def _do_almost_everything(self, localpath):
page_data = super(OneBoxAction, self)._do_almost_everything(
localpath + '/' + ONE_BOX_URL)
if page_data:
builders = self.request.GET.getall('builder')
if builders and len(builders):
one_box = BeautifulSoup(page_data['content'])
all_tds = one_box.findAll('td')
for td in all_tds:
if td.a and td.a['title'] not in builders:
td.extract()
page_data['content'] = self.ContentsToHtml(one_box)
return page_data
def _do_almost_everything(self, localpath):
page_data = super(OneBoxAction,
self)._do_almost_everything(localpath + '/' +
ONE_BOX_URL)
if page_data:
builders = self.request.GET.getall('builder')
if builders and len(builders):
one_box = BeautifulSoup(page_data['content'])
all_tds = one_box.findAll('td')
for td in all_tds:
if td.a and td.a['title'] not in builders:
td.extract()
page_data['content'] = self.ContentsToHtml(one_box)
return page_data
def test_html_format(self):
import gatekeeper_mailer
template = gatekeeper_mailer.MailTemplate(
self.build_data['waterfall_url'], self.build_data['build_url'],
self.build_data['project_name'], '*****@*****.**')
_, html_content, _ = template.genMessageContent(self.build_data)
expected_html = ' '.join(self.read_file('expected.html').splitlines())
saw = str(BeautifulSoup(html_content)).split()
expected = str(BeautifulSoup(expected_html)).split()
self.assertEqual(saw, expected)
def AddRow(self, row):
self.SawRevision(row['rev'], row['rev_number'])
revlink = BeautifulSoup(row['rev']).a['href']
self.SetLink(revlink)
name = BeautifulSoup(row['name'])
self.SetName(self.ContentsToHtml(name))
status = BeautifulSoup(row['status']).findAll('table')
for i, stat in enumerate(status):
self.SetStatus(self.category_order[self.lastMasterSeen][i],
unicode(stat))
comment = BeautifulSoup(row['comment'])
self.SetComment(self.ContentsToHtml(comment))
if row['details']:
details = BeautifulSoup(row['details'])
self.SetDetail(self.ContentsToHtml(details))
def update_status(master, status_html, status_dict):
"""Parses build status information and saves it to a status dictionary."""
builder_soup = BeautifulSoup(status_html)
builders_by_category = builder_soup.findAll('table')
for i, c in enumerate(data.ordered_categories[master]):
status_dict[master].setdefault(c, {})
statuses_by_builder = builders_by_category[i].findAll(
'td', 'DevStatusBox')
# If we didn't get anything, it's because we're parsing the overall
# summary, so look for Slave boxes instead of Status boxes.
if not statuses_by_builder:
statuses_by_builder = builders_by_category[i].findAll(
'td', 'DevSlaveBox')
for j, b in enumerate(data.ordered_builders[master][c]):
# Save the whole link as the status to keep ETA and build number info.
status = unicode(statuses_by_builder[j].a)
status_dict[master][c][b] = status
def update_status(master, status_html, status_dict):
"""Parses build status information and saves it to a status dictionary."""
builder_soup = BeautifulSoup(status_html)
builders_by_category = builder_soup.findAll('table')
for i, c in enumerate(data.ordered_categories[master]):
status_dict[master].setdefault(c, {})
statuses_by_builder = builders_by_category[i].findAll('td',
'DevStatusBox')
# If we didn't get anything, it's because we're parsing the overall
# summary, so look for Slave boxes instead of Status boxes.
if not statuses_by_builder:
statuses_by_builder = builders_by_category[i].findAll('td',
'DevSlaveBox')
for j, b in enumerate(data.ordered_builders[master][c]):
# Save the whole link as the status to keep ETA and build number info.
status = unicode(statuses_by_builder[j].a)
status_dict[master][c][b] = status
def test_html_format_status(self):
import gatekeeper_mailer
status_header = ('Perf alert for "%(steps)s" on "%(builder_name)s"')
template = gatekeeper_mailer.MailTemplate(
self.build_data['waterfall_url'],
self.build_data['build_url'],
self.build_data['project_name'],
'*****@*****.**',
status_header=status_header)
_, html_content, _ = template.genMessageContent(self.build_data)
expected_html = ' '.join(
self.read_file('expected_status.html').splitlines())
saw = str(BeautifulSoup(html_content)).split()
expected = str(BeautifulSoup(expected_html)).split()
self.assertEqual(saw, expected)
def bootstrap(self):
"""Fills an empty MergerData with 100 rows of data."""
# Populate the categories, masters, status, and failures data.
for m in self.ordered_masters:
for d in (self.ordered_builders, self.ordered_categories,
self.status, self.failures):
d.setdefault(m, {})
# Get the category data and construct the list of categories
# for this master.
category_data = app.get_and_cache_pagedata(
'%s/console/categories' % m)
if not category_data['content']:
category_list = [u'default']
else:
category_soup = BeautifulSoup(category_data['content'])
category_list = [
tag.string.strip()
for tag in category_soup.findAll('td', 'DevStatus')
]
self.ordered_categories[m] = category_list
# Get the builder status data.
builder_data = app.get_and_cache_pagedata('%s/console/summary' % m)
if not builder_data['content']:
continue
builder_soup = BeautifulSoup(builder_data['content'])
builders_by_category = builder_soup.tr.findAll('td',
'DevSlave',
recursive=False)
# Construct the list of builders for this category.
for i, c in enumerate(self.ordered_categories[m]):
self.ordered_builders[m].setdefault(c, {})
builder_list = [
tag['title'] for tag in builders_by_category[i].findAll(
'a', 'DevSlaveBox')
]
self.ordered_builders[m][c] = builder_list
# Fill in the status data for all of this master's builders.
update_status(m, builder_data['content'], self.status)
# Copy that status data over into the failures dictionary too.
for c in self.ordered_categories[m]:
self.failures[m].setdefault(c, {})
for b in self.ordered_builders[m][c]:
if self.status[m][c][b] not in ('success', 'running',
'notstarted'):
self.failures[m][c][b] = True
else:
self.failures[m][c][b] = False
# Populate the individual row data, saving status info in the same
# master/category/builder tree format constructed above.
latest_rev = int(app.get_and_cache_rowdata('latest_rev')['rev_number'])
if not latest_rev:
logging.error(
"MergerData.bootstrap(): Didn't get latest_rev. Aborting.")
return
n = latest_rev
num_rows_saved = num_rows_skipped = 0
while num_rows_saved < self.SIZE and num_rows_skipped < 10:
curr_row = RowData()
for m in self.ordered_masters:
update_row(n, m, curr_row)
# If we didn't get any data, that revision doesn't exist, so skip on.
if not curr_row.revision:
num_rows_skipped += 1
n -= 1
continue
self.rows[n] = curr_row
num_rows_skipped = 0
num_rows_saved += 1
n -= 1
self.latest_rev = max(self.rows.keys())
def notstarted(status):
"""Converts a DevSlave status box to a notstarted DevStatus box."""
status_soup = BeautifulSoup(status)
status_soup['class'] = 'DevStatusBox notstarted'
return unicode(status_soup)
def bootstrap(self):
"""Fills an empty MergerData with 100 rows of data."""
# Populate the categories, masters, status, and failures data.
for m in self.ordered_masters:
for d in (self.ordered_builders,
self.ordered_categories,
self.status,
self.failures):
d.setdefault(m, {})
# Get the category data and construct the list of categories
# for this master.
category_data = app.get_and_cache_pagedata('%s/console/categories' % m)
if not category_data['content']:
category_list = [u'default']
else:
category_soup = BeautifulSoup(category_data['content'])
category_list = [tag.string.strip() for tag in
category_soup.findAll('td', 'DevStatus')]
self.ordered_categories[m] = category_list
# Get the builder status data.
builder_data = app.get_and_cache_pagedata('%s/console/summary' % m)
if not builder_data['content']:
continue
builder_soup = BeautifulSoup(builder_data['content'])
builders_by_category = builder_soup.tr.findAll('td', 'DevSlave',
recursive=False)
# Construct the list of builders for this category.
for i, c in enumerate(self.ordered_categories[m]):
self.ordered_builders[m].setdefault(c, {})
builder_list = [tag['title'] for tag in
builders_by_category[i].findAll('a', 'DevSlaveBox')]
self.ordered_builders[m][c] = builder_list
# Fill in the status data for all of this master's builders.
update_status(m, builder_data['content'], self.status)
# Copy that status data over into the failures dictionary too.
for c in self.ordered_categories[m]:
self.failures[m].setdefault(c, {})
for b in self.ordered_builders[m][c]:
if self.status[m][c][b] not in ('success', 'running', 'notstarted'):
self.failures[m][c][b] = True
else:
self.failures[m][c][b] = False
# Populate the individual row data, saving status info in the same
# master/category/builder tree format constructed above.
latest_rev = int(app.get_and_cache_rowdata('latest_rev')['rev_number'])
if not latest_rev:
logging.error("MergerData.bootstrap(): Didn't get latest_rev. Aborting.")
return
n = latest_rev
num_rows_saved = num_rows_skipped = 0
while num_rows_saved < self.SIZE and num_rows_skipped < 10:
curr_row = RowData()
for m in self.ordered_masters:
update_row(n, m, curr_row)
# If we didn't get any data, that revision doesn't exist, so skip on.
if not curr_row.revision:
num_rows_skipped += 1
n -= 1
continue
self.rows[n] = curr_row
num_rows_skipped = 0
num_rows_saved += 1
n -= 1
self.latest_rev = max(self.rows.keys())
def parse_master(localpath, remoteurl, page_data=None):
"""Part of the new pipeline to store individual rows rather than
whole pages of html. Parses the master data into a set of rows,
and writes them out to the datastore in an easily retrievable format.
Doesn't modify page_data dict.
"""
ts = datetime.datetime.now()
page_data = page_data or {}
content = page_data.get('content')
if not content:
return page_data
content = content.decode('utf-8', 'replace')
# Split page into surroundings (announce, legend, footer) and data (rows).
surroundings = BeautifulSoup(content)
data = surroundings.find('table', 'ConsoleData')
if data is None:
raise Exception('parse_master: data can not be None')
new_data = Tag(surroundings, 'table', [('class', 'ConsoleData'),
('width', '96%')])
data.replaceWith(new_data)
surroundings_page = get_or_create_page(localpath + '/surroundings',
None, maxage=30)
surroundings_data = {}
surroundings_data['title'] = 'Surroundings for ' + localpath
surroundings_data['content'] = utf8_convert(surroundings)
save_page(surroundings_page, localpath + '/surroundings', ts,
surroundings_data)
rows = data.findAll('tr', recursive=False)
# The first table row can be special: the list of categories.
categories = None
# If the first row contains a DevStatus cell...
if rows[0].find('td', 'DevStatus') != None:
# ...extract it into the categories...
categories = rows[0]
# ...and get rid of the next (spacer) row too.
rows = rows[2:]
if categories:
category_page = get_or_create_page(localpath + '/categories',
None, maxage=30)
category_data = {}
category_data['title'] = 'Categories for ' + localpath
category_data['content'] = utf8_convert(categories)
save_page(category_page, localpath + '/categories', ts, category_data)
# The next table row is special, it's the summary one-box-per-builder.
summary = rows[0]
rows = rows[1:]
summary_page = get_or_create_page(localpath + '/summary', None, maxage=30)
summary_data = {}
summary_data['title'] = 'Summary for ' + localpath
summary_data['content'] = utf8_convert(summary)
save_page(summary_page, localpath + '/summary', ts, summary_data)
curr_row = {}
# Each table row is either a status row with a revision, name, and status,
# a comment row with the commit message, a details row with flakiness info,
# or a spacer row (in which case we finalize the row and save it).
for row in rows:
if row.find('td', 'DevComment'):
curr_row['comment'] = ''.join(utf8_convert(tag).strip()
for tag in row.td.contents)
elif row.find('td', 'DevDetails'):
curr_row['details'] = ''.join(utf8_convert(tag).strip()
for tag in row.td.contents)
elif row.find('td', 'DevStatus'):
curr_row['rev'] = ''.join(utf8_convert(tag).strip()
for tag in row.find('td', 'DevRev').contents)
curr_row['name'] = ''.join(utf8_convert(tag).strip()
for tag in row.find('td', 'DevName').contents)
curr_row['status'] = ''.join(utf8_convert(box.table).strip()
for box in row.findAll('td', 'DevStatus'))
else:
if 'details' not in curr_row:
curr_row['details'] = ''
curr_row['fetch_timestamp'] = ts
curr_row['rev_number'] = get_position_number(curr_row['comment'])
save_row(curr_row, localpath + '/' + curr_row['rev_number'])
curr_row = {}
return page_data
def console_merger(localpath, remoteurl, page_data,
masters_to_merge=None, num_rows_to_merge=None):
masters_to_merge = masters_to_merge or DEFAULT_MASTERS_TO_MERGE
num_rows_to_merge = num_rows_to_merge or 25
console_data = ConsoleData()
surroundings = get_and_cache_pagedata(
'%s/console/surroundings' % masters_to_merge[0])
merged_page = BeautifulSoup(surroundings['content'])
merged_tag = merged_page.find('table', 'ConsoleData')
if merged_tag is None:
msg = 'console_merger("%s", "%s", "%s"): merged_tag cannot be None.' % (
localpath, remoteurl, page_data)
logging.error(msg)
raise Exception(msg)
latest_rev = int(get_and_cache_rowdata('latest_rev')['rev_number'])
if not latest_rev:
logging.error('console_merger(\'%s\', \'%s\', \'%s\'): cannot get latest '
'revision number.' % (
localpath, remoteurl, page_data))
return
fetch_timestamp = datetime.datetime.now()
for master in masters_to_merge:
# Fetch the summary one-box-per-builder for the master.
# If we don't get it, something is wrong, skip the master entirely.
master_summary = get_and_cache_pagedata('%s/console/summary' % master)
if not master_summary['content']:
continue
console_data.SawMaster(master)
# Get the categories for this builder. If the builder doesn't have any
# categories, just use the default empty-string category.
category_list = []
master_categories = get_and_cache_pagedata('%s/console/categories' % master)
if not master_categories['content']:
category_list.append('')
else:
category_row = BeautifulSoup(master_categories['content'])
category_list = [c.text for c in category_row.findAll('td', 'DevStatus')]
# Get the corresponding summary box(es).
summary_row = BeautifulSoup(master_summary['content'])
summary_list = summary_row.findAll('table')
for category, summary in zip(category_list, summary_list):
console_data.AddCategory(category, summary)
# Fetch all of the rows that we need.
rows_fetched = 0
revs_skipped = 0
current_rev = latest_rev
while rows_fetched < num_rows_to_merge and current_rev >= 0:
# Don't get stuck looping backwards forever into data we don't have.
# How hard we try scales with how many rows the person wants.
if revs_skipped > max(num_rows_to_merge, 10):
break
row_data = get_and_cache_rowdata('%s/console/%s' % (master, current_rev))
if not row_data:
current_rev -= 1
revs_skipped += 1
continue
console_data.AddRow(row_data)
current_rev -= 1
revs_skipped = 0
rows_fetched += 1
# Convert the merged content into console content.
console_data.Finish()
template_environment = Environment()
template_environment.loader = FileSystemLoader('.')
def notstarted(builder_status):
"""Convert a BeautifulSoup Tag from builder status to a notstarted line."""
builder_status = re.sub(r'DevSlaveBox', 'DevStatusBox', str(builder_status))
builder_status = re.sub(r'class=\'([^\']*)\' target=',
'class=\'DevStatusBox notstarted\' target=',
builder_status)
builder_status = re.sub(r'class="([^"]*)" target=',
'class="DevStatusBox notstarted" target=',
builder_status)
return builder_status
template_environment.filters['notstarted'] = notstarted
merged_template = template_environment.from_string(console_template)
merged_console = merged_template.render(data=console_data)
# For debugging:
# logging.info('%r' % merged_console)
# import code
# code.interact(local=locals())
# Place merged console at |merged_tag|'s location in |merged_page|, and put
# the result in |merged_content|.
merged_tag.replaceWith(merged_console)
merged_content = utf8_convert(merged_page)
merged_content = re.sub(
r'\'\<a href="\'', '\'<a \' + attributes + \' href="\'', merged_content)
merged_content = re.sub(
r'\'\<table\>\'', r"'<table ' + attributes + '>'", merged_content)
merged_content = re.sub(
r'\'\<div\>\'', r"'<div ' + attributes + '>'", merged_content)
merged_content = re.sub(
r'\'\<td\>\'', r"'<td ' + attributes + '>'", merged_content)
merged_content = re.sub(
r'\<iframe\>\</iframe\>',
'<iframe \' + attributes + \' src="\' + url + \'"></iframe>',
merged_content)
# Update the merged console page.
merged_page = get_or_create_page(localpath, None, maxage=30)
logging.info('console_merger: saving merged console')
page_data = get_and_cache_pagedata(localpath)
page_data['title'] = 'BuildBot: Chromium'
page_data['offsite_base'] = 'http://build.chromium.org/p/chromium'
page_data['body_class'] = 'interface'
page_data['content'] = merged_content
save_page(merged_page, localpath, fetch_timestamp, page_data)
return
def parse_master(localpath, remoteurl, page_data=None):
"""Part of the new pipeline to store individual rows rather than
whole pages of html. Parses the master data into a set of rows,
and writes them out to the datastore in an easily retrievable format.
Doesn't modify page_data dict.
"""
ts = datetime.datetime.now()
page_data = page_data or {}
content = page_data.get('content')
if not content:
return page_data
content = content.decode('utf-8', 'replace')
# Split page into surroundings (announce, legend, footer) and data (rows).
surroundings = BeautifulSoup(content)
data = surroundings.find('table', 'ConsoleData')
if data is None:
raise Exception('parse_master: data can not be None')
new_data = Tag(surroundings, 'table', [('class', 'ConsoleData'),
('width', '96%')])
data.replaceWith(new_data)
surroundings_page = get_or_create_page(localpath + '/surroundings',
None, maxage=30)
surroundings_data = {}
surroundings_data['title'] = 'Surroundings for ' + localpath
surroundings_data['content'] = utf8_convert(surroundings)
save_page(surroundings_page, localpath + '/surroundings', ts,
surroundings_data)
rows = data.findAll('tr', recursive=False)
# The first table row can be special: the list of categories.
categories = None
# If the first row contains a DevStatus cell...
if rows[0].find('td', 'DevStatus') != None:
# ...extract it into the categories...
categories = rows[0]
# ...and get rid of the next (spacer) row too.
rows = rows[2:]
if categories:
category_page = get_or_create_page(localpath + '/categories',
None, maxage=30)
category_data = {}
category_data['title'] = 'Categories for ' + localpath
category_data['content'] = utf8_convert(categories)
save_page(category_page, localpath + '/categories', ts, category_data)
# The next table row is special, it's the summary one-box-per-builder.
summary = rows[0]
rows = rows[1:]
summary_page = get_or_create_page(localpath + '/summary', None, maxage=30)
summary_data = {}
summary_data['title'] = 'Summary for ' + localpath
summary_data['content'] = utf8_convert(summary)
save_page(summary_page, localpath + '/summary', ts, summary_data)
curr_row = {}
# Each table row is either a status row with a revision, name, and status,
# a comment row with the commit message, a details row with flakiness info,
# or a spacer row (in which case we finalize the row and save it).
for row in rows:
if row.find('td', 'DevComment'):
curr_row['comment'] = ''.join(utf8_convert(tag).strip()
for tag in row.td.contents)
elif row.find('td', 'DevDetails'):
curr_row['details'] = ''.join(utf8_convert(tag).strip()
for tag in row.td.contents)
elif row.find('td', 'DevStatus'):
curr_row['rev'] = ''.join(utf8_convert(tag).strip()
for tag in row.find('td', 'DevRev').contents)
curr_row['name'] = ''.join(utf8_convert(tag).strip()
for tag in row.find('td', 'DevName').contents)
curr_row['status'] = ''.join(utf8_convert(box.table).strip()
for box in row.findAll('td', 'DevStatus'))
else:
if 'details' not in curr_row:
curr_row['details'] = ''
curr_row['fetch_timestamp'] = ts
curr_row['rev_number'] = get_position_number(curr_row['comment'])
save_row(curr_row, localpath + '/' + curr_row['rev_number'])
curr_row = {}
return page_data
def console_merger(localpath, remoteurl, page_data,
masters_to_merge=None, num_rows_to_merge=None):
masters_to_merge = masters_to_merge or DEFAULT_MASTERS_TO_MERGE
num_rows_to_merge = num_rows_to_merge or 25
console_data = ConsoleData()
surroundings = get_and_cache_pagedata(
'%s/console/surroundings' % masters_to_merge[0])
merged_page = BeautifulSoup(surroundings['content'])
merged_tag = merged_page.find('table', 'ConsoleData')
if merged_tag is None:
msg = 'console_merger("%s", "%s", "%s"): merged_tag cannot be None.' % (
localpath, remoteurl, page_data)
logging.error(msg)
raise Exception(msg)
latest_rev = int(get_and_cache_rowdata('latest_rev')['rev_number'])
if not latest_rev:
logging.error('console_merger(\'%s\', \'%s\', \'%s\'): cannot get latest '
'revision number.' % (
localpath, remoteurl, page_data))
return
fetch_timestamp = datetime.datetime.now()
for master in masters_to_merge:
# Fetch the summary one-box-per-builder for the master.
# If we don't get it, something is wrong, skip the master entirely.
master_summary = get_and_cache_pagedata('%s/console/summary' % master)
if not master_summary['content']:
continue
console_data.SawMaster(master)
# Get the categories for this builder. If the builder doesn't have any
# categories, just use the default empty-string category.
category_list = []
master_categories = get_and_cache_pagedata('%s/console/categories' % master)
if not master_categories['content']:
category_list.append('')
else:
category_row = BeautifulSoup(master_categories['content'])
category_list = [c.text for c in category_row.findAll('td', 'DevStatus')]
# Get the corresponding summary box(es).
summary_row = BeautifulSoup(master_summary['content'])
summary_list = summary_row.findAll('table')
for category, summary in zip(category_list, summary_list):
console_data.AddCategory(category, summary)
# Fetch all of the rows that we need.
rows_fetched = 0
revs_skipped = 0
current_rev = latest_rev
while rows_fetched < num_rows_to_merge and current_rev >= 0:
# Don't get stuck looping backwards forever into data we don't have.
# How hard we try scales with how many rows the person wants.
if revs_skipped > max(num_rows_to_merge, 10):
break
row_data = get_and_cache_rowdata('%s/console/%s' % (master, current_rev))
if not row_data:
current_rev -= 1
revs_skipped += 1
continue
console_data.AddRow(row_data)
current_rev -= 1
revs_skipped = 0
rows_fetched += 1
# Convert the merged content into console content.
console_data.Finish()
template_environment = Environment()
template_environment.loader = FileSystemLoader('.')
def notstarted(builder_status):
"""Convert a BeautifulSoup Tag from builder status to a notstarted line."""
builder_status = re.sub(r'DevSlaveBox', 'DevStatusBox', str(builder_status))
builder_status = re.sub(r'class=\'([^\']*)\' target=',
'class=\'DevStatusBox notstarted\' target=',
builder_status)
builder_status = re.sub(r'class="([^"]*)" target=',
'class="DevStatusBox notstarted" target=',
builder_status)
return builder_status
template_environment.filters['notstarted'] = notstarted
merged_template = template_environment.from_string(console_template)
merged_console = merged_template.render(data=console_data)
# For debugging:
# logging.info('%r' % merged_console)
# import code
# code.interact(local=locals())
# Place merged console at |merged_tag|'s location in |merged_page|, and put
# the result in |merged_content|.
merged_tag.replaceWith(merged_console)
merged_content = utf8_convert(merged_page)
merged_content = re.sub(
r'\'\<a href="\'', '\'<a \' + attributes + \' href="\'', merged_content)
merged_content = re.sub(
r'\'\<table\>\'', r"'<table ' + attributes + '>'", merged_content)
merged_content = re.sub(
r'\'\<div\>\'', r"'<div ' + attributes + '>'", merged_content)
merged_content = re.sub(
r'\'\<td\>\'', r"'<td ' + attributes + '>'", merged_content)
merged_content = re.sub(
r'\<iframe\>\</iframe\>',
'<iframe \' + attributes + \' src="\' + url + \'"></iframe>',
merged_content)
# Update the merged console page.
merged_page = get_or_create_page(localpath, None, maxage=30)
logging.info('console_merger: saving merged console')
page_data = get_and_cache_pagedata(localpath)
page_data['title'] = 'BuildBot: Chromium'
page_data['offsite_base'] = 'http://build.chromium.org/p/chromium'
page_data['body_class'] = 'interface'
page_data['content'] = merged_content
save_page(merged_page, localpath, fetch_timestamp, page_data)
return
