_segments[-1].append(output_line)
 263
 264   output.write(autogenerated_note + '\n')
 265
 266   # Write the output.
 267   if source_segments:
 268     assert len(output_segments) == len(source_segments)
 269     for check_segment, source_segment in zip(output_segments, source_segments):
 270       for line in check_segment:
 271         output.write(line)
 272       for line in source_segment:
 273         output.write(line)
 274   else:
 275     for segment in output_segments:
 276       output.write('\n')
 277       for output_line in segment:
 278         output.write(output_line)
 279     output.write('\n')
 280   output.close()
 281
 282
 283 if __name__ == '__main__':
 284   main()