X-Git-Url: http://www.chiark.greenend.org.uk/ucgi/~yarrgweb/git?p=ypp-sc-tools.web-test.git;a=blobdiff_plain;f=yoweb-scrape;h=717b539131d8ffb0e4ba6abba5f6cddc3b38b199;hp=db5dc1338298a47a2c9120d2f93b168d07842ba5;hb=59824f48d09192d1c535adecf5ee2e4ee2ed47d5;hpb=e772c13a54236d87fede06c8f3364549b1c8e070

diff --git a/yoweb-scrape b/yoweb-scrape
index db5dc13..717b539 100755
--- a/yoweb-scrape
+++ b/yoweb-scrape
@@ -45,6 +45,7 @@ import curses
 import termios
 import random
 import subprocess
+import copy
 from optparse import OptionParser
 from StringIO import StringIO
 
@@ -107,6 +108,18 @@ def yppsc_dir():
 	os.environ["YPPSC_YARRG_SRCBASE"] = lib
 	return lib
 
+soup_massage = copy.copy(BeautifulSoup.MARKUP_MASSAGE)
+soup_massage.append(
+		(regexp.compile('(\<td.*") ("center")'),
+		 lambda m: m.group(1)+' align='+m.group(2))
+	)
+
+def make_soup(*args, **kwargs):
+	return BeautifulSoup(*args,
+		convertEntities=BeautifulSoup.HTML_ENTITIES,
+		markupMassage=soup_massage,
+			 **kwargs)
+
 #---------- caching and rate-limiting data fetcher ----------
 
 class Fetcher:
@@ -118,12 +131,24 @@ class Fetcher:
 			if oe.errno != errno.EEXIST: raise
 		self._cache_scan(time.time())
 
-	def _cache_scan(self, now):
+	def _match_url_normalise(self, url):
+		without_scheme = regexp.sub('^[a-z]+://', '', url)
+		without_tail = regexp.sub('/.*', '', without_scheme)
+		return without_tail
+
+	def _cache_scan(self, now, match_url=None):
 		# returns list of ages, unsorted
+		if match_url is not None:
+			match_url = self._match_url_normalise(match_url)
 		ages = []
 		debug('Fetcher   scan_cache')
 		for leaf in os.listdir(self.cachedir):
 			if not leaf.startswith('#'): continue
+			if match_url is not None:
+				leaf_url = urllib.unquote_plus(leaf.strip('#'))
+				leaf_url = self._match_url_normalise(leaf_url)
+				if leaf_url != match_url:
+					continue
 			path = self.cachedir + '/' + leaf
 			try: s = os.stat(path)
 			except (OSError,IOError), oe:
@@ -139,8 +164,8 @@ class Fetcher:
 			ages.append(age)
 		return ages
 
-	def need_wait(self, now, imaginary=[]):
-		ages = self._cache_scan(now)
+	def need_wait(self, now, imaginary=[], next_url=None):
+		ages = self._cache_scan(now, match_url=next_url)
 		ages += imaginary
 		ages.sort()
 		debug('Fetcher   ages ' + `ages`)
@@ -152,15 +177,14 @@ class Fetcher:
 					(min_age, age))
 				need_wait = max(need_wait, min_age - age)
 			min_age += 3
-			min_age *= 1.25
 		if need_wait > 0:
 			need_wait += random.random() - 0.5
 		return need_wait
 
-	def _rate_limit_cache_clean(self, now):
-		need_wait = self.need_wait(now)
+	def _rate_limit_cache_clean(self, now, next_url=None):
+		need_wait = self.need_wait(now, next_url=next_url)
 		if need_wait > 0:
-			debug('Fetcher   wait %d' % need_wait)
+			debug('Fetcher   wait %f' % need_wait)
 			sleep(need_wait)
 
 	def fetch(self, url, max_age):
@@ -186,7 +210,7 @@ class Fetcher:
 			return data
 
 		debug('Fetcher  fetch')
-		self._rate_limit_cache_clean(now)
+		self._rate_limit_cache_clean(now, next_url=url)
 
 		stream = urllib2.urlopen(url)
 		data = stream.read()
@@ -252,9 +276,7 @@ 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
-			)
+		self._soup = make_soup(html)
 
 #---------- scraper for pirate pages ----------
 
@@ -577,7 +599,7 @@ class IslandExtendedInfo(IslandBasicInfo):
 		self.islandid = None
 		self.yoweb_url = None
 
-		soup = BeautifulSoup(self.yppedia())
+		soup = make_soup(self.yppedia())
 		content = soup.find('div', attrs = {'id': 'content'})
 		yoweb_re = regexp.compile('^http://\w+\.puzzlepirates\.com/'+
 			'yoweb/island/info\.wm\?islandid=(\d+)$')
@@ -598,7 +620,7 @@ class IslandExtendedInfo(IslandBasicInfo):
 		debug('IEI COLLECT FLAGID '+`self.name`+' URL '+`yo`)
 		if yo is None: return None
 		dataf = fetcher.fetch(yo, 1800)
-		soup = BeautifulSoup(dataf)
+		soup = make_soup(dataf)
 		ruler_re = regexp.compile(
 			'/yoweb/flag/info\.wm\?flagid=(\d+)$')
 		ruler = soup.find('a', attrs = { 'href': ruler_re })