def update_pages(self):
bower = Content.get_by_id('bower', parent=self.version_key)
if bower is None:
return
bower_json = bower.get_json()
for _, path in bower_json.get('pages', {}).iteritems():
response = util.github_get('repos', self.owner, self.repo, 'contents/' + path, params={'ref': self.sha})
if response.status_code == 200:
response_json = json.loads(response.content)
markdown = None
# Ensure a file was returned
if isinstance(response_json, dict) and response_json.get('type') == 'file':
markdown = base64.b64decode(response_json.get('content'))
elif response.status_code == 404:
markdown = None
else:
return self.retry('error fetching page %s (%d)' % (path, response.status_code))
if markdown is not None:
response = util.github_markdown(markdown)
if response.status_code == 200:
Content(parent=self.version_key, id='page-' + path, content=response.content,
status=Status.ready, etag=response.headers.get('ETag', None)).put()
else:
return self.retry('error converting page to markdown %s (%d)' % (path, response.status_code))
def update_pages(self):
bower = Content.get_by_id('bower', parent=self.version_key)
if bower is None:
return
bower_json = bower.get_json()
for _, path in bower_json.get('pages', {}).iteritems():
response = util.github_get('repos', self.owner, self.repo, 'contents/' + path, params={'ref': self.sha})
if response.status_code == 200:
response_json = json.loads(response.content)
markdown = None
# Ensure a file was returned
if isinstance(response_json, dict) and response_json.get('type') == 'file':
markdown = base64.b64decode(response_json.get('content'))
elif response.status_code == 404:
markdown = None
else:
return self.retry('error fetching page %s (%d)' % (path, response.status_code))
if markdown is not None:
response = util.github_markdown(markdown)
if response.status_code == 200:
Content(parent=self.version_key, id='page-' + path, content=response.content,
status=Status.ready, etag=response.headers.get('ETag', None)).put()
else:
return self.retry('error converting page to markdown %s (%d)' % (path, response.status_code))
def test_analyze_leaves_existing_content_when_reanalyzing(self):
library_key = Library(id='owner/repo').put()
version_key = Version(id='v1.1.1',
parent=library_key,
sha='sha',
status='ready').put()
content = Content(id='analysis',
parent=version_key,
status=Status.pending)
content.content = 'existing data'
content.status = Status.ready
content.put()
response = self.app.get('/task/analyze/owner/repo',
headers={'X-AppEngine-QueueName': 'default'})
self.assertEqual(response.status_int, 200)
content = Content.get_by_id('analysis', parent=version_key)
self.assertEqual(content.content, 'existing data')
self.assertEqual(content.status, Status.ready)
tasks = self.tasks.get_filtered_tasks()
self.assertEqual([
util.ingest_analysis_task('owner', 'repo', 'v1.1.1'),
], [task.url for task in tasks])
def handle_post(self):
message_json = json.loads(
urllib.unquote(self.request.body).rstrip('='))
message = message_json['message']
data = base64.b64decode(str(message['data']))
attributes = message['attributes']
owner = attributes['owner']
repo = attributes['repo']
version = attributes['version']
error = attributes.get('error', None)
version_key = ndb.Key(Library, Library.id(owner, repo), Version,
version)
content = Content.get_by_id('analysis', parent=version_key)
if content is None:
return
if data == '':
content.set_json(None)
else:
content.set_json(json.loads(data))
if error is None:
content.status = Status.ready
content.error = None
else:
content.status = Status.error
content.error = error
content.put()
if version_key.id() == Library.default_version_for_key_async(
version_key.parent()).get_result():
task_url = util.update_indexes_task(owner, repo)
util.new_task(task_url, target='manage')
def update_search_index(self, owner, repo, version_key, library, bower):
metadata = json.loads(library.metadata)
registry_metadata = json.loads(library.registry_metadata) if library.registry_metadata else None
npm_description = registry_metadata.get('description', '') if registry_metadata else ''
npm_keywords = registry_metadata.get('keywords', []) if registry_metadata else []
fields = [
search.AtomField(name='owner', value=owner),
search.TextField(name='repo', value=repo),
search.AtomField(name='kind', value=library.kind),
search.AtomField(name='version', value=version_key.id()),
search.TextField(name='github_description', value=metadata.get('description', '')),
search.TextField(name='bower_description', value=bower.get('description', '')),
search.TextField(name='npm_description', value=npm_description),
search.TextField(name='bower_keywords', value=' '.join(bower.get('keywords', []))),
search.TextField(name='npm_keywords', value=' '.join(npm_keywords)),
search.TextField(name='prefix_matches', value=' '.join(util.generate_prefixes_from_list(
util.safe_split_strip(metadata.get('description')) + util.safe_split_strip(bower.get('description')) +
util.safe_split_strip(repo)))),
]
# Generate weighting field
weights = [(repo, 10)]
analysis = Content.get_by_id('analysis', parent=version_key)
if analysis is not None and analysis.status == Status.ready:
data = analysis.get_json()
if data.get('analyzerData', None) is not None:
# Use analyzer data for search index
element_objects = data.get('analyzerData', {}).get('elements', [])
elements = [element.get('tagname', '') or element.get('classname', '') for element in element_objects]
if elements != []:
fields.append(search.TextField(name='element', value=' '.join(elements)))
weights.append((' '.join(elements), 5))
behavior_objects = data.get('analyzerData', {}).get('metadata', {}).get('polymer', {}).get('behaviors', [])
behaviors = [behavior.get('name', '') for behavior in behavior_objects]
if behaviors != []:
fields.append(search.TextField(name='behavior', value=' '.join(behaviors)))
weights.append((' '.join(behaviors), 5))
else:
# Use hydrolysis data for search index
elements = data.get('elementsByTagName', {}).keys()
if elements != []:
fields.append(search.TextField(name='element', value=' '.join(elements)))
weights.append((' '.join(elements), 5))
behaviors = data.get('behaviorsByName', {}).keys()
if behaviors != []:
fields.append(search.TextField(name='behavior', value=' '.join(behaviors)))
weights.append((' '.join(behaviors), 5))
weighted = []
for value, weight in weights:
for _ in range(0, weight):
weighted.append(value)
fields.append(search.TextField(name='weighted_fields', value=' '.join(weighted)))
rank = int((library.updated - datetime.datetime(2016, 1, 1)).total_seconds())
document = search.Document(doc_id=Library.id(owner, repo), fields=fields, rank=rank)
index = search.Index('repo')
index.put(document)
def get(self, owner, repo, ver=None):
self.response.headers['Access-Control-Allow-Origin'] = '*'
owner = owner.lower()
repo = repo.lower()
library_key = ndb.Key(Library, Library.id(owner, repo))
if ver is None:
ver = yield Library.latest_version_for_key_async(library_key)
if ver is None:
self.response.set_status(404)
return
version_key = ndb.Key(Library, Library.id(owner, repo), Version, ver)
analysis = Content.get_by_id('analysis',
parent=version_key,
read_policy=ndb.EVENTUAL_CONSISTENCY)
if analysis is None:
self.response.set_status(404)
return
self.response.headers['Content-Type'] = 'application/json'
result = {}
result['status'] = analysis.status
if analysis.status == Status.ready:
result['content'] = json.loads(analysis.content)
if analysis.status == Status.error:
result['error'] = analysis.error
if result['status'] != Status.ready:
self.response.set_status(400)
self.response.headers['Content-Type'] = 'application/json'
self.response.write(json.dumps(result))
def get(self, owner, repo, tag, name=None, path=None):
self.response.headers['Access-Control-Allow-Origin'] = '*'
self.response.headers['Content-Type'] = 'application/json'
owner = owner.lower()
repo = repo.lower()
version_key = ndb.Key(Library, '%s/%s' % (owner, repo), Version, tag)
analysis = Content.get_by_id('analysis', parent=version_key, read_policy=ndb.EVENTUAL_CONSISTENCY)
if analysis is None:
self.response.set_status(404)
return
dependencies = json.loads(analysis.content).get('bowerDependencies', None)
if dependencies is None:
self.response.set_status(404)
return
config_map = {}
for dependency in dependencies:
config_map[dependency['name']] = '%s/%s/%s' % (dependency['owner'], dependency['repo'], dependency['version'])
# Ensure the repo serves its own version.
config_map[repo] = '%s/%s/%s' % (owner, repo, tag)
def resolve(name, path):
return 'https://cdn.rawgit.com/%s%s' % (config_map[name], path) if name in config_map else None
# debug mode
if name is None or path is None:
for k in config_map:
self.response.write('/%s/%s/%s/components/%s/... -> %s\n' % (owner, repo, tag, k, resolve(k, '/...')))
self.response.write('\n')
return
resolved = resolve(name, path)
if resolved is None:
self.response.write('%s is not a valid dependency for %s/%s#%s' % (name, owner, repo, tag))
self.response.set_status(400)
return
# TODO: Figure out what other types this is necessary for. eg. do we need it for CSS @import?
# We need to serve html files from the same origin, so that relative urls load correctly.
if path.endswith('.html'):
# TODO: Decide whether this should be memcached. Appengine's urlfetch already does caching.
response = urlfetch.fetch(resolved)
if response.status_code == 200:
self.response.write(response.content)
self.response.headers['cache-control'] = response.headers.get('cache-control', 'max-age=315569000')
self.response.headers['content-type'] = response.headers.get('content-type', 'text/html')
else:
self.response.write('could not fetch: %s' % resolved)
self.response.set_status(400)
else:
self.response.set_status(301)
self.response.headers['Location'] = str(resolved)
self.response.headers['cache-control'] = 'max-age=315569000'
def update_search_index(self, owner, repo, version_key, library, bower):
metadata = json.loads(library.metadata)
fields = [
search.AtomField(name='owner', value=owner),
search.TextField(name='repo', value=repo),
search.AtomField(name='kind', value=library.kind),
search.AtomField(name='version', value=version_key.id()),
search.TextField(name='github_description',
value=metadata.get('description', '')),
search.TextField(name='bower_description',
value=bower.get('description', '')),
search.TextField(name='bower_keywords',
value=' '.join(bower.get('keywords', []))),
search.TextField(
name='prefix_matches',
value=' '.join(
util.generate_prefixes_from_list(
util.safe_split_strip(metadata.get('description')) +
util.safe_split_strip(bower.get('description')) +
util.safe_split_strip(repo)))),
]
# Generate weighting field
weights = [(repo, 10)]
analysis = Content.get_by_id('analysis', parent=version_key)
if analysis is not None and analysis.status == Status.ready:
analysis = json.loads(analysis.content)
elements = analysis.get('elementsByTagName', {}).keys()
if elements != []:
fields.append(
search.TextField(name='element', value=' '.join(elements)))
weights.append((' '.join(elements), 5))
behaviors = analysis.get('behaviorsByName', {}).keys()
if behaviors != []:
fields.append(
search.TextField(name='behavior',
value=' '.join(behaviors)))
weights.append((' '.join(behaviors), 5))
weighted = []
for value, weight in weights:
for _ in range(0, weight):
weighted.append(value)
fields.append(
search.TextField(name='weighted_fields', value=' '.join(weighted)))
rank = int(
(library.updated - datetime.datetime(2016, 1, 1)).total_seconds())
document = search.Document(doc_id=Library.id(owner, repo),
fields=fields,
rank=rank)
index = search.Index('repo')
index.put(document)
def trigger_analysis(self, tag, sha, transactional=False):
analysis_sha = None
if self.library.kind == 'collection':
analysis_sha = sha
version_key = ndb.Key(Library, self.library.key.id(), Version, tag)
content = Content.get_by_id('analysis', parent=version_key)
if content is None or content.status == Status.error:
Content(id='analysis', parent=version_key, status=Status.pending).put()
task_url = util.ingest_analysis_task(self.scope, self.package, tag, analysis_sha)
util.new_task(task_url, target='analysis', transactional=transactional, queue_name='analysis')
def trigger_analysis(self, tag, sha, transactional=False):
analysis_sha = None
if self.library.kind == 'collection':
analysis_sha = sha
version_key = ndb.Key(Library, self.library.key.id(), Version, tag)
content = Content.get_by_id('analysis', parent=version_key)
if content is None or content.status == Status.error:
Content(id='analysis', parent=version_key, status=Status.pending).put()
task_url = util.ingest_analysis_task(self.owner, self.repo, tag, analysis_sha)
util.new_task(task_url, target='analysis', transactional=transactional, queue_name='analysis')
def test_analyze(self):
library_key = Library(id='owner/repo').put()
version_key = Version(id='v1.1.1', parent=library_key, sha='sha', status='ready').put()
response = self.app.get('/task/analyze/owner/repo', headers={'X-AppEngine-QueueName': 'default'})
self.assertEqual(response.status_int, 200)
content = Content.get_by_id('analysis', parent=version_key)
self.assertEqual(content.content, None)
self.assertEqual(content.status, Status.pending)
tasks = self.tasks.get_filtered_tasks()
self.assertEqual([
util.ingest_analysis_task('owner', 'repo', 'v1.1.1'),
], [task.url for task in tasks])
def get(self, owner, repo, ver=None):
use_analyzer_data = self.request.get('use_analyzer_data',
None) is not None
self.response.headers['Access-Control-Allow-Origin'] = '*'
owner = owner.lower()
repo = repo.lower()
library_key = ndb.Key(Library, Library.id(owner, repo))
if ver is None:
ver = yield Library.default_version_for_key_async(library_key)
if ver is None:
self.response.set_status(404)
return
version_key = ndb.Key(Library, Library.id(owner, repo), Version, ver)
analysis = Content.get_by_id('analysis',
parent=version_key,
read_policy=ndb.EVENTUAL_CONSISTENCY)
if analysis is None:
self.response.set_status(404)
return
self.response.headers['Content-Type'] = 'application/json'
result = {}
result['status'] = analysis.status
if analysis.status == Status.ready:
content = analysis.get_json()
has_analyzer_data = content.get('analyzerData', None) is not None
if use_analyzer_data and has_analyzer_data:
# Use the analyzer data fields
result['analysis'] = content['analyzerData']
else:
# Use the hydrolysis fields and delete the analyzer ones
if has_analyzer_data:
del content['analyzerData']
result['content'] = content
if analysis.status == Status.error:
result['error'] = analysis.error
if result['status'] != Status.ready:
self.response.set_status(400)
self.response.headers['Content-Type'] = 'application/json'
self.response.write(json.dumps(result))
def test_analyze_latest(self):
library_key = Library(id='owner/repo').put()
Version(id='v1.1.1', parent=library_key, sha='sha', status='ready').put()
version_key = Version(id='v1.1.2', parent=library_key, sha='sha', status='ready').put()
VersionCache.update(library_key)
response = self.app.get('/task/analyze/owner/repo/True', headers={'X-AppEngine-QueueName': 'default'})
self.assertEqual(response.status_int, 200)
content = Content.get_by_id('analysis', parent=version_key)
self.assertEqual(content.get_json(), None)
self.assertEqual(content.status, Status.pending)
tasks = self.tasks.get_filtered_tasks()
self.assertEqual([
util.ingest_analysis_task('owner', 'repo', 'v1.1.2'),
], [task.url for task in tasks])
def get(self, owner, repo, ver, path):
self.response.headers['Access-Control-Allow-Origin'] = '*'
version_key = ndb.Key(Library, Library.id(owner, repo), Version, ver)
if version_key is None:
self.response.set_status(404)
self.response.write('Invalid repo/version')
return
page = Content.get_by_id('page-' + path, parent=version_key, read_policy=ndb.EVENTUAL_CONSISTENCY)
if page is None:
self.response.set_status(404)
self.response.write('Cannot find page %s' % path)
return
self.response.write(page.content)
def get(self, owner, repo, ver=None):
self.response.headers['Access-Control-Allow-Origin'] = '*'
owner = owner.lower()
repo = repo.lower()
version_key = ndb.Key(Library, '%s/%s' % (owner, repo), Version, ver)
hydrolyzer = Content.get_by_id('hydrolyzer', parent=version_key, read_policy=ndb.EVENTUAL_CONSISTENCY)
if hydrolyzer is None:
self.response.set_status(404)
return
dependencies = json.loads(hydrolyzer.content).get('bowerDependencies', None)
if dependencies is None:
self.response.set_status(404)
return
self.response.headers['Content-Type'] = 'application/json'
self.response.write(json.dumps(dependencies))
def update_search_index(self, owner, repo, version_key, library, bower):
metadata = json.loads(library.metadata)
fields = [
search.AtomField(name='owner', value=owner),
search.TextField(name='repo', value=repo),
search.AtomField(name='kind', value=library.kind),
search.AtomField(name='version', value=version_key.id()),
search.TextField(name='github_description', value=metadata.get('description', '')),
search.TextField(name='bower_description', value=bower.get('description', '')),
search.TextField(name='bower_keywords', value=' '.join(bower.get('keywords', []))),
search.TextField(name='prefix_matches', value=' '.join(util.generate_prefixes_from_list(
util.safe_split_strip(metadata.get('description')) + util.safe_split_strip(bower.get('description')) +
util.safe_split_strip(repo)))),
]
# Generate weighting field
weights = [(repo, 10)]
analysis = Content.get_by_id('analysis', parent=version_key)
if analysis is not None and analysis.status == Status.ready:
analysis = json.loads(analysis.content)
elements = analysis.get('elementsByTagName', {}).keys()
if elements != []:
fields.append(search.TextField(name='element', value=' '.join(elements)))
weights.append((' '.join(elements), 5))
behaviors = analysis.get('behaviorsByName', {}).keys()
if behaviors != []:
fields.append(search.TextField(name='behavior', value=' '.join(behaviors)))
weights.append((' '.join(behaviors), 5))
weighted = []
for value, weight in weights:
for _ in range(0, weight):
weighted.append(value)
fields.append(search.TextField(name='weighted_fields', value=' '.join(weighted)))
rank = int((library.updated - datetime.datetime(2016, 1, 1)).total_seconds())
document = search.Document(doc_id=Library.id(owner, repo), fields=fields, rank=rank)
index = search.Index('repo')
index.put(document)
def update_search_index(self, owner, repo, version_key, library, bower):
metadata = json.loads(library.metadata)
fields = [
search.AtomField(name='owner', value=owner),
search.TextField(name='repo', value=repo),
search.AtomField(name='kind', value=library.kind),
search.AtomField(name='version', value=version_key.id()),
search.TextField(name='github_description',
value=metadata.get('description', '')),
search.TextField(name='bower_description',
value=bower.get('description', '')),
search.TextField(name='bower_keywords',
value=' '.join(bower.get('keywords', []))),
search.TextField(
name='prefix_matches',
value=' '.join(
util.generate_prefixes_from_list(
[repo] + util.safesplit(metadata.get('description')) +
util.safesplit(bower.get('description')) +
repo.replace("_", " ").replace("-", " ").split()))),
]
analysis = Content.get_by_id('analysis', parent=version_key)
if analysis is not None and analysis.status == Status.ready:
analysis = json.loads(analysis.content)
elements = analysis.get('elementsByTagName', {}).keys()
if elements != []:
fields.append(
search.TextField(name='element', value=' '.join(elements)))
behaviors = analysis.get('behaviorsByName', {}).keys()
if behaviors != []:
fields.append(
search.TextField(name='behavior',
value=' '.join(behaviors)))
document = search.Document(doc_id=Library.id(owner, repo),
fields=fields)
index = search.Index('repo')
index.put(document)
def get(self, owner, repo, ver=None):
# TODO: Share all of this boilerplate between GetDataMeta and GetHydroData
self.response.headers['Access-Control-Allow-Origin'] = '*'
owner = owner.lower()
repo = repo.lower()
library_key = ndb.Key(Library, '%s/%s' % (owner, repo))
# TODO: version shouldn't be optional here
if ver is None:
versions = Version.query(ancestor=library_key).map(lambda v: v.key.id())
versions.sort(versiontag.compare)
if versions == []:
self.response.set_status(404)
return
ver = versions[-1]
version_key = ndb.Key(Library, '%s/%s' % (owner, repo), Version, ver)
hydro = Content.get_by_id('hydrolyzer', parent=version_key, read_policy=ndb.EVENTUAL_CONSISTENCY)
if hydro is None:
self.response.set_status(404)
return
self.response.headers['Content-Type'] = 'application/json'
self.response.write(hydro.content)
def handle_post(self):
# Ignore payloads larger than 5 MB.
if len(self.request.body) > 1048487 * 5:
return
message_json = json.loads(urllib.unquote(self.request.body).rstrip('='))
message = message_json['message']
data = base64.b64decode(str(message['data']))
attributes = message['attributes']
if len(attributes) == 0:
logging.error(message)
return
owner = attributes['owner']
repo = attributes['repo']
version = attributes['version']
error = attributes.get('error', None)
version_key = ndb.Key(Library, Library.id(owner, repo), Version, version)
content = Content.get_by_id('analysis', parent=version_key)
if content is None:
return
if data == '':
content.set_json(None)
else:
content.set_json(json.loads(data))
if error is None:
content.status = Status.ready
content.error = None
else:
content.status = Status.error
content.error = error
content.put()
if version_key.id() == Library.default_version_for_key_async(version_key.parent()).get_result():
task_url = util.update_indexes_task(owner, repo)
util.new_task(task_url, target='manage')
def handle_post(self):
message_json = json.loads(urllib.unquote(self.request.body).rstrip('='))
message = message_json['message']
data = base64.b64decode(str(message['data']))
attributes = message['attributes']
owner = attributes['owner']
repo = attributes['repo']
version = attributes['version']
error = attributes.get('error', None)
version_key = ndb.Key(Library, Library.id(owner, repo), Version, version)
content = Content.get_by_id('analysis', parent=version_key)
if content is None:
return
if data == '':
content.content = None
elif len(data) > 500000:
# Max entity size is only 1MB.
logging.error('content was too large: %d %s %s', len(data), Library.id(owner, repo), version)
error = 'content was too large: %d' % len(data)
else:
content.content = data
if error is None:
content.status = Status.ready
content.error = None
else:
content.status = Status.error
content.error = error
content.put()
if version_key.id() == Library.default_version_for_key_async(version_key.parent()).get_result():
task_url = util.update_indexes_task(owner, repo)
util.new_task(task_url, target='manage')
def update_search_index(self, owner, repo, version_key, library, bower):
metadata = json.loads(library.metadata)
registry_metadata = json.loads(library.registry_metadata) if library.registry_metadata else None
npm_description = registry_metadata.get('description', '') if registry_metadata else ''
npm_keywords = registry_metadata.get('keywords', []) if registry_metadata else []
fields = [
search.AtomField(name='owner', value=owner),
search.AtomField(name='github_owner', value=library.github_owner),
search.TextField(name='repo', value=repo),
search.AtomField(name='kind', value=library.kind),
search.AtomField(name='version', value=version_key.id()),
search.TextField(name='github_description', value=metadata.get('description', '')),
search.TextField(name='bower_description', value=bower.get('description', '')),
search.TextField(name='npm_description', value=npm_description),
search.TextField(name='bower_keywords', value=' '.join(bower.get('keywords', []))),
search.TextField(name='npm_keywords', value=' '.join(npm_keywords)),
search.TextField(name='prefix_matches', value=' '.join(util.generate_prefixes_from_list(
util.safe_split_strip(metadata.get('description')) + util.safe_split_strip(bower.get('description')) +
util.safe_split_strip(repo)))),
]
# Generate weighting field
weights = [(repo, 10)]
analysis = Content.get_by_id('analysis', parent=version_key)
if analysis is not None and analysis.status == Status.ready:
data = analysis.get_json()
if data.get('analyzerData', None) is not None:
# Use analyzer data for search index
element_objects = data.get('analyzerData', {}).get('elements', [])
elements = [element.get('tagname', '') or element.get('classname', '') for element in element_objects]
if elements != []:
fields.append(search.TextField(name='element', value=' '.join(elements)))
weights.append((' '.join(elements), 5))
behavior_objects = data.get('analyzerData', {}).get('metadata', {}).get('polymer', {}).get('behaviors', [])
behaviors = [behavior.get('name', '') for behavior in behavior_objects]
if behaviors != []:
fields.append(search.TextField(name='behavior', value=' '.join(behaviors)))
weights.append((' '.join(behaviors), 5))
else:
# Use hydrolysis data for search index
elements = data.get('elementsByTagName', {}).keys()
if elements != []:
fields.append(search.TextField(name='element', value=' '.join(elements)))
weights.append((' '.join(elements), 5))
behaviors = data.get('behaviorsByName', {}).keys()
if behaviors != []:
fields.append(search.TextField(name='behavior', value=' '.join(behaviors)))
weights.append((' '.join(behaviors), 5))
weighted = []
for value, weight in weights:
for _ in range(0, weight):
weighted.append(value)
fields.append(search.TextField(name='weighted_fields', value=' '.join(weighted)))
rank = int((library.updated - datetime.datetime(2016, 1, 1)).total_seconds())
document = search.Document(doc_id=Library.id(owner, repo), fields=fields, rank=rank)
index = search.Index('repo')
index.put(document)
def get(self, owner, repo, ver=None):
owner = owner.lower()
repo = repo.lower()
library = Library.get_by_id('%s/%s' % (owner, repo), read_policy=ndb.EVENTUAL_CONSISTENCY)
if library is None or library.error is not None:
self.response.write(str(library))
self.response.set_status(404)
return
versions = library.versions()
if ver is None:
ver = versions[-1]
version = Version.get_by_id(ver, parent=library.key, read_policy=ndb.EVENTUAL_CONSISTENCY)
if version is None or version.error is not None:
self.response.write(str(version))
self.response.set_status(404)
return
metadata = json.loads(library.metadata)
dependencies = []
bower = Content.get_by_id('bower', parent=version.key, read_policy=ndb.EVENTUAL_CONSISTENCY)
if bower is not None:
try:
bower_json = json.loads(bower.content)
# TODO: Which exception is this for?
# pylint: disable=bare-except
except:
bower_json = {}
readme = Content.get_by_id('readme.html', parent=version.key, read_policy=ndb.EVENTUAL_CONSISTENCY)
full_name_match = re.match(r'(.*)/(.*)', metadata['full_name'])
result = {
'version': ver,
'versions': versions,
'readme': None if readme is None else readme.content,
'subscribers': metadata['subscribers_count'],
'stars': metadata['stargazers_count'],
'forks': metadata['forks'],
'contributors': library.contributor_count,
'open_issues': metadata['open_issues'],
'updated_at': metadata['updated_at'],
'owner': full_name_match.groups()[0],
'repo': full_name_match.groups()[1],
'bower': None if bower is None else {
'description': bower_json.get('description', ''),
'license': bower_json.get('license', ''),
'dependencies': bower_json.get('dependencies', []),
'keywords': bower_json.get('keywords', []),
},
'collections': []
}
for collection in library.collections:
if not versiontag.match(ver, collection.semver):
continue
collection_version = collection.version.id()
collection_library = collection.version.parent().get()
collection_metadata = json.loads(collection_library.metadata)
collection_name_match = re.match(r'(.*)/(.*)', collection_metadata['full_name'])
result['collections'].append({
'owner': collection_name_match.groups()[0],
'repo': collection_name_match.groups()[1],
'version': collection_version
})
if library.kind == 'collection':
dependencies = []
version_futures = []
for dep in version.dependencies:
parsed_dep = Dependency.fromString(dep)
dep_key = ndb.Key(Library, "%s/%s" % (parsed_dep.owner.lower(), parsed_dep.repo.lower()))
version_futures.append(Library.versions_for_key_async(dep_key))
for i, dep in enumerate(version.dependencies):
parsed_dep = Dependency.fromString(dep)
versions = version_futures[i].get_result()
versions.reverse()
while len(versions) > 0 and not versiontag.match(versions[0], parsed_dep.version):
versions.pop()
if len(versions) == 0:
dependencies.append({
'error': 'unsatisfyable dependency',
'owner': parsed_dep.owner,
'repo': parsed_dep.repo,
'versionSpec': parsed_dep.version
})
else:
dependencies.append(brief_metadata_from_datastore(parsed_dep.owner, parsed_dep.repo, versions[0]))
result['dependencies'] = dependencies
self.response.headers['Access-Control-Allow-Origin'] = '*'
self.response.headers['Content-Type'] = 'application/json'
self.response.write(json.dumps(result))
def get(self, owner, repo, tag, name=None, path=None):
self.response.headers['Access-Control-Allow-Origin'] = '*'
self.response.headers['Content-Type'] = 'application/json'
if name is None and path is None:
self.response.write(
'Invalid request. Try using a relative path if you are using an absolute path.'
)
self.response.set_status(400)
return
owner = owner.lower()
repo = repo.lower()
version_key = ndb.Key(Library, '%s/%s' % (owner, repo), Version, tag)
if path.endswith('/'):
path = path + 'index.html'
analysis = Content.get_by_id('analysis',
parent=version_key,
read_policy=ndb.EVENTUAL_CONSISTENCY)
if analysis is None or analysis.status != Status.ready:
self.response.write('could not find analysis for %s in %s/%s' %
(tag, owner, repo))
self.response.set_status(404)
return
dependencies = analysis.get_json().get('bowerDependencies', None)
if dependencies is None:
self.response.write('could not find dependencies for %s in %s/%s' %
(tag, owner, repo))
self.response.set_status(404)
return
config_map = {}
for dependency in dependencies:
if dependency['owner'] == owner and dependency['repo'] == repo:
continue
config_map[dependency['name']] = '%s/%s/%s' % (
dependency['owner'], dependency['repo'], dependency['version'])
# Ensure the repo serves its own version.
config_map[repo] = '%s/%s/%s' % (owner, repo, tag)
def resolve(name, path):
return 'https://cdn.rawgit.com/%s%s' % (
config_map[name], path) if name in config_map else None
# debug mode
if name is None or path is None:
for k in config_map:
self.response.write('/%s/%s/%s/components/%s/... -> %s\n' %
(owner, repo, tag, k, resolve(k, '/...')))
self.response.write('\n')
return
resolved = resolve(name, path)
if resolved is None:
self.response.write('%s is not a valid dependency for %s/%s#%s' %
(name, owner, repo, tag))
self.response.set_status(400)
return
# TODO: Figure out what other types this is necessary for. eg. do we need it for CSS @import?
# We need to serve html files from the same origin, so that relative urls load correctly.
if path.endswith('.html'):
# TODO: Decide whether this should be memcached. Appengine's urlfetch already does caching.
response = urlfetch.fetch(resolved, validate_certificate=True)
if response.status_code == 200:
self.response.write(response.content)
self.response.headers['cache-control'] = response.headers.get(
'cache-control', 'max-age=315569000')
self.response.headers['content-type'] = response.headers.get(
'content-type', 'text/html')
else:
self.response.write('could not fetch: %s' % resolved)
self.response.set_status(400)
else:
self.response.set_status(301)
self.response.headers['Location'] = str(resolved)
self.response.headers['cache-control'] = 'max-age=315569000'
