"""
MoinMoin - SearchInPagesAndSort Macro
A line-oriented search macro over multiple pages, with sorting
Pascal Bauermeister <pascal.bauermeister@smartdata.ch>
[[SearchInPages(PageRegex,TextRegex,SortKey)]]
Search for TextRegex in pages marching PagesRegex, and sort on
1) SortKey 2) TextRegex 3) found line
Original version:
* 2003/04/24 10:32:04
Updates:
* Wed Oct 29 14:48:02 CET 2003
works with MoinMoin Release 1.1 [Revision 1.173] and Python 2.3.2
"""
# Imports
import cgi, os, re, string, sys, time, urllib, cStringIO
from MoinMoin import config, user, util, wikiutil
from MoinMoin.Page import Page, webapi
from MoinMoin.parser.wiki import Parser
_arg_page = r'(?P<hquote1>[\'"])(?P<hpage>.+?)(?P=hquote1)'
_arg_text = r'(?P<hquote2>[\'"])(?P<htext>.+?)(?P=hquote2)'
_arg_key = r'(?P<hquote3>[\'"])(?P<hkey>.+?)(?P=hquote3)'
_args_re = re.compile(r'^(%s(,%s(,%s)?)?)?$' % (_arg_page, _arg_text, _arg_key))
def execute(macro, text, args_re=_args_re):
result = ""
# parse and check arguments
args = args_re.match(text)
if text is None or not args:
return ( '<p><strong class="error">Invalid SearchInPages arguments "%s"!' +
'</strong></p>' ) % text
text = args.group('htext')
pages = args.group('hpage')
key = args.group('hkey')
# get a list of pages matching the PageRegex
try:
pages_re = re.compile(pages, re.IGNORECASE)
except re.error:
pages_re = re.compile(re.escape(pages), re.IGNORECASE)
all_pages = wikiutil.getPageList(config.text_dir)
hits = filter(pages_re.search, all_pages)
hits.sort()
if len(hits) == 0:
return '<p><strong class="error">No page matching "%s"!</strong></p>' % (pages)
# compile all regex
try:
text_re = re.compile(text, re.IGNORECASE)
except:
text_re = re.compile(re.escape(text), re.IGNORECASE)
if key != None:
try:
key_re = re.compile(key, re.IGNORECASE)
except:
key_re = re.compile(re.escape(key), re.IGNORECASE)
# we will collect matching lines in each matching page
all_matches = []
for page_name in hits:
body = Page(page_name).get_raw_body()
pos = 0
last_start = -1
last_end = -1
while 1:
keep_line = 1
# search text
match = text_re.search(body, pos)
if not match: break
pos = match.end()+1
# cut before start of line
start_pos = match.start()
rev = 0
while body[start_pos] != '\n' and start_pos:
start_pos = start_pos - 1
rev = 1
if rev:
start_pos = start_pos + 1
# cut at end of line
end_pos = body.find("\n", match.end())
# extract line
line = body[start_pos:end_pos].strip()
# store this record if it differs from previous one
if start_pos == last_start or end_pos == last_end: keep_line = 0
# store this record if it it is not a comment
if line.startswith("##"): keep_line = 0
# store this record if it differs from previous one
if keep_line:
# remove possible list item leaders
for heading in ["*", "1.", "a.", "A.", "i.", "I.", "1.#n"]:
if line.startswith(heading):
line = line.replace(heading, "", 1)
line = line.strip()
# find the sort key
nbmatches = 0
keypos = 0
found = 0
while 1:
if key == None:
keyval = ""
else:
keymatch = key_re.search(line, keypos)
if keymatch:
keyval = line[keymatch.start():keymatch.end()]
keypos = keymatch.end()
nbmatches = nbmatches + 1
found = 1
else:
if nbmatches>0: break
keyval = "[unassigned]"
# store info
item = []
item.append(keyval) # key text
item.append(body[match.start():match.end()]) # search text
item.append(line) # line text
item.append(page_name) # page name
all_matches.append(item)
if found == 0: break
last_start = start_pos
last_end = end_pos
# sort and format records
bullet_list_open = macro.formatter.bullet_list(1)
bullet_list_close = macro.formatter.bullet_list(0)
listitem_open = macro.formatter.listitem(1)
listitem_close = macro.formatter.listitem(0)
all_matches.sort()
result = ""
result = result+"\n" + bullet_list_open
keyval = ""
head_count = 0
# treat records for output
for item in all_matches:
pagename = item[3]
text = item[2]
if key != None and item[0] != keyval:
# this is a new heading
keyval = item[0]
if head_count:
result = result+"\n " + bullet_list_close
result = result+"\n " + listitem_close
head_count = head_count +1
result = result+"\n " + listitem_open
result = result+ keyval
result = result+"\n " + bullet_list_open
# parse the text in wiki source format
stdout = sys.stdout
sys.stdout = cStringIO.StringIO()
Parser(text, macro.request).format(macro.formatter)
text_fmt = sys.stdout.getvalue()
sys.stdout = stdout
# correct the format
if text_fmt.startswith("\n<p>"):
text_fmt = text_fmt[4:]
# insert text
result = result+"\n " + listitem_open
result = result + text_fmt
result = result + " <font size=-1>"
result = result + wikiutil.link_tag(pagename)
result = result + "</font>"
result = result+"\n " + listitem_close
# all items done, close (hopefully) gracefully
if head_count:
result = result+"\n " + listitem_close
result = result+"\n " + bullet_list_close
if key != None:
result = result+"\n " + listitem_close
result = result+"\n" + bullet_list_close
# done
return result
