def parse_meta(json_spec: Dict) -> BotJsonMeta:
meta = BotJsonMeta()
meta.name = json_spec['name']
meta.race = PlayerRace[json_spec['race'].upper()]
bot_type = json_spec['botType']
if bot_type == "JAVA_JNI" or bot_type == "JAVA_MIRROR":
bot_type = "JAVA"
meta.botType = BotType[bot_type]
meta.description = json_spec[
'description'] if 'description' in json_spec else None
meta.update = parse_iso_date(
json_spec['update']) if 'update' in json_spec else None
meta.botBinary = json_spec[
'botBinary'] if 'botBinary' in json_spec else None
meta.bwapiDLL = json_spec[
'bwapiDLL'] if 'bwapiDLL' in json_spec else None
meta.botProfileURL = json_spec[
'botProfileURL'] if 'botProfileURL' in json_spec else None
meta.javaDebugPort = json_spec[
'javaDebugPort'] if 'javaDebugPort' in json_spec else None
meta.javaOpts = json_spec[
'javaOpts'] if 'javaOpts' in json_spec else None
return meta
def format_datetime_like(dt_object):
if dt_object is None:
result = dt_object
elif isinstance(dt_object, str):
try:
parse_iso_date(dt_object)
except Exception:
raise
else:
result = dt_object
elif isinstance(dt_object, dt.datetime):
result = dt_object.astimezone(
dt.timezone.utc).strftime(ISO8601_FORMAT_Z)
else:
result = dt_object.strftime(ISO8601_FORMAT_Z)
return result
def parse_meta(json_spec: Dict):
meta = BotJsonMeta()
meta.name = json_spec['name']
meta.race = PlayerRace[json_spec['race'].upper()]
meta.description = json_spec['description']
bot_type = json_spec['botType']
if bot_type == "JAVA_JNI" or bot_type == "JAVA_MIRROR":
bot_type = "JAVA"
meta.botType = BotType[bot_type]
meta.update = parse_iso_date(json_spec['update'])
meta.botBinary = json_spec['botBinary']
meta.bwapiDLL = json_spec['bwapiDLL']
meta.botProfileURL = json_spec['botProfileURL']
return meta
def task_post(self, grab, task):
print 'task_post', grab.response.url
self.platform.inc_api_calls()
title = content = date = None
# The Post's URL
url = grab.response.url
print "Checking url %s" % url
# The Post's Title
post_title = grab.doc.select(
'//div[contains(@id,"Blog")]//*[contains(@class,"post-title") or contains(@class,"postTitle")]'
)
if not post_title.count():
post_title = grab.doc.select(
'//div[contains(@id,"Blog")]//div[contains(@class,"Post")]//*[contains(@class,"PostHeader")]'
)
if not post_title.count():
post_title = grab.doc.select(
'//div[contains(@id,"Blog")]/div[@class="post"]/text()[1]')
if post_title.count():
title = post_title.text().strip()
# The Post's Content
post_content = grab.doc.select(
'//div[contains(@id,"Blog")]//div[(contains(@class,"post-body") or contains(@class,"postBody"))]'
)
if not post_content.count():
post_content = grab.doc.select(
'//div[contains(@id,"Blog")]//div[contains(@class,"Post")]//div[contains(@class,"PostContent")]'
)
if post_content.count():
content = post_content.html()
else:
return
# The post date can be present in different variations
# That's why we try to extract it from <abbr> (ISO format), after that from the date-header and finally from the URL
# In the header the date can be present in different formats: "Tuesday, 6 December 2011", "February 18, 2014", etc.
# Sometimes (http://www.tieandi.com/2014/02/valentines-day-wishlist.html) it's written not in English
# In this case we can only get year and month from the url ("/2014/02/"") and we set day=1
# We work with "date" over and over again, until this value is changed from None
# But if the date is None after all tries... well, I have no idea about
# when this post is published, may be a few thousands years ago?
post_date = grab.doc.select(
'//div[contains(@class,"post-footer")]//abbr[contains(@class,"published")]/@title'
)
# <abbr>
if post_date.count():
iso_date = post_date.text().strip()
try:
date = parse_iso_date(iso_date)
except ValueError:
# bad ISO-format? let's try another way
pass
# "date-header"
if not date:
post_date = grab.doc.select(
'//*[contains(@class,"date-header")]//span')
if not post_date.count():
post_date = grab.doc.select(
'//*[contains(@class,"date-header")]')
if not post_date.count():
post_date = grab.doc.select(
'//div[contains(@class,"post")]/*[contains(@class,"postAuthor")]/a'
)
if post_date.count():
YMDhm = self.parse_date(post_date.text())
if YMDhm:
date = datetime(*YMDhm)
# URL
if not date:
url_path = urlparse(grab.response.url).path
if url_path.startswith('/'):
url_path_parts = url_path.split('/')
month = year = None
year_from_url = url_path_parts[1]
month_from_url = url_path_parts[2]
if month_from_url.isdigit() and int(
month_from_url) in range(1, 13):
month = int(month_from_url)
if year_from_url.isdigit() and int(year_from_url) in range(
2000, 2020):
year = int(year_from_url)
if month and year:
date = datetime(year, month, 1, 0, 0)
post = models.Posts()
post.influencer = self.platform.influencer
post.show_on_search = self.platform.influencer.show_on_search
post.platform = self.platform
post.title = title
post.url = url
post.content = content
post.create_date = date
post.save()
self._inc('posts_saved')
print "Created post: %s " % post
self.posts[post] = []
# Comments
# This section also can be present in different variations
comments = []
# The first type
comments_blocks = grab.doc.select(
'//div[contains(@id,"comments")]//dl[contains(@id,"comments-block")]'
)
if comments_blocks.count():
comments_authors = comments_blocks.select(
'.//dt[contains(@class,"author")]')
for author in comments_authors:
author_name = author_url = url = None
author_a = author.select('./a[@rel="nofollow"]')
if not author_a.count():
author_a = author.select('.//a[@rel="nofollow"]')
if author_a.count():
author_name = author_a.text().strip()
try:
author_url = author_a.attr('href')
except Exception:
pass
comment_body = author.select(
'./following-sibling::dd[contains(@class,"comment-body")][1]'
)
if comment_body.count():
content = comment_body.html()
# content = ' '.join([_.text().strip() for _ in comment_body.select('.//text()')])
else:
continue
comment_footer = author.select(
'./following-sibling::dd[contains(@class,"comment-footer")][1]'
)
if comment_footer.count():
comment_timestamp = comment_footer.select(
'.//span[contains(@class,"comment-timestamp")]/a')
else:
comment_timestamp = author.select(
'.//span[contains(@class,"comment-timestamp")]/a')
if comment_timestamp.count():
try:
url = comment_timestamp.attr('href')
except Exception:
pass
timestamp = comment_timestamp.text()
if date and not str(
date.year)[:-1] in timestamp and ':' in timestamp:
timestamp = '.'.join([
str(_) for _ in (date.month, date.day, date.year)
]) + ' ' + timestamp
YMDhm = self.parse_date(timestamp)
if YMDhm:
date = datetime(*YMDhm)
comments.append(
dict(
author_name=author_name,
author_url=author_url,
content=content,
date=date,
url=url,
))
# The second type
else:
comments_blocks = grab.doc.select(
'//div[@*="comments"]//div[@*="comment-header"]')
for comment_header in comments_blocks:
author_name = author_url = url = None
author = comment_header.select(
'.//cite[contains(@class,"user")]')
if author.count():
author_a = author.select('./a')
if author_a.count():
author_name = author_a.text().strip()
try:
author_url = author_a.attr('href')
except Exception:
pass
else:
author_name = author.text().strip()
else:
author = comment_header.select(
'.//a[contains(@class, "autor-name")]')
if author.count():
author_name = author.text().strip()
try:
author_url = author.attr('href')
except Exception:
pass
comment_timestamp = comment_header.select(
'.//span[contains(@class,"datetime") or contains(@class,"timestamp") or contains(@id,"timestamp")]/a'
)
if comment_timestamp.count():
try:
url = comment_timestamp.attr('href')
except Exception:
pass
timestamp = comment_timestamp.text()
if date and not str(
date.year)[:-1] in timestamp and ':' in timestamp:
timestamp = '.'.join([
str(_) for _ in (date.month, date.day, date.year)
]) + ' ' + timestamp
YMDhm = self.parse_date(timestamp)
if YMDhm:
date = datetime(*YMDhm)
comment_body = comment_header.select(
'.//p[contains(@class,"comment-content") or contains(@class,"comment-body")]'
)
if not comment_body.count():
comment_body = comment_header.select(
'./following-sibling::p[contains(@class,"comment-content") or contains(@class,"comment-body")][1]'
)
if comment_body.count():
content = comment_body.html()
# content = ' '.join([_.text().strip() for _ in comment_body.select('.//text()')])
else:
continue
comments.append(
dict(
author_name=author_name,
author_url=author_url,
content=content,
date=date,
url=url,
))
print "got %d comments " % len(comments)
for comment in comments:
self.posts[post].append(comment)
# this will call self.fetch_post_interactions() first to see if we got any comments
# if not, disqus will be crawled
self.fetch_post_interactions_extra([post])
