rename_yaml_files.py

#!/usr/bin/env python3

## Usage: Run from directory `sv-benchmarks/c`.
## The script will scan all subdirectories for yml-Files and rename them if they contain any
## property or are duplicates.

import glob
import os

NAME_TO_PROP_AND_SUBPROP = {
    'unreach-call': ('properties/unreach-call.prp', None),
    'termination': ('properties/termination.prp', None),
    'no-overflow': ('properties/no-overflow.prp', None),
    'valid-memcleanup': ('properties/valid-memcleanup.prp', None),
    'valid-memsafety': ('properties/valid-memsafety.prp', None),
    'valid-deref': ('properties/valid-memsafety.prp', 'valid-deref'),
    'valid-free': ('properties/valid-memsafety.prp', 'valid-free'),
    'valid-memtrack': ('properties/valid-memsafety.prp', 'valid-memtrack'),
    'def-behavior': ('properties/def-behavior.prp', None)
}

all_tasks = glob.glob('**/*.yml', recursive=True)

tasks_to_new_names_and_yml = dict()
for task_file in all_tasks:
    properties = list()
    name_pcs_dot = task_file.split('.')
    new_name_pcs_dot = list()
    for pd in name_pcs_dot:
        name_pcs = pd.split('_')
        new_name_pcs = list()
        for p in name_pcs:
            offset = 0
            for name, prop in NAME_TO_PROP_AND_SUBPROP.items():
                if name not in p:
                    continue  # with next name_pc p
                if p.startswith('true'):
                    expected = 'true'
                    offset += len('true-')
                elif p.startswith('false'):
                    expected = 'false'
                    offset += len('false-')
                elif p.startswith('unknown-'):
                    expected = None
                    offset += len('unknown-')
                else:
                    continue  # with next name_pc p
                properties.append((prop, expected))
                offset += len(name)
                break  # for-loop over properties once one matches, because they are distinct
            new_p = p[offset:]
            if new_p:
                new_name_pcs.append(new_p)
        new_name_pcs_dot.append("_".join(new_name_pcs))

    yml_info = (task_file, properties)

    new_task_file = '.'.join(new_name_pcs_dot)
    if new_task_file.endswith('.c.yml'):
        suffix_offset = 5
    elif new_task_file.endswith('.c.i.yml'):
        suffix_offset = 7
    else:
        suffix_offset = 3
    new_task_file = new_task_file[:-suffix_offset] + "yml"
    tasks_to_new_names_and_yml[task_file] = (new_task_file, yml_info)

for old_name, (curr_task, yml_info) in tasks_to_new_names_and_yml.items():
    task_basename = os.path.basename(curr_task)
    collisions = [k
            for k, v in tasks_to_new_names_and_yml.items()
            if os.path.basename(v[0]).lower() == task_basename.lower()
            and k != old_name]
    if collisions:
        curr_task = curr_task[:-4] + '-1' + curr_task[-4:]
        for idx, other in enumerate(collisions, 2):
            new_name, content = tasks_to_new_names_and_yml.pop(other)
            new_name = new_name[:-4] + "-" + str(idx) + new_name[-4:]
            tasks_to_new_names_and_yml[other] = (new_name, content)

    task_basename = os.path.basename(curr_task)
    task_dir = os.path.dirname(curr_task)

    if old_name != curr_task:
        os.rename(old_name, curr_task)

        if os.path.exists(old_name[:-3] + 'i'):
            old_c = old_name[:-3] + 'i'
            curr_task_name = curr_task.split('/')[-1]
            new_c_basename = curr_task_name[:-3] + 'i'
            new_c_name = os.path.dirname(old_c) + '/' + new_c_basename
            os.rename(old_c, new_c_name)

            with open(curr_task) as inp:
                content = inp.readlines()
            new_content = list()
            old_c_basename = os.path.basename(old_c)
            for line in content:
                if 'input_file' in line:
                    line = line.replace(old_c_basename, new_c_basename)
                new_content.append(line)
            with open(curr_task, 'w+') as outp:
                outp.writelines(new_content)

        if os.path.exists(old_name[:-3] + 'c'):
            old_c = old_name[:-3] + 'c'
        elif os.path.exists(old_name[:-4]):
            old_c = old_name[:-4]
        # ldv-memsafety/memleaks*.i -> ldv-memsafety/memleaks-notpreprocessed/memleaks*.c
        elif old_name.startswith('ldv-memsafety/memleaks'):
            old_c = 'ldv-memsafety/memleaks-notpreprocessed/'\
                    + old_name.split('/')[-1][:-3] + 'c'
        else:
            old_c = None
        if old_c:
            assert old_c not in all_tasks
            curr_task_name = curr_task.split('/')[-1]
            new_c_basename = curr_task_name[:-3] + 'c'
            new_c_name = os.path.dirname(old_c) + '/' + new_c_basename
            with open(curr_task) as inp:
                content = inp.readlines()
            new_content = list()
            old_c_basename = os.path.basename(old_c)
            for line in content:
                if 'input_file' in line:
                    line = line.replace(old_c_basename, new_c_basename)
                new_content.append(line)
            with open(curr_task, 'w+') as outp:
                outp.writelines(new_content)

            os.rename(old_c, new_c_name)