#!/usr/bin/python #---------- setup ---------- import signal signal.signal(signal.SIGINT, signal.SIG_DFL) import os import time import urllib import urllib2 import errno import sys import re as regexp import random import curses from optparse import OptionParser from BeautifulSoup import BeautifulSoup opts = None #---------- YPP parameters and arrays ---------- puzzles = ('Swordfighting/Bilging/Sailing/Rigging/Navigating'+ '/Battle Navigation/Gunning/Carpentry/Rumble/Treasure Haul'+ '/Drinking/Spades/Hearts/Treasure Drop/Poker/Distilling'+ '/Alchemistry/Shipwrightery/Blacksmithing/Foraging').split('/') standingvals = ('Able/Distinguished/Respected/Master'+ '/Renowned/Grand-Master/Legendary/Ultimate').split('/') pirate_ref_re = regexp.compile('^/yoweb/pirate\\.wm') max_pirate_namelen = 12 #---------- general utilities ---------- def debug(m): if opts.debug > 0: print m def format_time_interval(ti): if ti < 120: return '%d:%02d' % (ti / 60, ti % 60) if ti < 7200: return '%2dm' % (ti / 60) if ti < 86400: return '%dh' % (ti / 3600) return '%dd' % (ti / 86400) #---------- caching and rate-limiting data fetcher ---------- class Fetcher: def __init__(self, ocean, cachedir): debug('Fetcher init %s' % cachedir) self.ocean = ocean self.cachedir = cachedir try: os.mkdir(cachedir) except (OSError,IOError), oe: if oe.errno != errno.EEXIST: raise self._cache_scan(time.time()) def default_ocean(self, ocean='ice'): if self.ocean is None: self.ocean = ocean def _cache_scan(self, now): # returns list of ages, unsorted ages = [] debug('Fetcher scan_cache') for leaf in os.listdir(self.cachedir): if not leaf.startswith('#'): continue path = self.cachedir + '/' + leaf try: s = os.stat(path) except (OSError,IOError), oe: if oe.errno != errno.ENOENT: raise continue age = now - s.st_mtime if age > opts.expire_age: debug('Fetcher expire %d %s' % (age, path)) try: os.remove(path) except (OSError,IOError), oe: if oe.errno != errno.ENOENT: raise continue ages.append(age) return ages def need_wait(self, now): ages = self._cache_scan(now) ages.sort() debug('Fetcher ages ' + `ages`) min_age = 1 need_wait = 0 for age in ages: if age < min_age and age < 300: debug('Fetcher morewait min=%d age=%d' % (min_age, age)) need_wait = max(need_wait, min_age - age) min_age += 3 min_age *= 1.25 return need_wait def _rate_limit_cache_clean(self, now): need_wait = self.need_wait(now) if need_wait > 0: debug('Fetcher wait %d' % need_wait) time.sleep(need_wait) def fetch(self, url, max_age): debug('Fetcher fetch %s' % url) cache_corename = urllib.quote_plus(url) cache_item = "%s/#%s#" % (self.cachedir, cache_corename) try: f = file(cache_item, 'r') except (OSError,IOError), oe: if oe.errno != errno.ENOENT: raise f = None now = time.time() max_age = max(opts.min_max_age, min(max_age, opts.expire_age)) if f is not None: s = os.fstat(f.fileno()) age = now - s.st_mtime if age > max_age: debug('Fetcher stale %d < %d'% (max_age, age)) f = None if f is not None: data = f.read() f.close() debug('Fetcher cached %d > %d' % (max_age, age)) return data debug('Fetcher fetch') self._rate_limit_cache_clean(now) stream = urllib2.urlopen(url) data = stream.read() cache_tmp = "%s/#%s~%d#" % ( self.cachedir, cache_corename, os.getpid()) f = file(cache_tmp, 'w') f.write(data) f.close() os.rename(cache_tmp, cache_item) debug('Fetcher stored') return data def yoweb(self, kind, tail, max_age): self.default_ocean() url = 'http://%s.puzzlepirates.com/yoweb/%s%s' % ( self.ocean, kind, tail) return self.fetch(url, max_age) #---------- logging assistance for troubled screenscrapers ---------- class SoupLog: def __init__(self): self.msgs = [ ] def msg(self, m): self.msgs.append(m) def soupm(self, obj, m): self.msg(m + '; in ' + `obj`) def needs_msgs(self, child_souplog): self.msgs += child_souplog.msgs child_souplog.msgs = [ ] def soup_text(obj): str = ''.join(obj.findAll(text=True)) return str.strip() class SomethingSoupInfo(SoupLog): def __init__(self, kind, tail, max_age): SoupLog.__init__(self) html = fetcher.yoweb(kind, tail, max_age) self._soup = BeautifulSoup(html, convertEntities=BeautifulSoup.HTML_ENTITIES ) #---------- scraper for pirate pages ---------- class PirateInfo(SomethingSoupInfo): # Public data members: # pi.standings = { 'Treasure Haul': 'Able' ... } # pi.name = name # pi.crew = (id, name) # pi.flag = (id, name) # pi.msgs = [ 'message describing problem with scrape' ] def __init__(self, pirate, max_age=300): SomethingSoupInfo.__init__(self, 'pirate.wm?target=', pirate, max_age) self.name = pirate self._find_standings() self.crew = self._find_crewflag('crew', '^/yoweb/crew/info\\.wm') self.flag = self._find_crewflag('flag', '^/yoweb/flag/info\\.wm') def _find_standings(self): imgs = self._soup.findAll('img', src=regexp.compile('/yoweb/images/stat.*')) re = regexp.compile( u'\\s*\\S*/([-A-Za-z]+)\\s*$|\\s*\\S*/\\S*\\s*\\(ocean\\-wide(?:\\s|\\xa0)+([-A-Za-z]+)\\)\\s*$' ) standings = { } for skill in puzzles: standings[skill] = [ ] skl = SoupLog() for img in imgs: try: puzzle = img['alt'] except KeyError: continue if not puzzle in puzzles: skl.soupm(img, 'unknown puzzle: "%s"' % puzzle) continue key = img.findParent('td') if key is None: skl.soupm(img, 'puzzle at root! "%s"' % puzzle) continue valelem = key.findNextSibling('td') if valelem is None: skl.soupm(key, 'puzzle missing sibling "%s"' % puzzle) continue valstr = soup_text(valelem) match = re.match(valstr) if match is None: skl.soupm(key, ('puzzle "%s" unparseable'+ ' standing "%s"') % (puzzle, valstr)) continue standing = match.group(match.lastindex) standings[puzzle].append(standing) self.standings = { } for puzzle in puzzles: sl = standings[puzzle] if len(sl) > 1: skl.msg('puzzle "%s" multiple standings %s' % (puzzle, `sl`)) continue if not sl: skl.msg('puzzle "%s" no standing found' % puzzle) continue standing = sl[0] for i in range(0, len(standingvals)-1): if standing == standingvals[i]: self.standings[puzzle] = i if not puzzle in self.standings: skl.msg('puzzle "%s" unknown standing "%s"' % (puzzle, standing)) all_standings_ok = True for puzzle in puzzles: if not puzzle in self.standings: self.needs_msgs(skl) def _find_crewflag(self, cf, yoweb_re): things = self._soup.findAll('a', href=regexp.compile(yoweb_re)) if len(things) != 1: self.msg('zero or several %s id references found' % cf) return None thing = things[0] id_re = '\\b%sid\\=(\\w+)$' % cf id_haystack = thing['href'] match = regexp.compile(id_re).search(id_haystack) if match is None: self.soupm(thing, ('incomprehensible %s id ref'+ ' (%s in %s)') % (cf, id_re, id_haystack)) return None name = soup_text(thing) return (match.group(1), name) def __str__(self): return `(self.crew, self.flag, self.standings, self.msgs)` #---------- scraper for crew pages ---------- class CrewInfo(SomethingSoupInfo): # Public data members: # ci.crew = [ ('Captain', ['Pirate', ...]), # ('Senior Officer', [...]), # ... ] # pi.msgs = [ 'message describing problem with scrape' ] def __init__(self, crewid, max_age=300): SomethingSoupInfo.__init__(self, 'crew/info.wm?crewid=', crewid, max_age) self._find_crew() def _find_crew(self): self.crew = [] capts = self._soup.findAll('img', src='/yoweb/images/crew-captain.png') if len(capts) != 1: self.msg('crew members: no. of captain images != 1') return tbl = capts[0] while not tbl.find('a', href=pirate_ref_re): tbl = tbl.findParent('table') if not tbl: self.msg('crew members: cannot find table') return current_rank_crew = None crew_rank_re = regexp.compile('/yoweb/images/crew') for row in tbl.contents: # findAll(recurse=False) if isinstance(row,basestring): continue is_rank = row.find('img', attrs={'src': crew_rank_re}) if is_rank: rank = soup_text(row) current_rank_crew = [] self.crew.append((rank, current_rank_crew)) continue for cell in row.findAll('a', href=pirate_ref_re): if current_rank_crew is None: self.soupm(cell, 'crew members: crew' ' before rank') continue current_rank_crew.append(soup_text(cell)) def __str__(self): return `(self.crew, self.msgs)` #---------- pretty-printer for tables of pirate puzzle standings ---------- class StandingsTable: def __init__(self, use_puzzles=None, col_width=6): if use_puzzles is None: if opts.ship_duty: use_puzzles=[ 'Navigating','Battle Navigation', 'Gunning', ['Sailing','Rigging'], 'Bilging', 'Carpentry', 'Treasure Haul' ] else: use_puzzles=puzzles self._puzzles = use_puzzles self.s = '' self._cw = col_width-1 def _pline(self, pirate, puzstrs, extra): self.s += ' %-*s' % (max(max_pirate_namelen, 14), pirate) for v in puzstrs: self.s += ' %-*.*s' % (self._cw,self._cw, v) if extra: self.s += ' ' + extra self.s += '\n' def _puzstr(self, pi, puzzle): if not isinstance(puzzle,list): puzzle = [puzzle] try: standing = max([pi.standings[p] for p in puzzle]) except KeyError: return '?' if not standing: return '' s = '' if self._cw > 4: c1 = standingvals[standing][0] if standing < 3: c1 = c1.lower() # 3 = Master s += `standing` if self._cw > 5: s += ' ' s += '*' * (standing / 2) s += '+' * (standing % 2) return s def headings(self): def puzn_redact(name): if isinstance(name,list): return '/'.join( ["%.*s" % (self._cw/2, puzn_redact(n)) for n in name]) spc = name.find(' ') if spc < 0: return name return name[0:min(4,spc)] + name[spc+1:] self._pline('', map(puzn_redact, self._puzzles), None) def literalline(self, line): self.s += line + '\n' def pirate_dummy(self, name, standingstring, extra=None): self._pline(name, standingstring * len(self._puzzles), extra) def pirate(self, pi, extra=None): puzstrs = [self._puzstr(pi,puz) for puz in self._puzzles] self._pline(pi.name, puzstrs, extra) def results(self): return self.s #---------- chat log parser ---------- class PirateAboard: # This is essentially a transparent, dumb, data class. # pa.v # pa.name # pa.last_time # pa.last_event # pa.gunner # pa.last_chat_time # pa.last_chat_chan # pa.pi def __init__(pa, pn, v, time, event): pa.name = pn pa.v = v pa.last_time = time pa.last_event = event pa.last_chat_time = None pa.last_chat_chan = None pa.gunner = False pa.pi = None def pirate_info(pa): if not pa.pi and not fetcher.need_wait(time.time()): pa.pi = PirateInfo(pa.name, 3600) return pa.pi class ChatLogTracker: # This is quite complex so we make it opaque. Use the # official invokers, accessors etc. def __init__(self, myself_pi, logfn): self._pl = {} # self._pl['Pirate'] = self._vl = {} # self._vl['Vessel']['Pirate'] = PirateAboard # self._vl['Vessel']['#lastaboard'] self._v = None # self._v = self._vessel = None # self._vl[self._vessel] self._date = None self._myself = myself_pi self._need_redisplay = False self._f = file(logfn) self._lbuf = '' self._progress = [0, os.fstat(self._f.fileno()).st_size] def force_redisplay(self): self._need_redisplay = True def _onboard_event(self,timestamp,pirate,event): try: pa = self._pl[pirate] except KeyError: pa = None if pa is not None and pa.v is self._v: pa.last_time = timestamp pa.last_event = event else: if pa is not None: del pa.v[pirate] pa = PirateAboard(pirate, self._v, timestamp, event) self._pl[pirate] = pa self._v[pirate] = pa self._v['#lastaboard'] = timestamp self.force_redisplay() return pa def _trash_vessel(self, v): for pn in v: if pn.startswith('#'): continue del self._pl[pn] self.force_redisplay() def expire_garbage(self, timestamp): for (vn,v) in list(self._vl.iteritems()): la = v['#lastaboard'] if timestamp - la > opts.ship_reboard_clearout: self._debug_line_disposition(timestamp,'', 'stale reset '+vn) self._trash_vessel(v) del self._vl[vn] def clear_vessel(self, timestamp): if self._v is not None: self._trash_vessel(self._v) self._v = {'#lastaboard': timestamp} self._vl[self._vessel] = self._v def _debug_line_disposition(self,timestamp,l,m): debug('CLT %13s %-30s %s' % (timestamp,m,l)) def chatline(self,l): rm = lambda re: regexp.match(re,l) d = lambda m: self._debug_line_disposition(timestamp,l,m) timestamp = None m = rm('=+ (\\d+)/(\\d+)/(\\d+) =+$') if m: self._date = [int(x) for x in m.groups()] self._previous_timestamp = None return d('date '+`self._date`) if self._date is None: return d('date unset') m = rm('\\[(\d\d):(\d\d):(\d\d)\\] ') if not m: return d('no timestamp') while True: time_tuple = (self._date + [int(x) for x in m.groups()] + [-1,-1,-1]) timestamp = time.mktime(time_tuple) if timestamp >= self._previous_timestamp: break self._date[2] += 1 self._debug_line_disposition(timestamp,'', 'new date '+`self._date`) self._previous_timestamp = timestamp l = l[l.find(' ')+1:] def ob_x(who,event): return self._onboard_event(timestamp, who, event) def ob1(did): ob_x(m.group(1), did); return d(did) def oba(did): return ob1('%s %s' % (did, m.group(2))) m = rm('Going aboard the (\\S.*\\S)\\.\\.\\.$') if m: pn = self._myself.name self._vessel = m.group(1) dm = 'boarding' try: self._v = self._vl[self._vessel] except KeyError: self._v = None; dm += ' new' if self._v is not None: la = self._v['#lastaboard'] else: la = 0; dm += ' ?la' if timestamp - la > opts.ship_reboard_clearout: self.clear_vessel(timestamp) dm += ' stale' ob_x(pn, 'we boarded') self.expire_garbage(timestamp) return d(dm) if self._v is None: return d('no vessel') m = rm('(\\w+) has come aboard\\.$') if m: return ob1('boarded'); m = rm('You have ordered (\\w+) to do some (\\S.*\\S)\\.$') if m: (who,what) = m.groups() pa = ob_x(who,'ord '+what) if what == 'Gunning': pa.gunner = True return d('duty order') m = rm('(\\w+) abandoned a (\\S.*\\S) station\\.$') if m: oba('stopped'); return d("end") def chat(what): who = m.group(1) try: pa = self._pl[who] except KeyError: return d('chat mystery') if pa.v is self._v: pa.last_chat_time = timestamp pa.last_chat_chan = what self.force_redisplay() return d(what+' chat') m = rm('(\\w+) (?:issued an order|ordered everyone) "') if m: return ob1('general order'); m = rm('(\\w+) says, "') if m: return chat('public') m = rm('(\\w+) tells ye, "') if m: return chat('private') m = rm('(\\w+) flag officer chats, "') if m: return chat('flag officer') m = rm('(\\w+) officer chats, "') if m: return chat('officer') m = rm('Game over\\. Winners: ([A-Za-z, ]+)\\.$') if m: pl = m.group(1).split(', ') if not self._myself.name in pl: return d('lost boarding battle') for pn in pl: if ' ' in pn: continue ob_x(pn,'won boarding battle') return d('won boarding battle') m = rm('(\\w+) is eliminated\\!') if m: return ob1('eliminated in fray'); m = rm('(\\w+) has left the vessel\.') if m: who = m.group(1) ob_x(who, 'disembarked') del self._v[who] del self._pl[who] return d('disembarked') return d('not matched') def _str_vessel(self, vn, v): s = ' vessel %s\n' % vn s += ' '*20 + "%-*s %13s\n" % ( max_pirate_namelen, '#lastaboard', v['#lastaboard']) for pn in sorted(v.keys()): if pn.startswith('#'): continue pa = v[pn] assert pa.v == v assert self._pl[pn] == pa s += ' '*20 + "%s %-*s %13s %-30s %13s %s\n" % ( (' ','G')[pa.gunner], max_pirate_namelen, pn, pa.last_time, pa.last_event, pa.last_chat_time, pa.last_chat_chan) return s def __str__(self): s = '''>self._f, 'Starting up, %s on the %s ocean' % ( pirate, ocean) def caughtup(self): self._f.write(' \r') self._f.flush() #----- modes which use the chat log parser are quite complex ----- def prep_chat_log(args, bu, progress=ProgressPrintPercentage(), max_myself_age=3600): if len(args) != 1: bu('this action takes only chat log filename') logfn = args[0] logfn_re = '(?:.*/)?([A-Z][a-z]+)_([a-z]+)_' match = regexp.match(logfn_re, logfn) if not match: bu('chat log filename is not in expected format') (pirate, ocean) = match.groups() fetcher.default_ocean(ocean) myself = PirateInfo(pirate,max_myself_age) progress.show_init(pirate, fetcher.ocean) track = ChatLogTracker(myself, logfn) opts.debug -= 1 track.catchup(progress) opts.debug += 1 track.force_redisplay() return (myself, track) def do_track_chat_log(args, bu): (myself, track) = prep_chat_log(args, bu) while True: track.catchup() if track.changed(): print track time.sleep(1) #----- ship management aid ----- class Display_dumb(ProgressPrintPercentage): def __init__(self): ProgressPrintPercentage.__init__(self) def show(self, s): print '\n\n', s; def realstart(self): pass class Display_overwrite(ProgressPrintPercentage): def __init__(self): ProgressPrintPercentage.__init__(self) null = file('/dev/null','w') curses.setupterm(fd=null.fileno()) self._clear = curses.tigetstr('clear') if not self._clear: self._debug('missing clear!') self.show = Display_dumb.show return self._t = {'el':'', 'ed':''} if not self._init_sophisticated(): for k in self._t.keys(): self._t[k] = '' self._t['ho'] = self._clear def _debug(self,m): debug('display overwrite: '+m) def _init_sophisticated(self): for k in self._t.keys(): s = curses.tigetstr(k) self._t[k] = s self._t['ho'] = curses.tigetstr('ho') if not self._t['ho']: cup = curses.tigetstr('cup') self._t['ho'] = curses.tparm(cup,0,0) missing = [k for k in self._t.keys() if not self._t[k]] if missing: self.debug('missing '+(' '.join(missing))) return 0 return 1 def show(self, s): w = sys.stdout.write def wti(k): w(self._t[k]) wti('ho') nl = '' for l in s.rstrip().split('\n'): w(nl) w(l) wti('el') nl = '\r\n' wti('ed') w(' ') sys.stdout.flush() def realstart(self): sys.stdout.write(self._clear) sys.stdout.flush() def do_ship_aid(args, bu): if opts.ship_duty is None: opts.ship_duty = True displayer = globals()['Display_'+opts.display]() rotate_nya = '/-\\' (myself, track) = prep_chat_log(args, bu, progress=displayer) def timeevent(t,e): if t is None: return ' ' * 22 return " %-4s %-16s" % (format_time_interval(now - t),e) displayer.realstart() while True: track.catchup() now = time.time() s = "%s" % track.myname() vn = track.vessel() if vn is None: s += " not on a vessel?!" else: s += " on board the %s" % vn s += " at %s\n" % time.strftime("%Y-%m-%d %H:%M:%S") tbl = StandingsTable() tbl.headings() for pa in track.aboard(): pi = pa.pirate_info() xs = '' if pa.gunner: xs += 'G ' else: xs += ' ' xs += timeevent(pa.last_time, pa.last_event) xs += timeevent(pa.last_chat_time, pa.last_chat_chan) if pi is None: tbl.pirate_dummy(pa.name, rotate_nya[0], xs) else: tbl.pirate(pi, xs) s += tbl.results() displayer.show(s) time.sleep(1) rotate_nya = rotate_nya[1:2] + rotate_nya[0] #---------- main program ---------- def main(): global opts, fetcher pa = OptionParser( '''usage: .../yoweb-scrape [OPTION...] ACTION [ARGS...] actions: yoweb-scrape [--ocean OCEAN ...] pirate PIRATE yoweb-scrape [--ocean OCEAN ...] crew-of PIRATE yoweb-scrape [--ocean OCEAN ...] standings-crew-of PIRATE yoweb-scrape [--ocean OCEAN ...] track-chat-log CHAT-LOG yoweb-scrape [options] ship-aid CHAT-LOG (must be .../PIRATE_OCEAN_chat-log*) display modes (for --display) apply to ship-aid: --display=dumb just print new information, scrolling the screen --display=overwrite use cursor motion, selective clear, etc. to redraw at top ''') ao = pa.add_option ao('-O','--ocean',dest='ocean', metavar='OCEAN', default=None, help='select ocean OCEAN') ao('--cache-dir', dest='cache_dir', metavar='DIR', default='~/.yoweb-scrape-cache', help='cache yoweb pages in DIR') ao('-D','--debug', action='count', dest='debug', default=0, help='enable debugging output') ao('--debug-fd', action='count', dest='debug_fd', help='write any debugging output to specified fd') ao('-q','--quiet', action='store_true', dest='quiet', help='suppress warning output') ao('--display', action='store', dest='display', type='choice', choices=['dumb','overwrite'], help='how to display ship aid') ao('--ship-duty', action='store_true', dest='ship_duty', help='show ship duty station puzzles') ao('--all-puzzles', action='store_false', dest='ship_duty', help='show all puzzles, not just ship duty stations') (opts,args) = pa.parse_args() random.seed() if len(args) < 1: pa.error('need a mode argument') if opts.debug_fd is not None: opts.debug_file = fdopen(opts.debug_fd, 'w') mode = args[0] mode_fn_name = 'do_' + mode.replace('_','#').replace('-','_') try: mode_fn = globals()[mode_fn_name] except KeyError: pa.error('unknown mode "%s"' % mode) # fixed parameters opts.min_max_age = 60 opts.expire_age = 3600 opts.ship_reboard_clearout = 3600 if opts.cache_dir.startswith('~/'): opts.cache_dir = os.getenv('HOME') + opts.cache_dir[1:] if opts.display is None: if ((opts.debug > 0 and opts.debug_fd is None) or not os.isatty(sys.stdout.fileno())): opts.display = 'dumb' else: opts.display = 'overwrite' fetcher = Fetcher(opts.ocean, opts.cache_dir) mode_fn(args[1:], pa.error) main()