#!/usr/bin/python # Copyright 2006 Lawrence Kesteloot # # Generates a vim help file which is a permuted index of all # options and ex commands. # # Usage: # # vim_permuted_index.py [vimdir] # # Where vimdir is where your vim installation files are. It # defaults to "/usr/share/vim/vim70". # # The output is to stdout. Put the output into your "doc" directory # and run ":helptags" on that directory. Then type ":help permuted-index" # to get the index. # import sys, re, os def get_ex_lines(doc): # Load lines for ex commands. lines = file(doc + "/index.txt").readlines() # Remove \n. lines = [line[:-1] for line in lines] # Find first line of ex reference. for i in range(len(lines)): if lines[i].find("*ex-cmd-index*") >= 0: del lines[:i] break else: sys.stderr.write("Did not find *ex-cmd-index*\n") sys.exit(-1) # Find first line with a command. for i in range(len(lines)): if len(lines[i]) > 0 and lines[i][0] == '|': del lines[:i] break else: sys.stderr.write("Did not find first ex command\n") sys.exit(-1) # Skip to next blank line. for i in range(len(lines)): if len(lines[i]) == 0: del lines[i:] break else: sys.stderr.write("Did not find end of ex commands\n") sys.exit(-1) # Unwrap lines. i = 0 while i < len(lines) - 1: next_line = lines[i + 1] if len(next_line) > 0 and \ (next_line[0] == ' ' or next_line[0] == '\t'): lines[i] = lines[i] + ' ' + next_line.strip() del lines[i + 1] else: i = i + 1 # Split at tabs. lines = [line.split(None, 2) for line in lines] # Remove second column. lines = [(line[0], line[2]) for line in lines] return lines def get_opt_lines(doc): # Load lines for options. lines = file(doc + "/quickref.txt").readlines() # Remove \n. lines = [line[:-1] for line in lines] # Find first line of options. for i in range(len(lines)): if lines[i].find("*option-list*") >= 0: del lines[:i + 1] break else: sys.stderr.write("Did not find *option-list*\n") sys.exit(-1) # Skip to end of section. for i in range(len(lines)): if lines[i].find("--------") >= 0: del lines[i:] break else: sys.stderr.write("Did not find end of options\n") sys.exit(-1) # Unwrap lines. Doesn't seem to happen with options. i = 0 while i < len(lines) - 1: next_line = lines[i + 1] if len(next_line) > 0 and \ (next_line[0] == ' ' or next_line[0] == '\t'): lines[i] = lines[i] + ' ' + next_line.strip() del lines[i + 1] else: i = i + 1 # Expand tabs, since tab usage is inconsistent. lines = [line.expandtabs() for line in lines] # Keep full opt and summary. This depends on specific layout. lines = [(line[:18].strip(), line[28:].strip()) for line in lines] return lines # Represents each line in the output. class Entry: def __init__(self, tag, summary, index): # The vim feature (ex command, option). self.tag = tag # The summary of its description. self.summary = summary # The location within the summary where we're sorting. self.index = index # Returns the part at and after the sorting word. def get_key(self): return self.summary[self.index:] # Compares by key, breaking ties with summary. Both # are case-insensitive. def __cmp__(self, rhs): c = cmp(self.get_key().lower(), rhs.get_key().lower()) if c != 0: return c return cmp(self.summary.lower(), rhs.summary.lower()) # Returns the sum of the tag word plus the index, basically # everything to the left of the sorted word. def get_prefix_length(self): return len(self.tag) + self.index # It's not useful to index these. def is_common_word(word): return word.lower() in ["a", "an", "the", "for", "to", "in", "of", "and", "go", "or", "but", "use", "like", "with", "that", "from", "be", "one", "as", "is", "when", "do", "how", "on", "not", "are"] def generate_permuted_index(lines): entries = [] words = {} for line in lines: # Find all words. matches = re.finditer(r'\b[a-zA-Z0-9]+', line[1]) for match in matches: word = match.group() if not is_common_word(word): entries.append(Entry(line[0], line[1], match.start())) # Count the number of times we see this. if word not in words: words[word] = 1 else: words[word] = words[word] + 1 # Sort entries by string at index. entries.sort() return entries, words # Display the most-used words. def display_words(words): # Turn the words into an array. words = words.items() # Sort by reverse count. words.sort(None, lambda word: word[1], True) # Only keep the top 40. del words[40:] # Print the rest for word in words: print word[0], word[1] # Find vim doc files. if len(sys.argv) > 1: vim = sys.argv[1] else: vim = "/usr/share/vim/vim70" # Quicky check. if not os.access(vim, os.F_OK): sys.stderr.write("No vim installation at \"" + vim + "\".\n") sys.exit(-1) doc = vim + "/doc" # Get the summary lines from the help files. try: ex = get_ex_lines(doc) opts = get_opt_lines(doc) except IOError: sys.stderr.write("Cannot open files in \"" + doc + "\".\n") sys.exit(-1) # Figure out all the permuted entries (and used words). entries, words = generate_permuted_index(ex + opts) # Get prefix length of each. prefix_length = [entry.get_prefix_length() for entry in entries] # Get max of that. max_prefix_length = max(prefix_length) # Print header. print "*permuted-index* (see http://www.teamten.com/lawrence/projects/vim_permuted_index/)" print # Print each entry. for entry in entries: spacing = max_prefix_length - entry.get_prefix_length() + 1 print entry.tag + (" " * spacing) + entry.summary # Print footer. We split the string so that vim doesn't interpret this # line when loading this source file. print print " vi" + "m:tw=78:fo=tcq2:isk=!-~,^*,^\|,^\":ts=8:ft=help:norl:nowrap" # For debugging is_common_word(): # display_words(words)