def get_predict_dsl():
request_data = request.json
request_data['query_filters'] = ['inference_dsl']
retcode, retmsg, data = model_utils.query_model_info_from_file(
**request_data)
if data:
for d in data:
if d.get("f_role") in {"guest", "host"}:
_data = d
break
else:
return error_response(
210,
"can not found guest or host model, please get predict dsl on guest or host."
)
if request_data.get("filename"):
os.makedirs(TEMP_DIRECTORY, exist_ok=True)
temp_filepath = os.path.join(TEMP_DIRECTORY,
request_data.get("filename"))
with open(temp_filepath, "w") as fout:
fout.write(json_dumps(_data['f_inference_dsl'], indent=4))
return send_file(open(temp_filepath, "rb"),
as_attachment=True,
attachment_filename=request_data.get("filename"))
else:
return get_json_result(data=_data['f_inference_dsl'])
return error_response(
210,
"No model found, please check if arguments are specified correctly.")
def component_output_data_download():
request_data = request.json
tasks = JobSaver.query_task(only_latest=True,
job_id=request_data['job_id'],
component_name=request_data['component_name'],
role=request_data['role'],
party_id=request_data['party_id'])
if not tasks:
raise ValueError(
f'no found task, please check if the parameters are correct:{request_data}'
)
import_component_output_depend(tasks[0].f_provider_info)
try:
output_tables_meta = get_component_output_tables_meta(
task_data=request_data)
except Exception as e:
stat_logger.exception(e)
return error_response(210, str(e))
limit = request_data.get('limit', -1)
if not output_tables_meta:
return error_response(response_code=210, retmsg='no data')
if limit == 0:
return error_response(response_code=210, retmsg='limit is 0')
tar_file_name = 'job_{}_{}_{}_{}_output_data.tar.gz'.format(
request_data['job_id'], request_data['component_name'],
request_data['role'], request_data['party_id'])
return TableStorage.send_table(output_tables_meta,
tar_file_name,
limit=limit,
need_head=request_data.get("head", True))
def validate_component_param():
if not request.json or not isinstance(request.json, dict):
return error_response(400, 'bad request')
required_keys = [
'component_name',
'component_module_name',
]
config_keys = ['role']
dsl_version = int(request.json.get('dsl_version', 0))
if dsl_version == 1:
config_keys += ['role_parameters', 'algorithm_parameters']
parser_class = DSLParser
elif dsl_version == 2:
config_keys += ['component_parameters']
parser_class = DSLParserV2
else:
return error_response(400, 'unsupported dsl_version')
try:
check_config(request.json, required_keys + config_keys)
except Exception as e:
return error_response(400, str(e))
try:
parser_class.validate_component_param(
get_federatedml_setting_conf_directory(),
{i: request.json[i]
for i in config_keys}, *[request.json[i] for i in required_keys])
except Exception as e:
return error_response(400, str(e))
return get_json_result()
def get_component_summary():
request_data = request.json
try:
required_params = ["job_id", "component_name", "role", "party_id"]
detect_utils.check_config(request_data, required_params)
tracker = Tracker(job_id=request_data["job_id"],
component_name=request_data["component_name"],
role=request_data["role"],
party_id=request_data["party_id"],
task_id=request_data.get("task_id", None),
task_version=request_data.get("task_version", None))
summary = tracker.read_summary_from_db()
if summary:
if request_data.get("filename"):
temp_filepath = os.path.join(TEMP_DIRECTORY,
request_data.get("filename"))
with open(temp_filepath, "w") as fout:
fout.write(json.dumps(summary, indent=4))
return send_file(
open(temp_filepath, "rb"),
as_attachment=True,
attachment_filename=request_data.get("filename"))
else:
return get_json_result(data=summary)
return error_response(
210,
"No component summary found, please check if arguments are specified correctly."
)
except Exception as e:
stat_logger.exception(e)
return error_response(210, str(e))
def get_config():
kwargs = {}
job_configuration = None
for i in ('job_id', 'role', 'party_id'):
if request.json.get(i) is None:
return error_response(400, f"'{i}' is required.")
kwargs[i] = str(request.json[i])
for i in ('component_name', 'task_id', 'task_version'):
if request.json.get(i) is None:
break
kwargs[i] = str(request.json[i])
else:
try:
job_configuration = job_utils.get_task_using_job_conf(**kwargs)
except Exception:
pass
if job_configuration is None:
job_configuration = job_utils.get_job_configuration(
kwargs['job_id'], kwargs['role'], kwargs['party_id'])
if job_configuration is None:
return error_response(404, 'Job not found.')
return get_json_result(data=job_configuration.to_dict())
def transfer_model():
party_model_id = request.json.get('namespace')
model_version = request.json.get('name')
if not party_model_id or not model_version:
return error_response(400, 'namespace and name are required')
model_data = publish_model.download_model(party_model_id, model_version)
if model_data is None:
return error_response(404, 'model not found')
return get_json_result(data=model_data)
def component_output_data_download():
request_data = request.json
output_data_table = get_component_output_data_table(task_data=request_data)
limit = request_data.get('limit', -1)
if not output_data_table:
return error_response(response_code=500, retmsg='no data')
if limit == 0:
return error_response(response_code=500, retmsg='limit is 0')
output_data_count = 0
have_data_label = False
output_tmp_dir = os.path.join(os.getcwd(), 'tmp/{}'.format(fate_uuid()))
output_file_path = '{}/output_%s'.format(output_tmp_dir)
output_data_file_path = output_file_path % 'data.csv'
os.makedirs(os.path.dirname(output_data_file_path), exist_ok=True)
with open(output_data_file_path, 'w') as fw:
for k, v in output_data_table.collect():
data_line, have_data_label = get_component_output_data_line(
src_key=k, src_value=v)
fw.write('{}\n'.format(','.join(map(lambda x: str(x), data_line))))
output_data_count += 1
if output_data_count == limit:
break
if output_data_count:
# get meta
header = get_component_output_data_meta(
output_data_table=output_data_table,
have_data_label=have_data_label)
output_data_meta_file_path = output_file_path % 'data_meta.json'
with open(output_data_meta_file_path, 'w') as fw:
json.dump({'header': header}, fw, indent=4)
if request_data.get('head', True):
with open(output_data_file_path, 'r+') as f:
content = f.read()
f.seek(0, 0)
f.write('{}\n'.format(','.join(header)) + content)
# tar
memory_file = io.BytesIO()
tar = tarfile.open(fileobj=memory_file, mode='w:gz')
tar.add(output_data_file_path,
os.path.relpath(output_data_file_path, output_tmp_dir))
tar.add(output_data_meta_file_path,
os.path.relpath(output_data_meta_file_path, output_tmp_dir))
tar.close()
memory_file.seek(0)
try:
shutil.rmtree(os.path.dirname(output_data_file_path))
except Exception as e:
# warning
stat_logger.warning(e)
tar_file_name = 'job_{}_{}_{}_{}_output_data.tar.gz'.format(
request_data['job_id'], request_data['component_name'],
request_data['role'], request_data['party_id'])
return send_file(memory_file,
attachment_filename=tar_file_name,
as_attachment=True)
def get_mysql_info():
if IS_STANDALONE:
return error_response(404, 'mysql only available on cluster mode')
try:
with DB.connection_context():
DB.random()
except Exception as e:
return error_response(503, str(e))
return error_response(200)
def list_job():
limit, offset = parse_limit_and_offset()
query = {
'tag': ('!=', 'submit_failed'),
}
for i in ('job_id', 'description'):
if request.json.get(i) is not None:
query[i] = ('contains', request.json[i])
if request.json.get('party_id') is not None:
try:
query['party_id'] = int(request.json['party_id'])
except Exception:
return error_response(400, f"Invalid parameter 'party_id'.")
query['party_id'] = ('contains', query['party_id'])
if request.json.get('partner') is not None:
query['roles'] = ('contains', query['partner'])
for i in ('role', 'status'):
if request.json.get(i) is None:
continue
if isinstance(request.json[i], str):
request.json[i] = [request.json[i]]
if not isinstance(request.json[i], list):
return error_response(400, f"Invalid parameter '{i}'.")
request.json[i] = set(request.json[i])
for j in request.json[i]:
if j not in valid_query_parameters[i]:
return error_response(400, f"Invalid parameter '{i}'.")
query[i] = ('in_', request.json[i])
jobs, count = job_utils.list_job(limit, offset, query,
parse_order_by(('create_time', 'desc')))
jobs = [job.to_human_model_dict() for job in jobs]
for job in jobs:
job['party_id'] = int(job['party_id'])
job['partners'] = set()
for i in ('guest', 'host', 'arbiter'):
job['partners'].update(job['roles'].get(i, []))
job['partners'].discard(job['party_id'])
job['partners'] = sorted(job['partners'])
return get_json_result(data={
'jobs': jobs,
'count': count,
})
def parse_order_by(default=None):
order_by = []
if request.json.get('order_by') is not None:
if request.json['order_by'] not in valid_query_parameters['order_by']:
abort(error_response(400, f"Invalid parameter 'order_by'."))
order_by.append(request.json['order_by'])
if request.json.get('order') is not None:
if request.json['order'] not in valid_query_parameters['order']:
abort(error_response(400, f"Invalid parameter order 'order'."))
order_by.append(request.json['order'])
return order_by or default
def get_eggroll_info():
if IS_STANDALONE:
return error_response(404, 'eggroll only available on cluster mode')
if PROXY != CoordinationProxyService.ROLLSITE:
return error_response(
404, 'coordination communication protocol is not rollsite')
conf = ServiceRegistry.FATE_ON_EGGROLL['rollsite']
with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
r = s.connect_ex((conf['host'], conf['port']))
if r != 0:
return error_response(503)
return error_response(200)
def dsl_generator():
data = request.json
cpn_str = data.get("cpn_str", "")
try:
if not cpn_str:
raise Exception("Component list should not be empty.")
if isinstance(cpn_str, list):
cpn_list = cpn_str
else:
if (cpn_str.find("/") and cpn_str.find("\\")) != -1:
raise Exception(
"Component list string should not contain '/' or '\\'.")
cpn_str = cpn_str.replace(" ", "").replace("\n", "").strip(",[]")
cpn_list = cpn_str.split(",")
train_dsl = json_loads(data.get("train_dsl"))
parser = schedule_utils.get_dsl_parser_by_version(
data.get("version", "2"))
predict_dsl = parser.deploy_component(cpn_list, train_dsl)
if data.get("filename"):
os.makedirs(TEMP_DIRECTORY, exist_ok=True)
temp_filepath = os.path.join(TEMP_DIRECTORY, data.get("filename"))
with open(temp_filepath, "w") as fout:
fout.write(json.dumps(predict_dsl, indent=4))
return send_file(open(temp_filepath, 'rb'),
as_attachment=True,
attachment_filename=data.get("filename"))
return get_json_result(data=predict_dsl)
except Exception as e:
stat_logger.exception(e)
return error_response(
210, "DSL generating failed. For more details, "
"please check logs/fate_flow/fate_flow_stat.log.")
def parse_limit_and_offset():
try:
limit = int(request.json.get('limit', 0))
page = int(request.json.get('page', 1)) - 1
except Exception:
abort(error_response(400, f"Invalid parameter 'limit' or 'page'."))
return limit, limit * page
def get_fateboard_info():
host = ServiceRegistry.FATEBOARD.get('host')
port = ServiceRegistry.FATEBOARD.get('port')
if not host or not port:
return error_response(404, 'fateboard is not configured')
return get_json_result(data={
'host': host,
'port': port,
})
def get_fateboard_info():
fateboard = get_base_config('fateboard', {})
host = fateboard.get('host')
port = fateboard.get('port')
if not host or not port:
return error_response(404, 'fateboard is not configured')
return get_json_result(data={
'host': host,
'port': port,
})
def get_checkpoint():
checkpoint_manager = load_checkpoints()
if 'step_index' in request.json:
try:
request.json['step_index'] = int(request.json['step_index'])
except Exception:
return error_response(400, "Invalid 'step_index'")
checkpoint = checkpoint_manager.get_checkpoint_by_index(
request.json['step_index'])
elif 'step_name' in request.json:
checkpoint = checkpoint_manager.get_checkpoint_by_name(
request.json['step_name'])
else:
return error_response(400, "'step_index' or 'step_name' is required")
if checkpoint is None:
return error_response(404, "The checkpoint was not found.")
return get_json_result(data=checkpoint.to_dict(True))
def check_job_log_dir():
job_id = str(request.json['job_id'])
job_log_dir = job_utils.get_job_log_directory(job_id=job_id)
if not os.path.exists(job_log_dir):
abort(
error_response(
404,
f"Log file path: '{job_log_dir}' not found. Please check if the job id is valid."
))
return job_id, job_log_dir
def list_task():
limit, offset = parse_limit_and_offset()
query = {}
for i in ('job_id', 'role', 'party_id', 'component_name'):
if request.json.get(i) is not None:
query[i] = request.json[i]
if query.get('role') is not None:
if query['role'] not in valid_query_parameters['role']:
return error_response(400, f"Invalid parameter 'role'.")
if query.get('party_id') is not None:
try:
query['party_id'] = int(query['party_id'])
except Exception:
return error_response(400, f"Invalid parameter 'party_id'.")
tasks, count = job_utils.list_task(limit, offset, query,
parse_order_by(('create_time', 'asc')))
return get_json_result(
data={
'tasks': [task.to_human_model_dict() for task in tasks],
'count': count,
})
def authentication():
if request.json and request.form:
return error_response(400)
if not (HTTP_APP_KEY and HTTP_SECRET_KEY):
return
for i in [
'TIMESTAMP',
'NONCE',
'APP_KEY',
'SIGNATURE',
]:
if not request.headers.get(i):
return error_response(401)
try:
timestamp = int(request.headers['TIMESTAMP']) / 1000
except Exception:
return error_response(400, 'Invalid TIMESTAMP')
now = time()
if not now - MAX_TIMESTAMP_INTERVAL < timestamp < now + MAX_TIMESTAMP_INTERVAL:
return error_response(
425,
f'TIMESTAMP is more than {MAX_TIMESTAMP_INTERVAL} seconds away from the server time'
)
if not request.headers['NONCE']:
return error_response(400, 'Invalid NONCE')
if request.headers['APP_KEY'] != HTTP_APP_KEY:
return error_response(401, 'Unknown APP_KEY')
signature = b64encode(
HMAC(
HTTP_SECRET_KEY.encode('ascii'),
b'\n'.join([
request.headers['TIMESTAMP'].encode('ascii'),
request.headers['NONCE'].encode('ascii'),
request.headers['APP_KEY'].encode('ascii'),
request.full_path.rstrip('?').encode('ascii'),
request.data if request.json else b'',
# quote_via: `urllib.parse.quote` replaces spaces with `%20`
# safe: unreserved characters from rfc3986
urlencode(
sorted(request.form.items()), quote_via=quote,
safe='-._~').encode('ascii') if request.form else b'',
]),
'sha1').digest()).decode('ascii')
if signature != request.headers['SIGNATURE']:
return error_response(403)
def load_checkpoints():
required_args = [
'role', 'party_id', 'model_id', 'model_version', 'component_name'
]
try:
check_config(request.json, required_args)
except Exception as e:
abort(error_response(400, str(e)))
checkpoint_manager = CheckpointManager(**{
i: request.json[i]
for i in required_args
},
mkdir=False)
checkpoint_manager.load_checkpoints_from_disk()
return checkpoint_manager
def table_download():
request_data = request.json
from fate_flow.component_env_utils.env_utils import import_component_output_depend
import_component_output_depend()
data_table_meta = storage.StorageTableMeta(
name=request_data.get("name"), namespace=request_data.get("namespace"))
if not data_table_meta:
return error_response(
response_code=210,
retmsg=
f'no found table:{request_data.get("namespace")}, {request_data.get("name")}'
)
tar_file_name = 'table_{}_{}.tar.gz'.format(request_data.get("namespace"),
request_data.get("name"))
return TableStorage.send_table(
output_tables_meta={"table": data_table_meta},
tar_file_name=tar_file_name,
need_head=request_data.get("head", True))
def get_predict_conf():
request_data = request.json
required_parameters = ['model_id', 'model_version']
check_config(request_data, required_parameters)
model_dir = os.path.join(get_project_base_directory(), 'model_local_cache')
model_fp_list = glob.glob(
model_dir +
f"/guest#*#{request_data['model_id']}/{request_data['model_version']}")
if model_fp_list:
fp = model_fp_list[0]
pipeline_model = PipelinedModel(model_id=fp.split('/')[-2],
model_version=fp.split('/')[-1])
pipeline = pipeline_model.read_component_model('pipeline',
'pipeline')['Pipeline']
predict_dsl = json_loads(pipeline.inference_dsl)
train_runtime_conf = json_loads(pipeline.train_runtime_conf)
parser = schedule_utils.get_dsl_parser_by_version(
train_runtime_conf.get('dsl_version', '1'))
predict_conf = parser.generate_predict_conf_template(
predict_dsl=predict_dsl,
train_conf=train_runtime_conf,
model_id=request_data['model_id'],
model_version=request_data['model_version'])
else:
predict_conf = ''
if predict_conf:
if request_data.get("filename"):
os.makedirs(TEMP_DIRECTORY, exist_ok=True)
temp_filepath = os.path.join(TEMP_DIRECTORY,
request_data.get("filename"))
with open(temp_filepath, "w") as fout:
fout.write(json_dumps(predict_conf, indent=4))
return send_file(open(temp_filepath, "rb"),
as_attachment=True,
attachment_filename=request_data.get("filename"))
else:
return get_json_result(data=predict_conf)
return error_response(
210,
"No model found, please check if arguments are specified correctly.")
def job_log():
job_id = request.json.get('job_id', '')
job_log_dir = job_utils.get_job_log_directory(job_id=job_id)
if os.path.exists(job_log_dir):
memory_file = io.BytesIO()
tar = tarfile.open(fileobj=memory_file, mode='w:gz')
for root, dir, files in os.walk(job_log_dir):
for file in files:
full_path = os.path.join(root, file)
rel_path = os.path.relpath(full_path, job_log_dir)
tar.add(full_path, rel_path)
tar.close()
memory_file.seek(0)
return send_file(
memory_file,
attachment_filename='job_{}_log.tar.gz'.format(job_id),
as_attachment=True)
else:
return error_response(
210,
"Log file path: {} not found. Please check if the job id is valid."
.format(job_log_dir))
def register():
info = request.json or request.form.to_dict()
if not Path(info["path"]).is_dir():
return error_response(400, "invalid path")
provider = ComponentProvider(
name=info["name"],
version=info["version"],
path=info["path"],
class_path=info.get("class_path",
ComponentRegistry.get_default_class_path()))
code, std = WorkerManager.start_general_worker(
worker_name=WorkerName.PROVIDER_REGISTRAR, provider=provider)
if code == 0:
ComponentRegistry.load()
if ComponentRegistry.get_providers().get(provider.name, {}).get(
provider.version, None) is None:
return get_json_result(retcode=RetCode.OPERATING_ERROR,
retmsg=f"not load into memory")
else:
return get_json_result()
else:
return get_json_result(retcode=RetCode.OPERATING_ERROR,
retmsg=f"register failed:\n{std}")
def operate_model(model_operation):
request_config = request.json or request.form.to_dict()
job_id = job_utils.generate_job_id()
if model_operation not in [
ModelOperation.STORE, ModelOperation.RESTORE,
ModelOperation.EXPORT, ModelOperation.IMPORT
]:
raise Exception(
'Can not support this operating now: {}'.format(model_operation))
required_arguments = ["model_id", "model_version", "role", "party_id"]
check_config(request_config, required_arguments=required_arguments)
request_config["model_id"] = gen_party_model_id(
model_id=request_config["model_id"],
role=request_config["role"],
party_id=request_config["party_id"])
if model_operation in [ModelOperation.EXPORT, ModelOperation.IMPORT]:
if model_operation == ModelOperation.IMPORT:
try:
file = request.files.get('file')
file_path = os.path.join(TEMP_DIRECTORY, file.filename)
# if not os.path.exists(file_path):
# raise Exception('The file is obtained from the fate flow client machine, but it does not exist, '
# 'please check the path: {}'.format(file_path))
try:
os.makedirs(os.path.dirname(file_path), exist_ok=True)
file.save(file_path)
except Exception as e:
shutil.rmtree(file_path)
raise e
request_config['file'] = file_path
model = pipelined_model.PipelinedModel(
model_id=request_config["model_id"],
model_version=request_config["model_version"])
model.unpack_model(file_path)
pipeline = model.read_component_model('pipeline',
'pipeline')['Pipeline']
train_runtime_conf = json_loads(pipeline.train_runtime_conf)
permitted_party_id = []
for key, value in train_runtime_conf.get('role', {}).items():
for v in value:
permitted_party_id.extend([v, str(v)])
if request_config["party_id"] not in permitted_party_id:
shutil.rmtree(model.model_path)
raise Exception(
"party id {} is not in model roles, please check if the party id is valid."
)
try:
adapter = JobRuntimeConfigAdapter(train_runtime_conf)
job_parameters = adapter.get_common_parameters().to_dict()
with DB.connection_context():
db_model = MLModel.get_or_none(
MLModel.f_job_id == job_parameters.get(
"model_version"),
MLModel.f_role == request_config["role"])
if not db_model:
model_info = model_utils.gather_model_info_data(model)
model_info['imported'] = 1
model_info['job_id'] = model_info['f_model_version']
model_info['size'] = model.calculate_model_file_size()
model_info['role'] = request_config["model_id"].split(
'#')[0]
model_info['party_id'] = request_config[
"model_id"].split('#')[1]
if model_utils.compare_version(
model_info['f_fate_version'], '1.5.1') == 'lt':
model_info['roles'] = model_info.get(
'f_train_runtime_conf', {}).get('role', {})
model_info['initiator_role'] = model_info.get(
'f_train_runtime_conf',
{}).get('initiator', {}).get('role')
model_info['initiator_party_id'] = model_info.get(
'f_train_runtime_conf',
{}).get('initiator', {}).get('party_id')
model_info[
'work_mode'] = adapter.get_job_work_mode()
model_info['parent'] = False if model_info.get(
'f_inference_dsl') else True
model_utils.save_model_info(model_info)
else:
stat_logger.info(
f'job id: {job_parameters.get("model_version")}, '
f'role: {request_config["role"]} model info already existed in database.'
)
except peewee.IntegrityError as e:
stat_logger.exception(e)
operation_record(request_config, "import", "success")
return get_json_result()
except Exception:
operation_record(request_config, "import", "failed")
raise
else:
try:
model = pipelined_model.PipelinedModel(
model_id=request_config["model_id"],
model_version=request_config["model_version"])
if model.exists():
archive_file_path = model.packaging_model()
operation_record(request_config, "export", "success")
return send_file(archive_file_path,
attachment_filename=os.path.basename(
archive_file_path),
as_attachment=True)
else:
operation_record(request_config, "export", "failed")
res = error_response(
response_code=210,
retmsg="Model {} {} is not exist.".format(
request_config.get("model_id"),
request_config.get("model_version")))
return res
except Exception as e:
operation_record(request_config, "export", "failed")
stat_logger.exception(e)
return error_response(response_code=210, retmsg=str(e))
else:
data = {}
job_dsl, job_runtime_conf = gen_model_operation_job_config(
request_config, model_operation)
submit_result = DAGScheduler.submit(
{
'job_dsl': job_dsl,
'job_runtime_conf': job_runtime_conf
},
job_id=job_id)
data.update(submit_result)
operation_record(data=job_runtime_conf,
oper_type=model_operation,
oper_status='')
return get_json_result(job_id=job_id, data=data)
def operate_model(model_operation):
request_config = request.json or request.form.to_dict()
job_id = job_utils.generate_job_id()
if model_operation not in [
ModelOperation.STORE, ModelOperation.RESTORE,
ModelOperation.EXPORT, ModelOperation.IMPORT
]:
raise Exception(
'Can not support this operating now: {}'.format(model_operation))
required_arguments = ["model_id", "model_version", "role", "party_id"]
check_config(request_config, required_arguments=required_arguments)
request_config["model_id"] = gen_party_model_id(
model_id=request_config["model_id"],
role=request_config["role"],
party_id=request_config["party_id"])
if model_operation in [ModelOperation.EXPORT, ModelOperation.IMPORT]:
if model_operation == ModelOperation.IMPORT:
try:
file = request.files.get('file')
file_path = os.path.join(TEMP_DIRECTORY, file.filename)
# if not os.path.exists(file_path):
# raise Exception('The file is obtained from the fate flow client machine, but it does not exist, '
# 'please check the path: {}'.format(file_path))
try:
os.makedirs(os.path.dirname(file_path), exist_ok=True)
file.save(file_path)
except Exception as e:
shutil.rmtree(file_path)
raise e
request_config['file'] = file_path
model = pipelined_model.PipelinedModel(
model_id=request_config["model_id"],
model_version=request_config["model_version"])
model.unpack_model(file_path)
pipeline = model.read_component_model('pipeline',
'pipeline')['Pipeline']
train_runtime_conf = json_loads(pipeline.train_runtime_conf)
permitted_party_id = []
for key, value in train_runtime_conf.get('role', {}).items():
for v in value:
permitted_party_id.extend([v, str(v)])
if request_config["party_id"] not in permitted_party_id:
shutil.rmtree(model.model_path)
raise Exception(
"party id {} is not in model roles, please check if the party id is valid."
)
try:
with DB.connection_context():
model = MLModel.get_or_none(
MLModel.f_job_id == train_runtime_conf[
"job_parameters"]["model_version"],
MLModel.f_role == request_config["role"])
if not model:
MLModel.create(
f_role=request_config["role"],
f_party_id=request_config["party_id"],
f_roles=train_runtime_conf["role"],
f_job_id=train_runtime_conf["job_parameters"]
["model_version"],
f_model_id=train_runtime_conf["job_parameters"]
["model_id"],
f_model_version=train_runtime_conf[
"job_parameters"]["model_version"],
f_initiator_role=train_runtime_conf[
"initiator"]["role"],
f_initiator_party_id=train_runtime_conf[
"initiator"]["party_id"],
f_runtime_conf=train_runtime_conf,
f_work_mode=train_runtime_conf[
"job_parameters"]["work_mode"],
f_dsl=json_loads(pipeline.train_dsl),
f_imported=1,
f_job_status='complete')
else:
stat_logger.info(
f'job id: {train_runtime_conf["job_parameters"]["model_version"]}, '
f'role: {request_config["role"]} model info already existed in database.'
)
except peewee.IntegrityError as e:
stat_logger.exception(e)
operation_record(request_config, "import", "success")
return get_json_result()
except Exception:
operation_record(request_config, "import", "failed")
raise
else:
try:
model = pipelined_model.PipelinedModel(
model_id=request_config["model_id"],
model_version=request_config["model_version"])
if model.exists():
archive_file_path = model.packaging_model()
operation_record(request_config, "export", "success")
return send_file(archive_file_path,
attachment_filename=os.path.basename(
archive_file_path),
as_attachment=True)
else:
operation_record(request_config, "export", "failed")
res = error_response(
response_code=210,
retmsg="Model {} {} is not exist.".format(
request_config.get("model_id"),
request_config.get("model_version")))
return res
except Exception as e:
operation_record(request_config, "export", "failed")
stat_logger.exception(e)
return error_response(response_code=210, retmsg=str(e))
else:
data = {}
job_dsl, job_runtime_conf = gen_model_operation_job_config(
request_config, model_operation)
job_id, job_dsl_path, job_runtime_conf_path, logs_directory, model_info, board_url = DAGScheduler.submit(
{
'job_dsl': job_dsl,
'job_runtime_conf': job_runtime_conf
},
job_id=job_id)
data.update({
'job_dsl_path': job_dsl_path,
'job_runtime_conf_path': job_runtime_conf_path,
'board_url': board_url,
'logs_directory': logs_directory
})
operation_record(data=job_runtime_conf,
oper_type=model_operation,
oper_status='')
return get_json_result(job_id=job_id, data=data)
def internal_server_error(e):
stat_logger.exception(e)
return error_response(500, str(e))
def download_model(party_model_id, model_version):
party_model_id = party_model_id.replace('~', '#')
model_data = publish_model.download_model(party_model_id, model_version)
if model_data is None:
return error_response(404, 'model not found')
return get_json_result(data=model_data)
def component_output_data_download():
request_data = request.json
try:
output_tables_meta = get_component_output_tables_meta(task_data=request_data)
except Exception as e:
stat_logger.exception(e)
return error_response(210, str(e))
limit = request_data.get('limit', -1)
if not output_tables_meta:
return error_response(response_code=210, retmsg='no data')
if limit == 0:
return error_response(response_code=210, retmsg='limit is 0')
have_data_label = False
output_data_file_list = []
output_data_meta_file_list = []
output_tmp_dir = os.path.join(os.getcwd(), 'tmp/{}'.format(fate_uuid()))
for output_name, output_table_meta in output_tables_meta.items():
output_data_count = 0
is_str = False
output_data_file_path = "{}/{}.csv".format(output_tmp_dir, output_name)
os.makedirs(os.path.dirname(output_data_file_path), exist_ok=True)
with open(output_data_file_path, 'w') as fw:
with storage.Session.build(name=output_table_meta.get_name(), namespace=output_table_meta.get_namespace()) as storage_session:
output_table = storage_session.get_table()
for k, v in output_table.collect():
data_line, have_data_label, is_str = get_component_output_data_line(src_key=k, src_value=v)
fw.write('{}\n'.format(','.join(map(lambda x: str(x), data_line))))
output_data_count += 1
if output_data_count == limit:
break
if output_data_count:
# get meta
output_data_file_list.append(output_data_file_path)
header = get_component_output_data_schema(output_table_meta=output_table_meta, have_data_label=have_data_label, is_str=is_str)
output_data_meta_file_path = "{}/{}.meta".format(output_tmp_dir, output_name)
output_data_meta_file_list.append(output_data_meta_file_path)
with open(output_data_meta_file_path, 'w') as fw:
json.dump({'header': header}, fw, indent=4)
if request_data.get('head', True) and header:
with open(output_data_file_path, 'r+') as f:
content = f.read()
f.seek(0, 0)
f.write('{}\n'.format(','.join(header)) + content)
# tar
memory_file = io.BytesIO()
tar = tarfile.open(fileobj=memory_file, mode='w:gz')
for index in range(0, len(output_data_file_list)):
tar.add(output_data_file_list[index], os.path.relpath(output_data_file_list[index], output_tmp_dir))
tar.add(output_data_meta_file_list[index], os.path.relpath(output_data_meta_file_list[index], output_tmp_dir))
tar.close()
memory_file.seek(0)
output_data_file_list.extend(output_data_meta_file_list)
for path in output_data_file_list:
try:
shutil.rmtree(os.path.dirname(path))
except Exception as e:
# warning
stat_logger.warning(e)
tar_file_name = 'job_{}_{}_{}_{}_output_data.tar.gz'.format(request_data['job_id'],
request_data['component_name'],
request_data['role'], request_data['party_id'])
return send_file(memory_file, attachment_filename=tar_file_name, as_attachment=True)
