Fixed bugs and improved logging in config recommendation tasks/views
This commit is contained in:
dvanaken
Dana Van Aken
parent
3786714093
commit
9393fb7aca
|
|
@ -338,7 +338,6 @@ def get_result(max_time_sec=180, interval_sec=5, upload_code=None):
|
|||
url = dconf.WEBSITE_URL + '/query_and_get/' + upload_code
|
||||
elapsed = 0
|
||||
response_dict = None
|
||||
response = ''
|
||||
rout = ''
|
||||
|
||||
while elapsed <= max_time_sec:
|
||||
|
|
@ -347,8 +346,9 @@ def get_result(max_time_sec=180, interval_sec=5, upload_code=None):
|
|||
assert response != 'null'
|
||||
rout = json.dumps(response, indent=4) if isinstance(response, dict) else response
|
||||
|
||||
LOG.debug('%s [status code: %d, content_type: %s, elapsed: %ds]', rout,
|
||||
rsp.status_code, rsp.headers.get('Content-Type', ''), elapsed)
|
||||
LOG.debug('%s\n\n[status code: %d, type(response): %s, elapsed: %ds, %s]', rout,
|
||||
rsp.status_code, type(response), elapsed,
|
||||
', '.join(['{}: {}'.format(k, v) for k, v in rsp.headers.items()]))
|
||||
|
||||
if rsp.status_code == 200:
|
||||
# Success
|
||||
|
|
@ -368,12 +368,12 @@ def get_result(max_time_sec=180, interval_sec=5, upload_code=None):
|
|||
|
||||
elif rsp.status_code == 500:
|
||||
# Failure
|
||||
if len(response) > 5000:
|
||||
with open('error.html', 'w') as f:
|
||||
f.write(response)
|
||||
msg = "Saved HTML error to 'error.html'."
|
||||
else:
|
||||
msg = rout
|
||||
if isinstance(response, str):
|
||||
savepath = os.path.join(dconf.LOG_DIR, 'error.html')
|
||||
with open(savepath, 'w') as f:
|
||||
f.write(response)
|
||||
msg = "Saved HTML error to '{}'.".format(os.path.relpath(savepath))
|
||||
raise Exception(
|
||||
"Failed to download the next config.\nStatus code: {}\nMessage: {}\n".format(
|
||||
rsp.status_code, msg))
|
||||
|
|
@ -385,12 +385,12 @@ def get_result(max_time_sec=180, interval_sec=5, upload_code=None):
|
|||
if not response_dict:
|
||||
assert elapsed > max_time_sec, \
|
||||
'response={} but elapsed={}s <= max_time={}s'.format(
|
||||
response, elapsed, max_time_sec)
|
||||
rout, elapsed, max_time_sec)
|
||||
raise Exception(
|
||||
'Failed to download the next config in {}s: {} (elapsed: {}s)'.format(
|
||||
max_time_sec, rout, elapsed))
|
||||
|
||||
LOG.info('Downloaded the next config in %ds: %s', elapsed, rout)
|
||||
LOG.info('Downloaded the next config in %ds', elapsed)
|
||||
|
||||
return response_dict
|
||||
|
||||
|
|
@ -792,3 +792,5 @@ def integration_tests():
|
|||
upload_code='ottertuneTestTuningGPR')
|
||||
response = get_result(upload_code='ottertuneTestTuningGPR')
|
||||
assert response['status'] == 'good'
|
||||
|
||||
LOG.info("\n\nIntegration Tests: PASSED!!\n")
|
||||
|
|
|
|||
|
|
@ -83,58 +83,56 @@ class TaskUtilTest(TestCase):
|
|||
# FIXME: Actually setup celery tasks instead of a dummy class?
|
||||
test_tasks = []
|
||||
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks)
|
||||
self.assertTrue(status is None and num_complete == 0)
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks, 1)
|
||||
self.assertTrue(status is 'UNAVAILABLE' and num_complete == 0)
|
||||
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks, 0)
|
||||
self.assertTrue(status is 'UNAVAILABLE' and num_complete == 0)
|
||||
|
||||
test_tasks2 = [VarType() for i in range(5)]
|
||||
for task in test_tasks2:
|
||||
task.status = "SUCCESS"
|
||||
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks2)
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks2, 5)
|
||||
self.assertTrue(status == "SUCCESS" and num_complete == 5)
|
||||
|
||||
test_tasks3 = test_tasks2
|
||||
test_tasks3[3].status = "FAILURE"
|
||||
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks3)
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks3, 5)
|
||||
self.assertTrue(status == "FAILURE" and num_complete == 3)
|
||||
|
||||
test_tasks4 = test_tasks3
|
||||
test_tasks4[2].status = "REVOKED"
|
||||
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks4)
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks4, 5)
|
||||
self.assertTrue(status == "REVOKED" and num_complete == 2)
|
||||
|
||||
test_tasks5 = test_tasks4
|
||||
test_tasks5[1].status = "RETRY"
|
||||
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks5)
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks5, 5)
|
||||
self.assertTrue(status == "RETRY" and num_complete == 1)
|
||||
|
||||
test_tasks6 = [VarType() for i in range(10)]
|
||||
for i, task in enumerate(test_tasks6):
|
||||
task.status = "PENDING" if i % 2 == 0 else "SUCCESS"
|
||||
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks6)
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks6, 10)
|
||||
self.assertTrue(status == "PENDING" and num_complete == 5)
|
||||
|
||||
test_tasks7 = test_tasks6
|
||||
test_tasks7[9].status = "STARTED"
|
||||
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks7)
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks7, 10)
|
||||
self.assertTrue(status == "STARTED" and num_complete == 4)
|
||||
|
||||
test_tasks8 = test_tasks7
|
||||
test_tasks8[9].status = "RECEIVED"
|
||||
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks8)
|
||||
(status, num_complete) = TaskUtil.get_task_status(test_tasks8, 10)
|
||||
self.assertTrue(status == "RECEIVED" and num_complete == 4)
|
||||
|
||||
with self.assertRaises(Exception):
|
||||
test_tasks9 = [VarType() for i in range(1)]
|
||||
test_tasks9[0].status = "attemped"
|
||||
TaskUtil.get_task_status(test_tasks9)
|
||||
|
||||
|
||||
class DataUtilTest(TestCase):
|
||||
|
||||
|
|
|
|||
|
|
@ -28,7 +28,7 @@ from analysis.constraints import ParamConstraintHelper
|
|||
from website.models import (PipelineData, PipelineRun, Result, Workload, KnobCatalog, SessionKnob,
|
||||
MetricCatalog)
|
||||
from website import db
|
||||
from website.types import PipelineTaskType, AlgorithmType
|
||||
from website.types import PipelineTaskType, AlgorithmType, VarType
|
||||
from website.utils import DataUtil, JSONUtil
|
||||
from website.settings import IMPORTANT_KNOB_NUMBER, NUM_SAMPLES, TOP_NUM_CONFIG # pylint: disable=no-name-in-module
|
||||
from website.settings import (USE_GPFLOW, DEFAULT_LENGTH_SCALE, DEFAULT_MAGNITUDE,
|
||||
|
|
@ -47,7 +47,6 @@ from website.settings import (USE_GPFLOW, DEFAULT_LENGTH_SCALE, DEFAULT_MAGNITUD
|
|||
GPR_MODEL_NAME, ENABLE_DUMMY_ENCODER, DNN_GD_ITER)
|
||||
|
||||
from website.settings import INIT_FLIP_PROB, FLIP_PROB_DECAY
|
||||
from website.types import VarType
|
||||
|
||||
|
||||
LOG = get_task_logger(__name__)
|
||||
|
|
@ -116,32 +115,59 @@ class ConfigurationRecommendation(UpdateTask): # pylint: disable=abstract-metho
|
|||
|
||||
def clean_knob_data(knob_matrix, knob_labels, session):
|
||||
# Makes sure that all knobs in the dbms are included in the knob_matrix and knob_labels
|
||||
knob_cat = SessionKnob.objects.get_knobs_for_session(session)
|
||||
knob_cat = [knob["name"] for knob in knob_cat if knob["tunable"]]
|
||||
matrix = np.array(knob_matrix)
|
||||
missing_columns = set(knob_cat) - set(knob_labels)
|
||||
unused_columns = set(knob_labels) - set(knob_cat)
|
||||
LOG.debug("clean_knob_data added %d knobs and removed %d knobs.", len(missing_columns),
|
||||
len(unused_columns))
|
||||
# If columns are missing from the matrix
|
||||
if missing_columns:
|
||||
for knob in missing_columns:
|
||||
knob_object = KnobCatalog.objects.get(dbms=session.dbms, name=knob, tunable=True)
|
||||
index = knob_cat.index(knob)
|
||||
knob_matrix = np.array(knob_matrix)
|
||||
session_knobs = SessionKnob.objects.get_knobs_for_session(session)
|
||||
knob_cat = [k['name'] for k in session_knobs]
|
||||
|
||||
if knob_cat == knob_labels:
|
||||
# Nothing to do!
|
||||
return knob_matrix, knob_labels
|
||||
|
||||
LOG.info("session_knobs: %s, knob_labels: %s, missing: %s, extra: %s", len(knob_cat),
|
||||
len(knob_labels), len(set(knob_cat) - set(knob_labels)),
|
||||
len(set(knob_labels) - set(knob_cat)))
|
||||
|
||||
nrows = knob_matrix.shape[0] # pylint: disable=unsubscriptable-object
|
||||
new_labels = []
|
||||
new_columns = []
|
||||
|
||||
for knob in session_knobs:
|
||||
knob_name = knob['name']
|
||||
if knob_name not in knob_labels:
|
||||
# Add missing column initialized to knob's default value
|
||||
default_val = knob['default']
|
||||
try:
|
||||
default_val = float(knob_object.default)
|
||||
if knob['vartype'] == VarType.ENUM:
|
||||
default_val = knob['enumvals'].split(',').index(default_val)
|
||||
elif knob['vartype'] == VarType.BOOL:
|
||||
default_val = str(default_val).lower() in ("on", "true", "yes", "0")
|
||||
else:
|
||||
default_val = float(default_val)
|
||||
except ValueError:
|
||||
default_val = 0
|
||||
matrix = np.insert(matrix, index, default_val, axis=1)
|
||||
knob_labels.insert(index, knob)
|
||||
# If they are useless columns in the matrix
|
||||
if unused_columns:
|
||||
indexes = [i for i, n in enumerate(knob_labels) if n in unused_columns]
|
||||
# Delete unused columns
|
||||
matrix = np.delete(matrix, indexes, 1)
|
||||
for i in sorted(indexes, reverse=True):
|
||||
del knob_labels[i]
|
||||
return matrix, knob_labels
|
||||
LOG.exception("Error parsing knob default: %s=%s", knob_name, default_val)
|
||||
new_col = np.ones((nrows, 1), dtype=float) * default_val
|
||||
new_lab = knob_name
|
||||
else:
|
||||
index = knob_labels.index(knob_name)
|
||||
new_col = knob_matrix[:, index]
|
||||
new_lab = knob_labels[index]
|
||||
|
||||
new_labels.append(new_lab)
|
||||
new_columns.append(new_col)
|
||||
|
||||
new_matrix = np.hstack(new_columns).reshape(nrows, -1)
|
||||
LOG.debug("Cleaned matrix: %s, knobs (%s): %s", new_matrix.shape,
|
||||
len(new_labels), new_labels)
|
||||
|
||||
assert new_labels == knob_cat, \
|
||||
"Expected knobs: {}\nActual knobs: {}\n".format(
|
||||
knob_cat, new_labels)
|
||||
assert new_matrix.shape == (nrows, len(knob_cat)), \
|
||||
"Expected shape: {}, Actual shape: {}".format(
|
||||
(nrows, len(knob_cat)), new_matrix.shape)
|
||||
|
||||
return new_matrix, new_labels
|
||||
|
||||
|
||||
def clean_metric_data(metric_matrix, metric_labels, session):
|
||||
|
|
@ -438,8 +464,7 @@ def configuration_recommendation_ddpg(result_info): # pylint: disable=invalid-n
|
|||
metric_data = metric_data.flatten()
|
||||
metric_scalar = MinMaxScaler().fit(metric_data.reshape(1, -1))
|
||||
normalized_metric_data = metric_scalar.transform(metric_data.reshape(1, -1))[0]
|
||||
cleaned_knob_data = clean_knob_data(agg_data['X_matrix'], agg_data['X_columnlabels'],
|
||||
session)
|
||||
cleaned_knob_data = clean_knob_data(agg_data['X_matrix'], agg_data['X_columnlabels'], session)
|
||||
knob_labels = np.array(cleaned_knob_data[1]).flatten()
|
||||
knob_num = len(knob_labels)
|
||||
metric_num = len(metric_data)
|
||||
|
|
@ -834,7 +859,6 @@ def map_workload(map_workload_input):
|
|||
knob_data["data"], knob_data["columnlabels"] = clean_knob_data(knob_data["data"],
|
||||
knob_data["columnlabels"],
|
||||
newest_result.session)
|
||||
|
||||
metric_data = load_data_helper(pipeline_data, unique_workload, PipelineTaskType.METRIC_DATA)
|
||||
X_matrix = np.array(knob_data["data"])
|
||||
y_matrix = np.array(metric_data["data"])
|
||||
|
|
|
|||
|
|
@ -12,9 +12,13 @@
|
|||
<td style="width: 50%"><div class="text-right">{{ labels.id }}</div></td>
|
||||
<td style="width: 50%">{{ result.pk }}</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><div class="text-right">{{ labels.creation_time }}</div></td>
|
||||
<td>{{ result.creation_time }}</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><div class="text-right">{{ labels.session }}</div></td>
|
||||
<td>{{ result.session.name }}</td>
|
||||
<td><a href="{% url 'session' project_id session_id %}">{{ result.session.name }}</a></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><div class="text-right">{{ labels.workload }}</div></td>
|
||||
|
|
@ -28,10 +32,6 @@
|
|||
<td><div class="text-top text-right">{{ labels.metric_data }}</div></td>
|
||||
<td><a href="{% url 'metric_data' project_id session_id result.metric_data.pk %}">{{ result.metric_data.name }}</a></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td><div class="text-right">{{ labels.creation_time }}</div></td>
|
||||
<td>{{ result.creation_time }}</td>
|
||||
</tr>
|
||||
{% for key, value in default_metrics.items %}
|
||||
<tr>
|
||||
<td><div class="text-right">{{ metric_meta|get_item:key|get_attr:"pprint" }}</div></td>
|
||||
|
|
@ -47,8 +47,16 @@
|
|||
{% endif %}
|
||||
{% if next_conf_available %}
|
||||
<tr>
|
||||
<td><div class="text-right">{{ labels.next_conf_available }}</div></td>
|
||||
<td><a href="/get_result_data_file/?id={{ result.pk }}&type=next_conf">Download</a></td>
|
||||
<td>
|
||||
<div class="text-right">
|
||||
{{ labels.next_conf }}<br>
|
||||
<button type="submit" style="margin-top:15px;" class="btn btn-default btn-lg alert alert-info"
|
||||
onclick="location.href='/get_result_data_file/?id={{ result.pk }}&type=next_conf'">
|
||||
<span class="glyphicon glyphicon-download-alt" aria-hidden="true"></span>
|
||||
</button>
|
||||
</div>
|
||||
</td>
|
||||
<td><pre style="display:inline-block">{{ next_conf }}</pre></td>
|
||||
</tr>
|
||||
{% endif %}
|
||||
</tbody>
|
||||
|
|
|
|||
|
|
@ -82,11 +82,8 @@ class TaskUtil(object):
|
|||
return TaskMeta.objects.filter(task_id__in=task_ids).order_by(preserved)
|
||||
|
||||
@staticmethod
|
||||
def get_task_status(tasks):
|
||||
if not tasks:
|
||||
return None, 0
|
||||
|
||||
overall_status = 'SUCCESS'
|
||||
def get_task_status(tasks, num_tasks):
|
||||
overall_status = 'UNAVAILABLE'
|
||||
num_completed = 0
|
||||
for task in tasks:
|
||||
status = task.status
|
||||
|
|
@ -101,6 +98,9 @@ class TaskUtil(object):
|
|||
task.id, status, task.task_id)
|
||||
overall_status = status
|
||||
|
||||
if num_tasks > 0 and num_tasks == num_completed:
|
||||
overall_status = 'SUCCESS'
|
||||
|
||||
return overall_status, num_completed
|
||||
|
||||
|
||||
|
|
@ -484,3 +484,13 @@ def delete_user(username):
|
|||
deleted = False
|
||||
|
||||
return delete_info, deleted
|
||||
|
||||
|
||||
def model_to_dict2(m, exclude=None):
|
||||
exclude = exclude or []
|
||||
d = {}
|
||||
for f in m._meta.fields: # pylint: disable=protected-access
|
||||
fname = f.name
|
||||
if fname not in exclude:
|
||||
d[fname] = getattr(m, fname, None)
|
||||
return d
|
||||
|
|
|
|||
|
|
@ -398,18 +398,33 @@ def result_view(request, project_id, session_id, result_id):
|
|||
# default_metrics = {mname: metric_data[mname] * metric_meta[mname].scale
|
||||
# for mname in default_metrics}
|
||||
|
||||
status = None
|
||||
if target.task_ids is not None:
|
||||
tasks = TaskUtil.get_tasks(target.task_ids)
|
||||
status, _ = TaskUtil.get_task_status(tasks)
|
||||
if status is None:
|
||||
status = 'UNAVAILABLE'
|
||||
task_ids = [t for t in (target.task_ids or '').split(',') if t.strip() != '']
|
||||
tasks = TaskUtil.get_tasks(task_ids)
|
||||
status, _ = TaskUtil.get_task_status(tasks, len(task_ids))
|
||||
|
||||
next_conf_available = True if status == 'SUCCESS' else False
|
||||
next_conf = ''
|
||||
cfg = target.next_configuration
|
||||
LOG.debug("status: %s, next_conf_available: %s, next_conf: %s, type: %s",
|
||||
status, next_conf_available, cfg, type(cfg))
|
||||
|
||||
if next_conf_available:
|
||||
try:
|
||||
cfg = JSONUtil.loads(cfg)['recommendation']
|
||||
kwidth = max(len(k) for k in cfg.keys())
|
||||
vwidth = max(len(str(v)) for v in cfg.values())
|
||||
next_conf = ''
|
||||
for k, v in cfg.items():
|
||||
next_conf += '{: <{kwidth}} = {: >{vwidth}}\n'.format(
|
||||
k, v, kwidth=kwidth, vwidth=vwidth)
|
||||
except Exception as e: # pylint: disable=broad-except
|
||||
LOG.exception("Failed to format the next config (type=%s): %s.\n\n%s\n",
|
||||
type(cfg), cfg, e)
|
||||
|
||||
form_labels = Result.get_labels()
|
||||
form_labels.update(LabelUtil.style_labels({
|
||||
'status': 'status',
|
||||
'next_conf_available': 'next configuration'
|
||||
'next_conf': 'next configuration',
|
||||
}))
|
||||
form_labels['title'] = 'Result Info'
|
||||
context = {
|
||||
|
|
@ -417,6 +432,7 @@ def result_view(request, project_id, session_id, result_id):
|
|||
'metric_meta': metric_meta,
|
||||
'status': status,
|
||||
'next_conf_available': next_conf_available,
|
||||
'next_conf': next_conf,
|
||||
'labels': form_labels,
|
||||
'project_id': project_id,
|
||||
'session_id': session_id
|
||||
|
|
@ -843,14 +859,15 @@ def download_debug_info(request, project_id, session_id): # pylint: disable=unu
|
|||
def tuner_status_view(request, project_id, session_id, result_id): # pylint: disable=unused-argument
|
||||
res = Result.objects.get(pk=result_id)
|
||||
|
||||
tasks = TaskUtil.get_tasks(res.task_ids)
|
||||
task_ids = [t for t in (res.task_ids or '').split(',') if t.strip() != '']
|
||||
tasks = TaskUtil.get_tasks(task_ids)
|
||||
|
||||
overall_status, num_completed = TaskUtil.get_task_status(tasks)
|
||||
if overall_status in ['PENDING', 'RECEIVED', 'STARTED', None]:
|
||||
overall_status, num_completed = TaskUtil.get_task_status(tasks, len(task_ids))
|
||||
if overall_status in ['PENDING', 'RECEIVED', 'STARTED', 'UNAVAILABLE']:
|
||||
completion_time = 'N/A'
|
||||
total_runtime = 'N/A'
|
||||
else:
|
||||
completion_time = tasks[-1].date_done
|
||||
completion_time = tasks.reverse()[0].date_done
|
||||
total_runtime = (completion_time - res.creation_time).total_seconds()
|
||||
total_runtime = '{0:.2f} seconds'.format(total_runtime)
|
||||
|
||||
|
|
@ -1066,7 +1083,6 @@ def get_timeline_data(request):
|
|||
|
||||
# get the lastest result
|
||||
def give_result(request, upload_code): # pylint: disable=unused-argument
|
||||
|
||||
try:
|
||||
session = Session.objects.get(upload_code=upload_code)
|
||||
except Session.DoesNotExist:
|
||||
|
|
@ -1074,17 +1090,19 @@ def give_result(request, upload_code): # pylint: disable=unused-argument
|
|||
return HttpResponse("Invalid upload code: " + upload_code, status=400)
|
||||
|
||||
latest_result = Result.objects.filter(session=session).latest('creation_time')
|
||||
tasks = TaskUtil.get_tasks(latest_result.task_ids)
|
||||
overall_status, num_completed = TaskUtil.get_task_status(tasks)
|
||||
response = {
|
||||
'celery_status': overall_status,
|
||||
'result_id': latest_result.pk,
|
||||
'message': '',
|
||||
'errors': [],
|
||||
}
|
||||
task_ids = [t for t in (latest_result.task_ids or '').split(',') if t.strip() != '']
|
||||
tasks = TaskUtil.get_tasks(task_ids)
|
||||
overall_status, num_completed = TaskUtil.get_task_status(tasks, len(task_ids))
|
||||
next_config = latest_result.next_configuration
|
||||
|
||||
LOG.debug("result_id: %s, overall_status: %s, tasks_completed: %s/%s, "
|
||||
"next_config: %s\n", latest_result.pk, overall_status, num_completed,
|
||||
len(task_ids), next_config)
|
||||
|
||||
response = dict(celery_status=overall_status, result_id=latest_result.pk, message='', errors=[])
|
||||
if overall_status == 'SUCCESS':
|
||||
response.update(JSONUtil.loads(latest_result.next_configuration),
|
||||
next_config = JSONUtil.loads(next_config)
|
||||
response.update(next_config,
|
||||
message='Celery successfully recommended the next configuration')
|
||||
status_code = 200
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue