Lib/idlelib/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 re
  18 from idlelib.configHandler import idleConf
  19
  20 class FormatParagraph:
  21
  22     menudefs = [
  23         ('format', [   # /s/edit/format   dscherer@cmu.edu
  24             ('Format Paragraph', '<<format-paragraph>>'),
  25          ])
  26     ]
  27
  28     def __init__(self, editwin):
  29         self.editwin = editwin
  30
  31     def close(self):
  32         self.editwin = None
  33
  34     def format_paragraph_event(self, event):
  35         maxformatwidth = int(idleConf.GetOption('main','FormatParagraph','paragraph'))
  36         text = self.editwin.text
  37         first, last = self.editwin.get_selection_indices()
  38         if first and last:
  39             data = text.get(first, last)
  40             comment_header = ''
  41         else:
  42             first, last, comment_header, data = \
  43                     find_paragraph(text, text.index("insert"))
  44         if comment_header:
  45             # Reformat the comment lines - convert to text sans header.
  46             lines = data.split("\n")
  47             lines = map(lambda st, l=len(comment_header): st[l:], lines)
  48             data = "\n".join(lines)
  49             # Reformat to maxformatwidth chars or a 20 char width, whichever is greater.
  50             format_width = max(maxformatwidth - len(comment_header), 20)
  51             newdata = reformat_paragraph(data, format_width)
  52             # re-split and re-insert the comment header.
  53             newdata = newdata.split("\n")
  54             # If the block ends in a \n, we dont want the comment
  55             # prefix inserted after it. (Im not sure it makes sense to
  56             # reformat a comment block that isnt made of complete
  57             # lines, but whatever!)  Can't think of a clean soltution,
  58             # so we hack away
  59             block_suffix = ""
  60             if not newdata[-1]:
  61                 block_suffix = "\n"
  62                 newdata = newdata[:-1]
  63             builder = lambda item, prefix=comment_header: prefix+item
  64             newdata = '\n'.join(map(builder, newdata)) + block_suffix
  65         else:
  66             # Just a normal text format
  67             newdata = reformat_paragraph(data, maxformatwidth)
  68         text.tag_remove("sel", "1.0", "end")
  69         if newdata != data:
  70             text.mark_set("insert", first)
  71             text.undo_block_start()
  72             text.delete(first, last)
  73             text.insert(first, newdata)
  74             text.undo_block_stop()
  75         else:
  76             text.mark_set("insert", last)
  77         text.see("insert")
  78         return "break"
  79
  80 def find_paragraph(text, mark):
  81     lineno, col = map(int, mark.split("."))
  82     line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
  83     while text.compare("%d.0" % lineno, "<", "end") and is_all_white(line):
  84         lineno = lineno + 1
  85         line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
  86     first_lineno = lineno
  87     comment_header = get_comment_header(line)
  88     comment_header_len = len(comment_header)
  89     while get_comment_header(line)==comment_header and \
  90               not is_all_white(line[comment_header_len:]):
  91         lineno = lineno + 1
  92         line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
  93     last = "%d.0" % lineno
  94     # Search back to beginning of paragraph
  95     lineno = first_lineno - 1
  96     line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
  97     while lineno > 0 and \
  98               get_comment_header(line)==comment_header and \
  99               not is_all_white(line[comment_header_len:]):
 100         lineno = lineno - 1
 101         line = text.get("%d.0" % lineno, "%d.0 lineend" % lineno)
 102     first = "%d.0" % (lineno+1)
 103     return first, last, comment_header, text.get(first, last)
 104
 105 def reformat_paragraph(data, limit):
 106     lines = data.split("\n")
 107     i = 0
 108     n = len(lines)
 109     while i < n and is_all_white(lines[i]):
 110         i = i+1
 111     if i >= n:
 112         return data
 113     indent1 = get_indent(lines[i])
 114     if i+1 < n and not is_all_white(lines[i+1]):
 115         indent2 = get_indent(lines[i+1])
 116     else:
 117         indent2 = indent1
 118     new = lines[:i]
 119     partial = indent1
 120     while i < n and not is_all_white(lines[i]):
 121         # XXX Should take double space after period (etc.) into account
 122         words = re.split("(\s+)", lines[i])
 123         for j in range(0, len(words), 2):
 124             word = words[j]
 125             if not word:
 126                 continue # Can happen when line ends in whitespace
 127             if len((partial + word).expandtabs()) > limit and \
 128                partial != indent1:
 129                 new.append(partial.rstrip())
 130                 partial = indent2
 131             partial = partial + word + " "
 132             if j+1 < len(words) and words[j+1] != " ":
 133                 partial = partial + " "
 134         i = i+1
 135     new.append(partial.rstrip())
 136     # XXX Should reformat remaining paragraphs as well
 137     new.extend(lines[i:])
 138     return "\n".join(new)
 139
 140 def is_all_white(line):
 141     return re.match(r"^\s*$", line) is not None
 142
 143 def get_indent(line):
 144     return re.match(r"^(\s*)", line).group()
 145
 146 def get_comment_header(line):
 147     m = re.match(r"^(\s*#*)", line)
 148     if m is None: return ""
 149     return m.group(1)