def add_repo(project_id):
data = request.get_json()
project = Project.query.get(project_id)
if project is None:
return jsonify(message="Project not found"), 404
if not project.course.is_user(current_user.id):
return jsonify(message="Unauthorized to add project"), 403
repo = project.get_repo_for_user(current_user.id)
if repo:
return jsonify(message="Already added repo for project"), 400
repo = Repo(project_id=project.id, name=data['repo_name'])
repo.user_id = current_user.id
if project.type is not ProjectTypes.IND:
g = project.get_group_for_user(current_user.id)
if g is None:
return jsonify(message="Not part of group"), 400
repo.group_id = g.id
webhook = add_webhook(repo.name, current_user.oauth_token)
if 'id' not in webhook:
return jsonify(error="Could not create webhook", data=webhook), 401
repo.webhook_id = webhook['id']
db.session.add(repo)
db.session.commit()
return jsonify(repo=ps.RepoSchema().dump(repo)), 201
def test_add_topics(self):
aggregator = Aggregator()
aggregator.add(Repo(False, 10, "python", ["Flask"]))
self.assertEqual(aggregator.topics, {"flask": 1})
aggregator.add(Repo(False, 10, "python", ["flask"]))
self.assertEqual(aggregator.topics, {"flask": 2})
aggregator.add(Repo(False, 10, "python", ["flask", "numpy"]))
self.assertEqual(aggregator.topics, {"flask": 3, "numpy": 1})
def test_add_languages(self):
aggregator = Aggregator()
aggregator.add(Repo(False, 10, "Python", ["Flask"]))
self.assertEqual(aggregator.languages, {"python": 1})
aggregator.add(Repo(False, 10, "python", ["Flask"]))
self.assertEqual(aggregator.languages, {"python": 2})
aggregator.add(Repo(False, 10, "java", ["Flask"]))
self.assertEqual(aggregator.languages, {"python": 2, "java": 1})
aggregator.add(Repo(False, 10, None, ["Flask"]))
self.assertEqual(aggregator.languages, {"python": 2, "java": 1})
def create(self, request, *args, **kwargs):
""" Generates db repo object """
user = request.data['user']
path = request.data['path']
version = request.data['version']
res = RepoManager(user).create()
if not version:
version = res
r = Repo(user=user, path=path, version=version)
r.save()
rs = RepoSerializer(r)
return Response(rs.data)
def create(self, request, *args, **kwargs):
""" Generates db repo object """
user = request.data['user']
path = request.data['path']
version = request.data['version']
res = RepoManager(user).create()
if not version:
version = res
r = Repo(user=user, path=path, version=version)
r.save()
rs = RepoSerializer(r)
return Response(rs.data)
def test_asdict(self):
aggregator = Aggregator()
expected = {
"public_repo_count": 0,
"public_repo_breakdown": {
"original_repo_count": 0,
"forked_repo_count": 0
},
"watcher_count": 0,
"languages": [],
"topics": []
}
self.assertEqual(aggregator.asdict(), expected)
expected2 = {
"public_repo_count": 1,
"public_repo_breakdown": {
"original_repo_count": 1,
"forked_repo_count": 0
},
"watcher_count": 10,
"languages": [{
"name": "python",
"count": 1
}],
"topics": [{
"name": "flask",
"count": 1
}]
}
aggregator.add(Repo(True, 10, "python", ["flask"]))
self.assertEqual(aggregator.asdict(), expected2)
def set_random_day_for_monthy_job():
for repo in Repo.select().where((Repo.random_job_day == None)):
repo.random_job_day = random.randint(1, 28)
task_logger.info(
f"set random day for monthly job of repo '{repo.name}' at '{repo.random_job_day}'"
)
repo.save()
def __init__(self, sample_dict=None, class_names=None):
global _sample
if self.feature_name is None:
raise Exception('Provide feature_name field in subclass.')
if sample_dict is None:
if _sample is None:
_sample = Repo.load_sample(separate=True)
sample_dict = _sample
if class_names is None:
class_names = sorted(classes)
self.class_names = class_names
self.figure, self.ax = plt.subplots()
hist_data = []
for clsname in class_names:
# ie 'class feature data'
cfd = np.array([getattr(repo, self.feature_name) for repo in sample_dict[clsname]])
if self.in_class_percentiles is not None:
min_val, max_val = [np.percentile(cfd, i) for i in self.in_class_percentiles]
cfd = np.array([e for e in cfd if min_val < e < max_val])
hist_data.append(cfd)
self.hist_data = hist_data
def get(self, username, reponame):
try:
repo = (Repo.select().join(User).alias("user").where(
(User.name == username) & (Repo.name == reponame)).get())
title = repo.user.name + "/" + repo.name
timemap = self.get_query_argument("timemap", "false") == "true"
datetime = self.get_query_argument("datetime", None)
key = self.get_query_argument("key", None)
if key and not timemap:
self.render("repo/memento.html",
repo=repo,
key=key,
datetime=datetime)
elif key and timemap:
self.render("repo/history.html", repo=repo, key=key)
else:
cs = (CSet.select(fn.distinct(
CSet.hkey)).where(CSet.repo == repo).limit(5).alias("cs"))
samples = (HMap.select(HMap.val).join(
cs, on=(HMap.sha == cs.c.hkey_id)))
self.render("repo/show.html",
title=title,
repo=repo,
samples=list(samples))
except Repo.DoesNotExist:
raise HTTPError(404)
def get(self, username, reponame):
try:
repo = (Repo.select().join(User).alias("user")
.where((User.name == username) & (Repo.name == reponame))
.get())
title = repo.user.name + "/" + repo.name
timemap = self.get_query_argument("timemap", "false") == "true"
datetime = self.get_query_argument("datetime", None)
key = self.get_query_argument("key", None)
if key and not timemap:
self.render("repo/memento.html", repo=repo, key=key,
datetime=datetime)
elif key and timemap:
self.render("repo/history.html", repo=repo, key=key)
else:
cs = (CSet.select(fn.distinct(CSet.hkey))
.where(CSet.repo == repo).limit(5).alias("cs"))
samples = (HMap.select(HMap.val)
.join(cs, on=(HMap.sha == cs.c.hkey_id)))
self.render("repo/show.html", title=title, repo=repo,
samples=list(samples))
except Repo.DoesNotExist:
raise HTTPError(404)
def get(self, username, reponame):
try:
repo = (Repo.select().join(User).alias("user")
.where((User.name == username) & (Repo.name == reponame))
.get())
title = repo.user.name + "/" + repo.name
timemap = self.get_query_argument("timemap", "false") == "true"
datetime = self.get_query_argument("datetime", None)
key = self.get_query_argument("key", None)
index = self.get_query_argument("index", "false") == "true"
if self.get_query_argument("datetime", None):
datestr = self.get_query_argument("datetime")
try:
ts = date(datestr, QSDATEFMT)
except ValueError:
raise HTTPError(reason="Invalid format of datetime param", status_code=400)
elif "Accept-Datetime" in self.request.headers:
datestr = self.request.headers.get("Accept-Datetime")
ts = date(datestr, RFC1123DATEFMT)
else:
ts = now()
if key and not timemap:
chain = revision_logic.get_chain_at_ts(repo, key, ts)
# use ts of cset instead of now(), to make prev work
if len(chain) != 0:
ts = chain[-1].time
cs_prev = revision_logic.get_cset_prev_before_ts(repo, key, ts)
cs_next = revision_logic.get_cset_next_after_ts(repo, key, ts)
if cs_prev:
cs_prev_str = self.request.protocol + "://" + self.request.host + self.request.path + "?key=" + key + "&datetime=" + cs_prev.time.strftime(QSDATEFMT)
else:
cs_prev_str = ""
if cs_next:
cs_next_str = self.request.protocol + "://" + self.request.host + self.request.path + "?key=" + key + "&datetime=" + cs_next.time.strftime(QSDATEFMT)
else:
cs_next_str = ""
commit_message = revision_logic.get_commit_message(repo, key, ts)
self.render("repo/memento.html", repo=repo, key=key, datetime=datetime, cs_next_str=cs_next_str, cs_prev_str=cs_prev_str, commit_message=commit_message)
elif key and timemap:
self.render("repo/history.html", repo=repo, key=key)
elif index:
cs = (CSet.select(fn.distinct(CSet.hkey)).where((CSet.repo == repo) & (CSet.time <= ts)).alias("cs"))
key_count = (HMap.select(HMap.val).join(cs, on=(HMap.sha == cs.c.hkey_id))).count()
page = int(self.get_query_argument("page", "1"))
hm = revision_logic.get_repo_index(repo, ts, page)
self.render("repo/index.html", repo=repo, title=title, key_count=key_count, page_size=revision_logic.INDEX_PAGE_SIZE, hm=hm, current_page=page)
else:
hm = list(revision_logic.get_repo_index(repo, ts, 1, 5))
# cs = (CSet.select(fn.distinct(CSet.hkey)).where(CSet.repo == repo).limit(5).alias("cs"))
# samples = (HMap.select(HMap.val).join(cs, on=(HMap.sha == cs.c.hkey_id)))
self.render("repo/show.html", title=title, repo=repo, hm=hm)
except Repo.DoesNotExist:
raise HTTPError(reason="Repo not found.", status_code=404)
async def html_app(request, app_name):
app = Repo.select().where(Repo.name == app_name)
if app.count == 0:
raise NotFound()
return {"app": app[0], 'relative_path_to_root': '../../', 'path': request.path}
async def launch_monthly_job():
today = date.today().day
for repo in Repo.select().where(Repo.random_job_day == today):
task_logger.info(
f"Launch monthly job for {repo.name} on day {today} of the month ")
await create_job(repo.name, repo.url)
def test_repo_commits(self):
repo = Repo.create(**self.TEST_REPO)
repo.save()
response = self.fetch(self.get_app().reverse_url('view', repo.id))
self.assertIn(self.MESSAGES['no_records'], response.body.decode())
self.assertIn(self.MESSAGES['get_more'], response.body.decode())
for commit in range(self.TEST_COUNT):
commit_data = self.TEST_COMMIT
commit_data.update({'repo': repo})
c = Commit(**commit_data)
c.save()
response = self.fetch(self.get_app().reverse_url('view', repo.id))
self.assertEqual(response.body.decode().count(self.TEST_COMMIT['message']),
self.TEST_COUNT)
self.assertIn(self.MESSAGES['get_more'], response.body.decode())
repo.next_page = None
repo.save()
response = self.fetch(self.get_app().reverse_url('view', repo.id))
self.assertNotIn(self.MESSAGES['get_more'], response.body.decode())
def reconcile_db_with_gh(*args, **kwargs):
ghc = GitHubConnector()
issues = ghc.get_all_issues()
repos = ghc.get_all_repos()
for repo in repos:
r = Repo(github_id=repo.id, name=repo.name)
r.save()
for issue in issues:
i = Issue(github_id=issue.id)
i.title = issue.title
i.number = issue.number
i.repo = Repo.objects.get(name=issue.repository[1])
i.save()
print "Not only did your task run successfully, but you're damned good looking too."
def add_repository(url, comment, user_id):
repo = Repo(url=url,
timestamp=datetime.datetime.now(),
user_id=user_id,
comment=comment)
db.session.add(repo)
db.session.commit()
return repo
def get(self, username, reponame):
try:
repo = Repo.select().join(User).alias("user").where((User.name == username) & (Repo.name == reponame)).get()
if not repo.private:
self._get(repo)
else:
self._getAuth(repo)
except Repo.DoesNotExist:
raise HTTPError(reason="Repo not found.", status_code=404)
def post(self):
reponame = self.get_argument("reponame", None)
desc = self.get_argument("description", None)
user = self.current_user
if not reponame:
self.redirect(self.reverse_url("web:create-repo"))
return
repo = Repo.create(user=user, name=reponame, desc=desc)
self.redirect(self.reverse_url("web:repo", user.name, repo.name))
def post(self):
reponame = self.get_argument("reponame", None)
desc = self.get_argument("description", None)
user = self.current_user
if not reponame:
self.redirect(self.reverse_url("web:create-repo"))
return
repo = Repo.create(user=user, name=reponame, desc=desc)
self.redirect(self.reverse_url("web:repo", user.name, repo.name))
def test_init(self):
is_original = True
watcher_count = 10
language = "Python"
topics = ["Flask"]
repo = Repo(is_original, watcher_count, language, topics)
self.assertEqual(repo.is_original, is_original)
self.assertEqual(repo.watcher_count, watcher_count)
self.assertEqual(repo.language, language)
self.assertEqual(repo.topics, topics)
def __init__(self, username, repo_name):
self.repo = Repo(repo_path(username, repo_name))
self.seen_files = {}
try:
print "Repo exists"
self.repo_model = RepoModel.objects.get(username = username, name = repo_name)
except RepoModel.DoesNotExist:
self.repo_model = RepoModel(username = username, name = repo_name)
self.repo_model.save()
self.cached_data = {}
def get(self):
query = tornado.escape.url_unescape(self.get_argument("q", ""))
if query:
pattern = "%" + query + "%"
repos = Repo.select().join(User).alias("user").where(Repo.name ** pattern, Repo.private == False)
users = User.select().where(User.name ** pattern)
else:
repos = []
users = []
self.render("search/show.html", query=query, repos=repos, users=users)
def change_status(target_repo, state):
from models import Repo
# class Repo(Document):
# url = StringField(max_length=100,default='Not set yet')
# last_used = DateTimeField(default=datetime.now())
# created_on = DateTimeField(default=datetime.now())
# monitoring = StringField(max_length=100,default='Not set yet')
# state = StringField(max_length=50,default='Ready')
# owner = StringField(max_length=50,default='no')
if not use_database:
return ''
try:
repo = Repo.objects.get(url=target_repo)
repo.last_used = datetime.today()
repo.state = state
repo.owner = parent_folder
repo.save()
except DoesNotExist:
repo = Repo()
repo.url=target_repo
repo.state = state
repo.owner = parent_folder
repo.save()
except Exception as e:
print 'database_exception: '+str(e)
def get(self):
query = tornado.escape.url_unescape(self.get_argument("q", ""))
if query:
pattern = "%" + query + "%"
repos = (Repo.select().join(User).alias("user").where(
Repo.name**pattern))
users = User.select().where(User.name**pattern)
else:
repos = []
users = []
self.render("search/show.html", query=query, repos=repos, users=users)
async def launch_monthly_job(type):
# XXX DRY
job_command_last_part = ""
if type == "arm":
job_command_last_part = " (~ARM~)"
elif type == "testing-unstable":
job_command_last_part = [" (testing)", " (unstable)"]
today = date.today().day
for repo in Repo.select().where(Repo.random_job_day == today):
task_logger.info(f"Launch monthly job for {repo.name} on day {today} of the month ")
await create_job(repo.name, repo.app_list, repo, job_command_last_part)
async def ws_app(request, websocket, app_name):
# XXX I don't check if the app exists because this websocket is supposed to
# be only loaded from the app page which does this job already
app = Repo.select().where(Repo.name == app_name)[0]
subscribe(websocket, f"app-jobs-{app.url}")
await websocket.send(ujson.dumps({
"action": "init_jobs",
"data": Job.select().where(Job.url_or_path ==
app.url).order_by(-Job.id),
}))
await websocket.wait_closed()
async def html_job(request, job_id):
job = Job.select().where(Job.id == job_id)
if job.count == 0:
raise NotFound()
job = job[0]
app = Repo.select().where(Repo.url == job.url_or_path)
app = app[0] if app else None
return {
"job": job,
'app': app,
'relative_path_to_root': '../',
'path': request.path
}
def delete(self, username, reponame):
# Check whether the key exists and if maybe the last change already is
# a delete, else insert a `CSet.DELETE` entry without any blob data.
key = self.get_query_argument("key")
if username != self.current_user.name:
raise HTTPError(403)
if not key:
raise HTTPError(400)
datestr = self.get_query_argument("datetime", None)
ts = datestr and date(datestr, QSDATEFMT) or now()
try:
repo = (Repo.select(Repo.id).join(
User).where((User.name == username)
& (Repo.name == reponame)).naive().get())
except Repo.DoesNotExist:
raise HTTPError(404)
sha = shasum(key.encode("utf-8"))
try:
last = (CSet.select(
CSet.time,
CSet.type).where((CSet.repo == repo)
& (CSet.hkey == sha)).order_by(
CSet.time.desc()).limit(1).naive().get())
except CSet.DoesNotExist:
# No changeset was found for the given key -
# the resource does not exist.
raise HTTPError(400)
if not ts > last.time:
# Appended timestamps must be monotonically increasing!
raise HTTPError(400)
if last.type == CSet.DELETE:
# The resource was deleted already, return instantly.
return self.finish()
# Insert the new "delete" change.
CSet.create(repo=repo, hkey=sha, time=ts, type=CSet.DELETE, len=0)
def test_commits_create(self, get_commits):
fixtures = self.data.buffer.read()
jsn = json.loads(fixtures.decode())
request = HTTPRequest(self.TEST_REPO['href'])
response = HTTPResponse(request, client.OK, buffer=io.BytesIO(fixtures))
future = Future()
future.set_result(response)
get_commits.return_value = future
body = {'href': self.TEST_REPO['href']}
response = self.fetch(self.get_app().reverse_url('create'), method='POST',
body=urlencode(body).encode())
self.assertIn('всего {}'.format(len(jsn)), response.body.decode())
self.assertEqual(len(jsn), Commit.select().count())
self.assertEqual(1, Repo.select().count())
def change_status(target_repo, state):
from models import Repo
if not use_database:
return ''
try:
repo = Repo.objects.get(url=target_repo)
repo.last_used = datetime.today()
repo.state = state
repo.owner = parent_folder
repo.save()
except DoesNotExist:
repo = Repo()
repo.url = target_repo
repo.state = state
repo.owner = parent_folder
repo.save()
except Exception as e:
print 'database_exception: ' + str(e)
def get_samples():
"""Return a {'class': [reponames]}."""
repos = Repo.load()
fetch_dates = [datetime.datetime(*(r.fetch_ymd)) for r in repos]
print 'number of repos:', len(repos)
latest_fetch = max(fetch_dates)
print 'fetched between %s and %s' % (min(fetch_dates), latest_fetch)
print
filtered = [r for r in repos if
30720 > r.size > 0 and # not foolproof to avoid big repos
r.stars > 1 and
not r.fork and
not 'dotfile' in r.name.lower() and
not 'sublime' in r.name.lower() # avoid SublimeText config
]
print 'after noise filter:', len(filtered)
filtered = [r for r in filtered if
((latest_fetch - r.creation_date) >
datetime.timedelta(30))
]
print 'exluding very new:', len(filtered)
filtered = [r for r in filtered if
r.stars > 5 and
classes.score(r) > (1 / 30)
]
print 'exluding very unpopular:', len(filtered)
score_pairs = [(classes.score(r), r) for r in filtered]
score_pairs.sort(key=lambda x: x[0])
# top 1k, bottom 1k.
return {'high': [r.name for (score, r) in score_pairs[-1000:]],
'low': [r.name for (score, r) in score_pairs[:1000]],
}
def get(self, username):
try:
user = User.select().where(User.name == username).get()
except User.DoesNotExist:
raise HTTPError(reason="User not found.", status_code=404)
repos = Repo.select().where(Repo.user == user)
reposit = repos.iterator()
# TODO: Paginate?
first = None
try:
first = reposit.next()
except StopIteration:
# No repos for user
# No need to raise an error, just return empty list in json
pass
accept = self.request.headers.get("Accept", "")
user_url = (self.request.protocol + "://" + self.request.host)
if "application/json" in accept or "*/*" in accept:
self.set_header("Content-Type", "application/json")
self.write('{"username": '******', "repositories": {"list":[')
m = ('{{"name": "{0}", "uri": "' + user_url +
'/'+username+'/{0}"}}')
if first:
self.write(m.format(first.name))
for repo in reposit:
self.write(', ' + m.format(repo.name))
self.write(']}')
self.write('}')
def cargar_datos(self):
"""Parsea o contido dun ficheiro JSON."""
datos_json = self.cargar_json()
for k, v in datos_json.items():
if (k == "usuario"):
usuario = v
self.usuario = UsuarioGit(usuario["usuario"], usuario["email"])
elif (k == "repos"):
for item in v:
aux = item
repo = Repo(
aux["nome"],
aux["rama"],
aux["remoto"],
aux["uri"],
aux["directorio"],
aux["ten_submodulos"],
aux["e_submodulo"]
)
self.add_repo(repo)
def change_status(target_repo, state):
from models import Repo
if not use_database:
return ''
try:
repo = Repo.objects.get(url=target_repo)
repo.last_used = datetime.today()
repo.state = state
repo.owner = parent_folder
repo.save()
except DoesNotExist:
repo = Repo()
repo.url = target_repo
repo.state = state
repo.owner = parent_folder
repo.save()
except Exception as e:
print 'database_exception: ' + str(e)
def test_add_none(self):
aggregator = Aggregator()
aggregator.add(Repo(None, 10, "Python", ["Flask"]))
self.assertEqual(aggregator.count, 1)
self.assertEqual(aggregator.original_count, 0)
self.assertEqual(aggregator.forked_count, 0)
def RunWithPars(pars, uid):
startTime = Common.getCurrentTimeMil()
reposFol = 'SavePath/Repos/'
if not os.path.exists(config.basedir + '/' + reposFol):
os.makedirs(config.basedir + '/' + reposFol)
fname = reposFol + Common.gen_rnd_filename()
os.makedirs(config.basedir + '/' + fname)
##ADD Field
e = Repo()
e.cloneFinishDate = "--RUNNING--"
e.cloneStartDate = str(startTime)
e.repoInfo = ''
e.isPrivate = int(pars['isPrivate'])
e.path = fname
e.repoName = pars['repoName']
e.url = pars['url']
e.userId = uid
db.session.add(e)
db.session.commit()
try:
porcelain.clone(pars['url'], config.basedir + '/' + fname)
endTime = Common.getCurrentTimeMil()
e.cloneFinishDate = str(endTime)
db.session.commit()
except Exception as ex:
print(ex)
e.delete()
db.session.commit()
import classes
from models import Repo
import utils
def get_classifier(X, y):
return RandomForestClassifier(
n_estimators=100, max_depth=None, min_samples_split=1,
random_state=0, # random seed is static for comparison
compute_importances=True,
)
if __name__ == '__main__':
repos = Repo.load_sample()
class_to_id, id_to_class = utils.create_bimap(classes.classes)
dict_repos = []
for r in repos:
d = {mod: False for mod in utils.stdlib_module_names()}
for mod in r.imported_stdlib_modules:
d[mod] = True
dict_repos.append(d)
vectorizer = DictVectorizer(sparse=False)
y = np.array([class_to_id[classes.classify(r)] for r in repos])
X = vectorizer.fit_transform(dict_repos)
def load_data(args):
start_time = time.time()
users = {}
repos = {}
popular_repos = []
superprojects = {}
print "Loading user to repo map"
data = open(args[0], 'r')
for line in data.readlines():
user_id, repo_id = line.strip().split(':')
user_id = int(user_id)
repo_id = int(repo_id)
if user_id in users:
user = users[user_id]
else:
user = User(user_id)
users[user_id] = user
if repo_id in repos:
repo = repos[repo_id]
else:
repo = Repo(repo_id)
repos[repo_id] = repo
repo.is_watched_by(user)
user.is_watching(repo)
data.close()
print "Ordering repos by popularity"
popular_repos = sorted(repos.values(), reverse=True,
key=lambda x: x.popularity)
owners = {}
print "Reading repo details"
repo_txt = open(args[1], 'r')
for line in repo_txt.readlines():
id, other = line.strip().split(':')
id = int(id)
if id not in repos:
continue
parts = other.split(',')
repo = repos[id]
owner, repo.name = parts[0].split('/', 1)
repo.creation_date = parts[1]
if owner in owners:
repo.owner = owners[owner]
else:
repo.owner = RepoOwner(owner)
owners[owner] = repo.owner
repo.owner.owns.add(repo)
if len(parts) > 2 and int(parts[2]) in repos:
repo.forked_from = repos[int(parts[2])]
repo.forked_from.forked_by.append(repo)
repo_txt.close()
# print "Grouping superprojects"
# superproject_keys = ['gnome', 'django', 'ruby', 'perl', 'rails']
# for repo in repos.values():
# for key in superproject_keys:
# if key in repo.name.lower():
# if key not in superprojects:
# superprojects[key] = []
# superprojects[key].append(repo)
print "Reading repo language"
lang = open(args[2], 'r')
for line in lang.readlines():
id, other = line.strip().split(':')
id = int(id)
if id not in repos:
continue
parts = other.split(',')
repo = repos[id]
for part in parts:
lang_name, count = part.split(';')
repo.langs.append((lang_name, int(count)))
lang.close()
print "Data read in %d seconds" % (time.time() - start_time)
return users, repos, popular_repos, superprojects
for class_name in class_names:
repos = class_map[class_name]
feature_data = [getattr(repo, feature_name) for repo in repos]
class_summaries.append([f(feature_data) for f in funcs])
feature_summaries[feature_name] = np.array(class_summaries)
for feature_name, summary in feature_summaries.items():
print feature_name
for i, class_name in enumerate(class_names):
print ' ', class_name
print ' ', '\t'.join(str(e) for e in summary[i])
print '-----'
print
if __name__ == '__main__':
class_map = Repo.load_sample(separate=True)
summarize_features(class_map, ['with_stmt_usage',
'compr_usage',
'lambda_usage',
'global_usage',
'gen_exp_usage',
'print_usage',
], sorted(classes))
#summarize_imports(class_map)
def workflow(changed_files, repo_str):
global current_repo, current_user
current_repo = Repo(name=repo_str, user=current_user, status="starting", started_at=datetime.now(), progress=10)
current_repo.save()
dolog("deleting the repo if exists")
change_status("preparation", 10)
delete_forked_repo(repo_str)
time.sleep(60) # for GitHub to avoid the unknown issues
dolog("forking the repo %s" % repo_str)
change_status("forking", 20)
repo_url = fork_repo(repo_str)
dolog("cloning the repo %s" % repo_url)
change_status("cloning", 30)
clone_repo(repo_url)
# fork_cleanup()
change_status("updating the fork", 40)
update_fork(repo_str) # update from upstream as the cloned repo is an old fork due to Github limitation
dolog("getting jar configurations")
change_status("getting configurations", 50)
target_files, jar_command = get_jar_config(os.path.join(get_repo_abs_path(), 'jar.cfg'))
if target_files is None or jar_command is None:
dolog("get jar config failed")
change_status("Error getting configurations", 100)
delete_local_copy()
return "get jar config failed"
else:
change_status("configuration parsed", 60)
dolog("running if target")
change_status("running if target", 70)
is_found, msg = run_if_target(changed_files, target_files, jar_command)
dolog("after running")
if is_found:
dolog("is found")
change_status("pushing changes", 80)
push_changes()
dolog("after pushing the changes")
change_status("creating pull request", 90)
if create_pull_request(repo_str):
dolog("pull request is True")
change_status("pull request created", 100)
current_repo.completed_at = datetime.now()
current_repo.save()
msg += " And pull request is created"
dolog("deleting the forked repo attempt")
else:
dolog("pull request is False")
change_status("pull failed to be created", 100)
current_repo.completed_at = datetime.now()
current_repo.save()
msg += " And pull request failed to be created"
else:
change_status("not found", 100)
dolog("not found")
current_repo.completed_at = datetime.now()
current_repo.save()
return msg
async def api_list_app(request):
query = Repo.select()
return response.json([model_to_dict(x) for x in query.order_by(Repo.name)])
async def ws_apps(request, websocket):
subscribe(websocket, "jobs")
subscribe(websocket, "apps")
# I need to do this because peewee strangely f**k up on join and remove the
# subquery fields which breaks everything
repos = Repo.raw('''
SELECT
"id",
"name",
"url",
"revision",
"state",
"random_job_day",
"job_id",
"job_name",
"job_state",
"created_time",
"started_time",
"end_time"
FROM
"repo" AS "t1"
INNER JOIN (
SELECT
"t1"."id" as "job_id",
"t1"."name" as "job_name",
"t1"."url_or_path",
"t1"."state" as "job_state",
"t1"."created_time",
"t1"."started_time",
"t1"."end_time"
FROM
"job" AS "t1"
INNER JOIN (
SELECT
Max("t2"."id") AS "max_id"
FROM
"job" AS "t2"
GROUP BY
"t2"."url_or_path"
)
AS
"t3"
ON
("t1"."id" = "t3"."max_id")
) AS
"t5"
ON
("t5"."url_or_path" = "t1"."url")
ORDER BY
"name"
''')
repos = [{
"id":
x.id,
"name":
x.name,
"url":
x.url,
"revision":
x.revision,
"state":
x.state,
"random_job_day":
x.random_job_day,
"job_id":
x.job_id,
"job_name":
x.job_name,
"job_state":
x.job_state,
"created_time":
datetime.strptime(x.created_time.split(".")[0], '%Y-%m-%d %H:%M:%S')
if x.created_time else None,
"started_time":
datetime.strptime(x.started_time.split(".")[0], '%Y-%m-%d %H:%M:%S')
if x.started_time else None,
"end_time":
datetime.strptime(x.end_time.split(".")[0], '%Y-%m-%d %H:%M:%S')
if x.end_time else None,
} for x in repos]
# add apps without jobs
selected_repos = {x["id"] for x in repos}
for repo in Repo.select().where(Repo.id.not_in(selected_repos)):
repos.append({
"id": repo.id,
"name": repo.name,
"url": repo.url,
"revision": repo.revision,
"state": repo.state,
"random_job_day": repo.random_job_day,
"job_id": None,
"job_name": None,
"job_state": None,
"created_time": None,
"started_time": None,
"end_time": None,
})
repos = sorted(repos, key=lambda x: x["name"])
await websocket.send(ujson.dumps({
"action": "init_apps",
"data": repos,
}))
await websocket.wait_closed()
def calculate(f_to_calc, f_to_overwrite, console, download):
"""Calculate a list of features."""
sys.stdout.write('loading')
sys.stdout.flush()
repos = Repo.load_sample()
seen = 0
total = len(repos)
dl_failures = []
calc_failures = []
last_write = datetime.datetime.now()
if f_to_calc or f_to_overwrite or download:
for repo in repos:
seen += 1
success = True
if download:
success = utils.clone(repo)
if not success:
dl_failures.append(repo)
continue
try:
if f_to_calc:
logging.info("calc: %s", repo)
repo.calculate_features(f_to_calc)
if f_to_overwrite:
logging.info("calc: %s", repo)
repo.calculate_features(f_to_overwrite, overwrite=True)
repo._clear_support_features() # we're done with this repo now
except:
print # from status line
logging.exception("!problem: %s", repo)
calc_failures.append(repo)
print
progress_bar(seen, total)
since_write = datetime.datetime.now() - last_write
if since_write > datetime.timedelta(minutes=5):
sys.stdout.write("\r(writing results)")
sys.stdout.flush()
Repo.write_update(repos)
last_write = datetime.datetime.now()
print # from progress bar line
if dl_failures:
print "%s failed to download:" % len(dl_failures)
for f in dl_failures:
print " %s" % f
print
if calc_failures:
print "%s failed during calc:" % len(calc_failures)
for f in calc_failures:
print " %s" % f
print
if console:
message = ('`repos` contains results;\n'
'use ^d to write out or `exit()` to cancel')
code.interact(message, local=locals())
print 'writing out...'
Repo.write_update(repos)
async def monitor_apps_lists(monitor_git=False,
monitor_only_good_quality_apps=False):
"parse apps lists every hour or so to detect new apps"
# only support github for now :(
async def get_master_commit_sha(url):
command = await asyncio.create_subprocess_shell(
f"git ls-remote {url} master",
stdout=asyncio.subprocess.PIPE,
stderr=asyncio.subprocess.PIPE)
data = await command.stdout.read()
commit_sha = data.decode().strip().replace("\t", " ").split(" ")[0]
return commit_sha
async with aiohttp.ClientSession() as session:
task_logger.info(f"Downloading applist...")
async with session.get(APPS_LIST) as resp:
data = await resp.json()
data = data["apps"]
repos = {x.name: x for x in Repo.select()}
for app_id, app_data in data.items():
commit_sha = await get_master_commit_sha(app_data["git"]["url"])
if app_data["state"] != "working":
task_logger.debug(
f"skip {app_id} because state is {app_data['state']}")
continue
if monitor_only_good_quality_apps:
if app_data.get("level") in [None, "?"] or app_data["level"] <= 4:
task_logger.debug(
f"skip {app_id} because app is not good quality")
continue
# already know, look to see if there is new commits
if app_id in repos:
repo = repos[app_id]
# but first check if the URL has changed
if repo.url != app_data["git"]["url"]:
task_logger.info(
f"Application {app_id} has changed of url from {repo.url} to {app_data['git']['url']}"
)
repo.url = app_data["git"]["url"]
repo.save()
await broadcast(
{
"action": "update_app",
"data": model_to_dict(repo),
}, "apps")
# change the url of all jobs that used to have this URL I
# guess :/
# this isn't perfect because that could overwrite added by
# hand jobs but well...
for job in Job.select().where(Job.url_or_path == repo.url,
Job.state == "scheduled"):
job.url_or_path = repo.url
job.save()
task_logger.info(
f"Updating job {job.name} #{job.id} for {app_id} to {repo.url} since the app has changed of url"
)
await broadcast(
{
"action": "update_job",
"data": model_to_dict(job),
}, [
"jobs", f"job-{job.id}",
f"app-jobs-{job.url_or_path}"
])
# we don't want to do anything else
if not monitor_git:
continue
repo_is_updated = False
if repo.revision != commit_sha:
task_logger.info(
f"Application {app_id} has new commits on github "
f"({repo.revision} → {commit_sha}), schedule new job")
repo.revision = commit_sha
repo.save()
repo_is_updated = True
await create_job(app_id, repo.url)
repo_state = "working" if app_data[
"state"] == "working" else "other_than_working"
if repo.state != repo_state:
repo.state = repo_state
repo.save()
repo_is_updated = True
if repo.random_job_day is None:
repo.random_job_day = random.randint(1, 28)
repo.save()
repo_is_updated = True
if repo_is_updated:
await broadcast(
{
"action": "update_app",
"data": model_to_dict(repo),
}, "apps")
# new app
elif app_id not in repos:
task_logger.info(f"New application detected: {app_id} " +
(", scheduling a new job" if monitor_git else ""))
repo = Repo.create(
name=app_id,
url=app_data["git"]["url"],
revision=commit_sha,
state="working"
if app_data["state"] == "working" else "other_than_working",
random_job_day=random.randint(1, 28),
)
await broadcast(
{
"action": "new_app",
"data": model_to_dict(repo),
}, "apps")
if monitor_git:
await create_job(app_id, repo.url)
await asyncio.sleep(1)
# delete apps removed from the list
unseen_repos = set(repos.keys()) - set(data.keys())
for repo_name in unseen_repos:
repo = repos[repo_name]
# delete scheduled jobs first
task_logger.info(
f"Application {repo_name} has been removed from the app list, start by removing its scheduled job if there are any..."
)
for job in Job.select().where(Job.url_or_path == repo.url,
Job.state == "scheduled"):
await api_stop_job(None, job.id) # not sure this is going to work
job_id = job.id
task_logger.info(
f"Delete scheduled job {job.name} #{job.id} for application {repo_name} because the application is being deleted."
)
data = model_to_dict(job)
job.delete_instance()
await broadcast({
"action": "delete_job",
"data": data,
}, ["jobs", f"job-{job_id}", f"app-jobs-{job.url_or_path}"])
task_logger.info(
f"Delete application {repo_name} because it has been removed from the apps list."
)
data = model_to_dict(repo)
repo.delete_instance()
await broadcast({
"action": "delete_app",
"data": data,
}, "apps")
def create_new_repo(url,branch):
repo = Repo(url = url, branch = branch)
repo.save()
return repo
#benchmark(PassiveAggressiveClassifier(n_iter=50), X, y, feature_names)
#print 'kNN'
#benchmark(KNeighborsClassifier(n_neighbors=10), X, y, feature_names)
#print 'SGD'
#benchmark(SGDClassifier(n_jobs=-1, alpha=.0001, n_iter=np.ceil(10**3), penalty="elasticnet", shuffle=True),
# X, y, feature_names)
#print 'nearest centroid'
#benchmark(NearestCentroid(), X, y, feature_names)
#print 'naive bayes'
#benchmark(MultinomialNB(alpha=.01), X, y, feature_names)
#print 'naive bayes (bernoulli)'
#benchmark(BernoulliNB(alpha=.01), X, y, feature_names)
#classify(X, y, id_to_class, vec)
# classify(select_by_pca(X, y), y, id_to_class, vec)
if __name__ == '__main__':
ignore = ['imported_modules']
#ignore += ['imported_stdlib_modules']
features = [f for f in all_features if f not in ignore]
#features = ['imported_stdlib_modules']
_run(Repo.load_sample(), features)
class Analyzer(object):
def __init__(self, username, repo_name):
self.repo = Repo(repo_path(username, repo_name))
self.seen_files = {}
try:
print "Repo exists"
self.repo_model = RepoModel.objects.get(username = username, name = repo_name)
except RepoModel.DoesNotExist:
self.repo_model = RepoModel(username = username, name = repo_name)
self.repo_model.save()
self.cached_data = {}
def is_python_file(self, diff):
if diff.a_blob and re.search('.*\.py$', diff.a_blob.name):
return True
return False
def get_classes(self, entities):
return [e for e in entities if type(e) == ast.ClassDef]
def get_functions(self, entities):
return [e for e in entities if type(e) == ast.FunctionDef]
def get_all_funcs_from_body(self, body):
funcs = self.get_functions(body)
classes = self.get_classes(body)
for c in classes:
funcs = funcs + self.get_all_funcs_from_body(c.body)
for f in funcs:
funcs = funcs + self.get_all_funcs_from_body(f.body)
return funcs
def read_diffs(self, diffs):
new_funcs = []
files_changed = []
for diff in diffs:
if diff.a_blob and diff.b_blob:
a_blob_text = diff.a_blob.data_stream.read()
b_blob_text = diff.b_blob.data_stream.read()
try:
a_syntax_tree = ast.parse(a_blob_text)
except (ValueError, SyntaxError, TypeError) as e:
#Someone has committed some crap that's not valid python,
#carry on...
continue
a_entities = a_syntax_tree.body
a_funcs = self.get_all_funcs_from_body(a_entities)
a_func_names = [f.name for f in a_funcs]
file_name = diff.a_blob.abspath + diff.a_blob.name
files_changed.append(file_name)
if not self.seen_files.get(file_name, False):
#This is a new file, so ALL functions contained within it are new
self.seen_files[file_name] = True
new_funcs = new_funcs + [(diff.a_blob.abspath, fname) for fname in a_func_names]
print "New file!"
print new_funcs
else:
#Not a new file, get what has changed, so get the next blob, parse it, and get the
#functions from it.
#Get the syntax_tree for the second blob
try:
b_syntax_tree = ast.parse(b_blob_text)
except (ValueError, SyntaxError, TypeError) as e:
#Someone has committed some crap that's not valid python,
#carry on...
continue
b_entities = b_syntax_tree.body
b_funcs = self.get_all_funcs_from_body(b_entities)
b_func_names = [f.name for f in b_funcs]
#xor the functions
new_in_this_diff = list(set(a_func_names) ^ set(b_func_names))
new_funcs = new_funcs + [(diff.a_blob.abspath, fname) for fname in new_in_this_diff]
return new_funcs, files_changed
def store(self, commit, new_funcs, files_changed):
name = commit.author.name
date = commit.committed_date
self.cached_data[commit.hexsha] = {
'name' : name,
'date' : date,
'funcs' : new_funcs,
'files_changed' : files_changed
}
if(len(self.cached_data.keys()) > 30):
with transaction.commit_on_success():
self.do_save()
def do_save(self):
for hexsha in self.cached_data:
val = self.cached_data[hexsha]
try:
actor = Actor.objects.get(full_name = val['name'])
except Actor.DoesNotExist:
actor = Actor(full_name = val['name'])
actor.save()
#Create the actor
try:
commit = Commit.objects.get(hexsha = hexsha)
except Commit.DoesNotExist:
commit = Commit(hexsha = hexsha, repo = self.repo_model, actor = actor)
commit.save()
for path, fun in val['funcs']:
if not Function.objects.filter(name = fun, path = path).exists():
fmodel = Function(name = fun, commit = commit, path = path)
fmodel.save()
print "Saved `%s` : `%s`" % (path[-16:], fun)
for file_name in val['files_changed']:
FileChange(path = file_name, actor = actor, commit = commit).save()
self.cached_data.clear()
def walk_commits(self):
#This uses a lot of memory, but...I don't see another way to go backwards
#in git python
commits = []
for c in self.repo.iter_commits():
commits.append(c)
#pop the first commit off, so that all commits in the loop will have a parent
commits.pop()
while len(commits) > 0:
commit = commits.pop()
#Create a list of diffs based on the parent (aka commit before this commit)
try:
diffs = commit.diff(commit.parents[0])
diffs = [d for d in diffs if self.is_python_file(d)]
new_funcs, files_changed = self.read_diffs(diffs)
self.store(commit, new_funcs, files_changed)
except LookupError:
#This seems to be a bug in PyGit maybe?
#seems to throw this sometimes, not much we can do here...
continue
def get_repo_count(): return Repo.select().count()
def get(self, username, reponame):
timemap = self.get_query_argument("timemap", "false") == "true"
index = self.get_query_argument("index", "false") == "true"
key = self.get_query_argument("key", None)
if (index and timemap) or (index and key) or (timemap and not key):
raise HTTPError(400)
if self.get_query_argument("datetime", None):
datestr = self.get_query_argument("datetime")
ts = date(datestr, QSDATEFMT)
elif "Accept-Datetime" in self.request.headers:
datestr = self.request.headers.get("Accept-Datetime")
ts = date(datestr, RFC1123DATEFMT)
else:
ts = now()
try:
repo = (Repo.select(Repo.id).join(
User).where((User.name == username)
& (Repo.name == reponame)).naive().get())
except Repo.DoesNotExist:
raise HTTPError(404)
if key and not timemap:
# Recreate the resource for the given key in its latest state -
# if no `datetime` was provided - or in the state it was in at
# the time indicated by the passed `datetime` argument.
self.set_header("Content-Type", "application/n-quads")
self.set_header("Vary", "accept-datetime")
sha = shasum(key.encode("utf-8"))
# Fetch all relevant changes from the last "non-delta" onwards,
# ordered by time. The returned delta-chain consists of either:
# a snapshot followed by 0 or more deltas, or
# a single delete.
chain = list(
CSet.select(CSet.time, CSet.type).where((CSet.repo == repo) & (
CSet.hkey == sha) & (CSet.time <= ts) & (CSet.time >= SQL(
"COALESCE((SELECT time FROM cset "
"WHERE repo_id = %s "
"AND hkey_id = %s "
"AND time <= %s "
"AND type != %s "
"ORDER BY time DESC "
"LIMIT 1), 0)", repo.id, sha, ts, CSet.DELTA))).
order_by(CSet.time).naive())
if len(chain) == 0:
# A resource does not exist for the given key.
raise HTTPError(404)
timegate_url = (self.request.protocol + "://" + self.request.host +
self.request.path)
timemap_url = (self.request.protocol + "://" + self.request.host +
self.request.uri + "&timemap=true")
self.set_header(
"Link", '<%s>; rel="original"'
', <%s>; rel="timegate"'
', <%s>; rel="timemap"' % (key, timegate_url, timemap_url))
self.set_header("Memento-Datetime",
chain[-1].time.strftime(RFC1123DATEFMT))
if chain[0].type == CSet.DELETE:
# The last change was a delete. Return a 404 response with
# appropriate "Link" and "Memento-Datetime" headers.
raise HTTPError(404)
# Load the data required in order to restore the resource state.
blobs = (Blob.select(Blob.data).where(
(Blob.repo == repo) & (Blob.hkey == sha)
& (Blob.time << map(lambda e: e.time, chain))).order_by(
Blob.time).naive())
if len(chain) == 1:
# Special case, where we can simply return
# the blob data of the snapshot.
snap = blobs.first().data
return self.finish(decompress(snap))
stmts = set()
for i, blob in enumerate(blobs.iterator()):
data = decompress(blob.data)
if i == 0:
# Base snapshot for the delta chain
stmts.update(data.splitlines())
else:
for line in data.splitlines():
mode, stmt = line[0], line[2:]
if mode == "A":
stmts.add(stmt)
else:
stmts.discard(stmt)
self.write(join(stmts, "\n"))
elif key and timemap:
# Generate a timemap containing historic change information
# for the requested key. The timemap is in the default link-format
# or as JSON (http://mementoweb.org/guide/timemap-json/).
sha = shasum(key.encode("utf-8"))
csets = (CSet.select(
CSet.time).where((CSet.repo == repo)
& (CSet.hkey == sha)).order_by(
CSet.time.desc()).naive())
# TODO: Paginate?
csit = csets.iterator()
try:
first = csit.next()
except StopIteration:
# Resource for given key does not exist.
raise HTTPError(404)
req = self.request
base = req.protocol + "://" + req.host + req.path
accept = self.request.headers.get("Accept", "")
if "application/json" in accept or "*/*" in accept:
self.set_header("Content-Type", "application/json")
self.write('{"original_uri": ' + json_encode(key))
self.write(', "mementos": {"list":[')
m = ('{{"datetime": "{0}", "uri": "' + base + '?key=' +
url_escape(key) + '&datetime={1}"}}')
self.write(
m.format(first.time.isoformat(),
first.time.strftime(QSDATEFMT)))
for cs in csit:
self.write(', ' + m.format(cs.time.isoformat(),
cs.time.strftime(QSDATEFMT)))
self.write(']}')
self.write('}')
else:
m = (',\n'
'<' + base + '?key=' + url_escape(key) + '&datetime={0}>'
'; rel="memento"'
'; datetime="{1}"'
'; type="application/n-quads"')
self.set_header("Content-Type", "application/link-format")
self.write('<' + key + '>; rel="original"')
self.write(
m.format(first.time.strftime(QSDATEFMT),
first.time.strftime(RFC1123DATEFMT)))
for cs in csit:
self.write(
m.format(cs.time.strftime(QSDATEFMT),
cs.time.strftime(RFC1123DATEFMT)))
elif index:
# Generate an index of all URIs contained in the dataset at the
# provided point in time or in its current state.
self.set_header("Vary", "accept-datetime")
self.set_header("Content-Type", "text/plain")
page = int(self.get_query_argument("page", "1"))
# Subquery for selecting max. time per hkey group
mx = (CSet.select(
CSet.hkey,
fn.Max(CSet.time).alias("maxtime")).where(
(CSet.repo == repo) & (CSet.time <= ts)).group_by(
CSet.hkey).order_by(CSet.hkey).paginate(
page, INDEX_PAGE_SIZE).alias("mx"))
# Query for all the relevant csets (those with max. time values)
cs = (CSet.select(CSet.hkey, CSet.time).join(
mx,
on=((CSet.hkey == mx.c.hkey_id) & (CSet.time == mx.c.maxtime)
)).where((CSet.repo == repo)
& (CSet.type != CSet.DELETE)).alias("cs"))
# Join with the hmap table to retrieve the plain key values
hm = (HMap.select(HMap.val).join(
cs, on=(HMap.sha == cs.c.hkey_id)).naive())
for h in hm.iterator():
self.write(h.val + "\n")
else:
raise HTTPError(400)
def fetch():
sys.stdout.write('loading')
sys.stdout.flush()
repos = Repo.load_sample()
authors = {author.login: author for author in Author.load(FILE)}
seen = 0
total = len(repos)
failures = []
last_write = datetime.datetime.now()
el = Elaborator()
for repo in repos:
seen += 1
if repo.username in authors:
logging.info("already fetched %s", repo.username)
continue
try:
gh_data = el._gh_request(
'GET',
'/users/' + repo.username
)
except:
#loop really needs to keep running
logging.exception("problem! %s", repo)
failures.append(repo)
continue
authors[repo.username] = Author(**{key: gh_data.get(key, None) for key in
['login', # "octocat"
'id', # 1
'avatar_url', # "https://github.com/images/error/octocat_happy.gif"
'gravatar_id', # "somehexcode"
'url', # "https://api.github.com/users/octocat"
'name', # "monalisa octocat"
'company', # "GitHub"
'blog', # "https://github.com/blog"
'location', # "San Francisco"
'email', # "*****@*****.**"
'hireable', # false
'bio', # "There once was..."
'public_repos', # 2
'public_gists', # 1
'followers', # 20
'following', # 0
'html_url', # "https://github.com/octocat"
'created_at', # "2008-01-14T04:33:35Z"
'type', # "User"
]})
logging.info("fetched %s", repo.username)
progress_bar(seen, total)
since_write = datetime.datetime.now() - last_write
if since_write > datetime.timedelta(minutes=5):
sys.stdout.write("\r(writing results)")
sys.stdout.flush()
Author.dump(authors.values(), FILE)
last_write = datetime.datetime.now()
print # from progress bar line
if failures:
print "%s failures:" % len(failures)
for f in failures:
print " %s" % f
print
print 'writing out...'
Author.dump(authors.values(), FILE)
def post(self):
parsed_args = parser.parse_args()
repo = Repo(creator=parsed_args['creator'], name=parsed_args['name'])
session.add(repo)
session.commit()
return repo, 201
def put(self, username, reponame):
# Create a new revision of the resource specified by `key`.
fmt = self.request.headers.get("Content-Type", "application/n-triples")
key = self.get_query_argument("key", None)
if username != self.current_user.name:
raise HTTPError(403)
if not key:
raise HTTPError(400)
datestr = self.get_query_argument("datetime", None)
ts = datestr and date(datestr, QSDATEFMT) or now()
try:
repo = (Repo.select(Repo.id).join(
User).where((User.name == username)
& (Repo.name == reponame)).naive().get())
except Repo.DoesNotExist:
raise HTTPError(404)
sha = shasum(key.encode("utf-8"))
chain = list(
CSet.select(CSet.time, CSet.type, CSet.len).where(
(CSet.repo == repo) & (CSet.hkey == sha) & (CSet.time >= SQL(
"COALESCE((SELECT time FROM cset "
"WHERE repo_id = %s "
"AND hkey_id = %s "
"AND type != %s "
"ORDER BY time DESC "
"LIMIT 1), 0)", repo.id, sha, CSet.DELTA))).order_by(
CSet.time).naive())
if len(chain) > 0 and not ts > chain[-1].time:
# Appended timestamps must be monotonically increasing!
raise HTTPError(400)
if len(chain) == 0:
# Mapping for `key` likely does not exist:
# Store the SHA-to-KEY mapping in HMap,
# looking out for possible collisions.
try:
HMap.create(sha=sha, val=key)
except IntegrityError:
val = HMap.select(HMap.val).where(HMap.sha == sha).scalar()
if val != key:
raise HTTPError(500)
# Parse and normalize into a set of N-Quad lines
stmts = parse(self.request.body, fmt)
snapc = compress(join(stmts, "\n"))
if len(chain) == 0 or chain[0].type == CSet.DELETE:
# Provide dummy value for `patch` which is never stored.
# If we get here, we always store a snapshot later on!
patch = ""
else:
# Reconstruct the previous state of the resource
prev = set()
blobs = (Blob.select(Blob.data).where(
(Blob.repo == repo) & (Blob.hkey == sha)
& (Blob.time << map(lambda e: e.time, chain))).order_by(
Blob.time).naive())
for i, blob in enumerate(blobs.iterator()):
data = decompress(blob.data)
if i == 0:
# Base snapshot for the delta chain
prev.update(data.splitlines())
else:
for line in data.splitlines():
mode, stmt = line[0], line[2:]
if mode == "A":
prev.add(stmt)
else:
prev.discard(stmt)
if stmts == prev:
# No changes, nothing to be done. Bail out.
return self.finish()
patch = compress(
join(
map(lambda s: "D " + s, prev - stmts) +
map(lambda s: "A " + s, stmts - prev), "\n"))
# Calculate the accumulated size of the delta chain including
# the (potential) patch from the previous to the pushed state.
acclen = reduce(lambda s, e: s + e.len, chain[1:], 0) + len(patch)
blen = len(chain) > 0 and chain[0].len or 0 # base length
if (len(chain) == 0 or chain[0].type == CSet.DELETE
or len(snapc) <= len(patch) or SNAPF * blen <= acclen):
# Store the current state as a new snapshot
Blob.create(repo=repo, hkey=sha, time=ts, data=snapc)
CSet.create(repo=repo,
hkey=sha,
time=ts,
type=CSet.SNAPSHOT,
len=len(snapc))
else:
# Store a directed delta between the previous and current state
Blob.create(repo=repo, hkey=sha, time=ts, data=patch)
CSet.create(repo=repo,
hkey=sha,
time=ts,
type=CSet.DELTA,
len=len(patch))