add tool to fix LAMMPS homepage URLs

tools/coding_standard/homepage.py

@@ -0,0 +1,131 @@
+#!/usr/bin/env python3
+# Utility for detecting incorrect LAMMPS homepage URLs
+#
+# Written by Richard Berger (Temple University)
+import os
+import sys
+import glob
+import re
+import yaml
+import argparse
+import shutil
+
+DEFAULT_CONFIG = """
+recursive: true
+include:
+    - cmake/**
+    - doc
+    - doc/src/**
+    - python
+    - src/**
+    - tools/coding_standard
+patterns:
+    - "*.c"
+    - "*.cmake"
+    - "*.cpp"
+    - "*.h"
+    - "*.md"
+    - "*.py"
+    - "*.rst"
+    - "*.sh"
+    - ".gitignore"
+    - "README"
+    - "requirements.txt"
+"""
+
+def check_homepage(f):
+    pattern = re.compile(r'.*lammps.sandia.gov.*')
+    lineno = 1
+    errors = set()
+
+    for line in f:
+        if pattern.match(line):
+            errors.add(lineno)
+        lineno += 1
+
+    return errors
+
+def check_file(path):
+    if path.find('homepage.py') >= 0: return { 'homepage_errors' : '' }
+    encoding = 'UTF-8'
+    whitespace_errors = set()
+    try:
+        with open(path, 'r') as f:
+            homepage_errors = check_homepage(f)
+    except UnicodeDecodeError:
+        encoding = 'ISO-8859-1'
+        try:
+            with open(path, 'r', encoding=encoding) as f:
+                homepage_errors = check_homepage(f)
+        except Exception:
+            encoding = 'unknown'
+
+    return {
+        'homepage_errors': homepage_errors,
+        'encoding': encoding
+    }
+
+def fix_file(path, check_result):
+    if path.find('homepage.py') >= 0: return
+    newfile = path + ".modified"
+    with open(newfile, 'w', encoding='UTF-8') as out:
+        with open(path, 'r', encoding=check_result['encoding']) as src:
+            for line in src:
+                newline = line.replace("lammps.sandia.gov/doc/","docs.lammps.org/")
+                newline = newline.replace("http://lammps.sandia.gov,","https://www.lammps.org/")
+                newline = newline.replace("lammps.sandia.gov","www.lammps.org")
+                print(newline, end='', file=out)
+    shutil.copymode(path, newfile)
+    shutil.move(newfile, path)
+
+def check_folder(directory, config, fix=False, verbose=False):
+    success = True
+    files = []
+
+    for base_path in config['include']:
+        for pattern in config['patterns']:
+            path = os.path.join(directory, base_path, pattern)
+            files += glob.glob(path, recursive=config['recursive'])
+
+    for f in files:
+        path = os.path.normpath(f)
+
+        if verbose:
+            print("Checking file:", path)
+
+        result = check_file(path)
+
+        has_resolvable_errors = False
+
+        for lineno in result['homepage_errors']:
+            print("[Error] Incorrect LAMMPS homepage @ {}:{}".format(path, lineno))
+            has_resolvable_errors = True
+
+        if has_resolvable_errors:
+            if fix:
+                print("Applying automatic fixes to file:", path)
+                fix_file(path, result)
+            else:
+                success = False
+
+    return success
+
+def main():
+    parser = argparse.ArgumentParser(description='Utility for detecting and fixing whitespace issues in LAMMPS')
+    parser.add_argument('-c', '--config', metavar='CONFIG_FILE', help='location of a optional configuration file')
+    parser.add_argument('-f', '--fix', action='store_true', help='automatically fix URLs')
+    parser.add_argument('-v', '--verbose', action='store_true', help='verbose output')
+    parser.add_argument('DIRECTORY', help='directory that should be checked')
+    args = parser.parse_args()
+
+    if args.config:
+        with open(args.config, 'r') as cfile:
+            config = yaml.load(cfile, Loader=yaml.FullLoader)
+    else:
+        config = yaml.load(DEFAULT_CONFIG, Loader=yaml.FullLoader)
+
+    if not check_folder(args.DIRECTORY, config, args.fix, args.verbose):
+        sys.exit(1)
+
+if __name__ == "__main__":
+    main()