Tools/idle/FormatParagraph.py

   1 # Extension to format a paragraph
   2
   3 # Does basic, standard text formatting, and also understands Python
   4 # comment blocks.  Thus, for editing Python source code, this
   5 # extension is really only suitable for reformatting these comment
   6 # blocks or triple-quoted strings.
   7
   8 # Known problems with comment reformatting:
   9 # * If there is a selection marked, and the first line of the
  10 #   selection is not complete, the block will probably not be detected
  11 #   as comments, and will have the normal "text formatting" rules
  12 #   applied.
  13 # * If a comment block has leading whitespace that mixes tabs and
  14 #   spaces, they will not be considered part of the same block.
  15 # * Fancy comments, like this bulleted list, arent handled :-)
  16
  17 import string
  18 import re
  19
  20 class FormatParagraph:
  21
  22     menudefs = [
  23         ('edit', [
  24             ('Format Paragraph', '<<format-paragraph>>'),
  25          ])
  26     ]
  27
  28     keydefs = {
  29         '<<format-paragraph>>': ['<Alt-q>'],
  30     }
  31
  32     unix_keydefs = {
  33         '<<format-paragraph>>': ['<Meta-q>'],
  34     }
  35
  36     def __init__(self, editwin):
  37         self.editwin = editwin
  38
  39     def close(self):
  40         self.editwin = None
  41
  42     def format_paragraph_event(self, event):
  43         text = self.editwin.text
  44         first, last = self.editwin.get_selection_indices()
  45         if first and last:
  46             data = text.get(first, last)
  47             comment_header = ''
  48         else:
  49             first, last, comment_header, data = \
  50                     find_paragraph(text, text.index("insert"))
  51         if comment_header:
  52             # Reformat the comment lines - convert to text sans header.
  53             lines = string.split(data, "\n")
  54             lines = map(lambda st, l=len(comment_header): st[l:], lines)
  55             data = string.join(lines, "\n")
  56             # Reformat to 70 chars or a 20 char width, whichever is greater.
  57             format_width = max(70-len(comment_header), 20)
  58             newdata = reformat_paragraph(data, format_width)
  59             # re-split and re-insert the comment header.
  60             newdata = string.split(newdata, "\n")
  61             # If the block ends in a \n, we dont want the comment
  62             # prefix inserted after it. (Im not sure it makes sense to
  63             # reformat a comment block that isnt made of complete
  64             # lines, but whatever!)  Can't think of a clean soltution,
  65             # so we hack away
  66             block_suffix = ""
  67             if not newdata[-1]:
  68                 block_suffix = "\n"
  69                 newdata = newdata[:-1]
  70             builder = lambda item, prefix=comment_header: prefix+item
  71             newdata = string.join(map(builder, newdata), '\n') + block_suffix
  72         else:
  73             # Just a normal text format
  74             newdata = reformat_paragraph(data)
  75         text.tag_remove("sel", "1.0", "end")
  76         if newdata != data:
  77             text.mark_set("insert", first)
  78             text.undo_block_start()
  79             text.delete(first, last)
  80             text.insert(first, newdata)
  81             text.undo_block_stop()
  82         else:
  83             text.mark_set("insert", last)
  84         text.see("insert")
  85
  86 def find_paragraph(text, mark):
  87     lineno, col = map(int, string.split(mark, "."))
  88     line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
  89     while text.compare("%d.0" % lineno, "<", "end") and is_all_white(line):
  90         lineno = lineno + 1
  91         line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
  92     first_lineno = lineno
  93     comment_header = get_comment_header(line)
  94     comment_header_len = len(comment_header)
  95     while get_comment_header(line)==comment_header and \
  96               not is_all_white(line[comment_header_len:]):
  97         lineno = lineno + 1
  98         line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
  99     last = "%d.0" % lineno
 100     # Search back to beginning of paragraph
 101     lineno = first_lineno - 1
 102     line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
 103     while lineno > 0 and \
 104               get_comment_header(line)==comment_header and \
 105               not is_all_white(line[comment_header_len:]):
 106         lineno = lineno - 1
 107         line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
 108     first = "%d.0" % (lineno+1)
 109     return first, last, comment_header, text.get(first, last)
 110
 111 def reformat_paragraph(data, limit=70):
 112     lines = string.split(data, "\n")
 113     i = 0
 114     n = len(lines)
 115     while i < n and is_all_white(lines[i]):
 116         i = i+1
 117     if i >= n:
 118         return data
 119     indent1 = get_indent(lines[i])
 120     if i+1 < n and not is_all_white(lines[i+1]):
 121         indent2 = get_indent(lines[i+1])
 122     else:
 123         indent2 = indent1
 124     new = lines[:i]
 125     partial = indent1
 126     while i < n and not is_all_white(lines[i]):
 127         # XXX Should take double space after period (etc.) into account
 128         words = re.split("(\s+)", lines[i])
 129         for j in range(0, len(words), 2):
 130             word = words[j]
 131             if not word:
 132                 continue # Can happen when line ends in whitespace
 133             if len(string.expandtabs(partial + word)) > limit and \
 134                partial != indent1:
 135                 new.append(string.rstrip(partial))
 136                 partial = indent2
 137             partial = partial + word + " "
 138             if j+1 < len(words) and words[j+1] != " ":
 139                 partial = partial + " "
 140         i = i+1
 141     new.append(string.rstrip(partial))
 142     # XXX Should reformat remaining paragraphs as well
 143     new.extend(lines[i:])
 144     return string.join(new, "\n")
 145
 146 def is_all_white(line):
 147     return re.match(r"^\s*$", line) is not None
 148
 149 def get_indent(line):
 150     return re.match(r"^(\s*)", line).group()
 151
 152 def get_comment_header(line):
 153     m = re.match(r"^(\s*#*)", line)
 154     if m is None: return ""
 155     return m.group(1)