class WordsegMarkJobListResource(Resource):
@parse({
"is_superuser": fields.Boolean(missing=False),
"query": fields.String(missing=''),
"offset": fields.Integer(missing=0),
"limit": fields.Integer(missing=10),
"doc_type_id": fields.Integer(missing=None),
'order_by': fields.String(missing='-created_time'),
})
def get(self: Resource,
args: typing.Dict) -> typing.Tuple[typing.Dict, int]:
count, result = MarkJobService().get_mark_job_list_by_nlp_task(
args, nlp_task=NlpTaskEnum.wordseg)
return {
"message": "请求成功",
"result": result,
"count": count,
}, 200
@parse(
{
"mark_job_name":
fields.String(required=True),
"mark_job_type":
fields.String(required=True),
"mark_job_desc":
fields.String(),
"doc_type_id":
fields.Integer(required=True),
"files":
fields.List(fields.File(), required=True),
"assign_mode":
fields.String(required=True,
validate=lambda x: x in ['average', 'together']),
"assessor_id":
fields.Integer(missing=0),
"labeler_ids":
fields.List(fields.Integer(), required=True),
},
locations=('form', 'files'))
def post(self: Resource,
args: typing.Dict) -> typing.Tuple[typing.Dict, int]:
files = args["files"]
assign_mode = args["assign_mode"]
if assign_mode == AssignModeEnum.together:
abort(400, message="不支持共同标注")
job_type = Common().check_job_type_by_files(files)
if job_type != "text":
abort(400, message="请上传纯文本文档（txt/csv）")
else:
args['mark_job_type'] = job_type
try:
result = MarkJobService().create_mark_job(files,
NlpTaskEnum.wordseg,
args)
return {"message": "创建成功", "result": result}, 201
except TypeError:
abort(400, message="上传文件类型错误")
class PredictJobSchema(Schema):
doc_type = fields.Nested(DocTypeSchema)
task_list = fields.List(fields.Nested(PredictTaskSchema))
extract_job_id = fields.Integer(attribute="predict_job_id")
extract_job_name = fields.String(attribute="predict_job_name")
extract_job_type = fields.String(attribute="predict_job_type.value")
extract_job_state = fields.Function(lambda obj: StatusEnum(obj.predict_job_status).name)
extract_job_desc = fields.String(attribute="predict_job_desc")
is_batch = fields.Boolean()
created_time = fields.DateTime()
class RelationMarkJobListResource(Resource):
@parse({
"is_superuser": fields.Boolean(missing=False),
"query": fields.String(missing=''),
"offset": fields.Integer(missing=0),
"limit": fields.Integer(missing=10),
"doc_type_id": fields.Integer(missing=None),
'order_by': fields.String(missing='-created_time'),
})
def get(self: Resource,
args: typing.Dict) -> typing.Tuple[typing.Dict, int]:
count, result = MarkJobService().get_mark_job_list_by_nlp_task(
args, NlpTaskEnum.relation)
return {
"message": "请求成功",
"result": result,
"count": count,
}, 200
@parse(
{
"mark_job_name":
fields.String(required=True),
"mark_job_type":
fields.String(required=True),
"mark_job_desc":
fields.String(),
"doc_type_id":
fields.Integer(required=True),
"files":
fields.List(fields.File(), required=True),
"assign_mode":
fields.String(required=True,
validate=lambda x: x in ['average', 'together']),
"assessor_id":
fields.Integer(),
"labeler_ids":
fields.List(fields.Integer(), required=True),
"use_rule":
fields.Integer(missing=1) # 默认使用规则
},
locations=('form', 'files'))
def post(self: Resource,
args: typing.Dict) -> typing.Tuple[typing.Dict, int]:
files = args['files']
job_type = Common().check_job_type_by_files(files)
if not job_type:
abort(400, message='请上传全部纯文本文档（txt/csv）或者全部电子文档（pdf/word文档）')
else:
args['mark_job_type'] = job_type
result = MarkJobService().create_mark_job(files, NlpTaskEnum.relation,
args)
return {"message": "创建成功", "result": result}, 201
class DocTypeSchema(Schema):
doc_terms = fields.List(fields.Integer())
doc_term_list = fields.List(fields.Nested(DocTermSchema))
doc_relation_list = fields.List(fields.Nested(EntityDocRelationSchema))
doc_lexicon_list = fields.List(fields.Nested(WordsegDocLexiconSchema),
attribute='doc_rules')
doc_type_id = fields.Integer()
doc_type_name = fields.String()
doc_type_desc = fields.String()
is_top = fields.Boolean(attribute="is_favorite")
created_time = fields.DateTime()
group_id = fields.Integer()
status = fields.Function(lambda obj: not obj.is_deleted)
class UpdateModelTrainResource(Resource):
@parse({
"model_version": fields.String(required=True),
"check_train_terms": fields.Boolean(required=True),
"model_train_state": fields.String(),
"model_train_result": fields.Dict(),
"model_type": fields.String(required=True,
validate=lambda x: x in ('extract', 'classify', 'relation', 'wordseg'))
})
def put(self, args):
"""
修改模型的状态
"""
update_params = {}
if args.get("model_train_state"):
# 后端返回结果转换，失败时后端目前还返回failed
update_params.update(train_status=status_str2int_mapper()[args["model_train_state"]])
train_task = ModelTrainService().update_train_task_by_model_version(model_version=args["model_version"], is_check_train_terms=args["check_train_terms"], args=update_params)
result = TrainTaskSchema().dump(train_task)
return {
"message": "更新成功",
"result": result,
}, 201
class ModelTrainItemResource(Resource):
@parse({
"model_type": fields.String(required=True, validate=lambda x: x in ("extract", "classify", "relation", "wordseg"))
})
def get(
self: Resource,
args: typing.Dict,
model_id: int,
model_train_id: int
) -> typing.Tuple[typing.Dict, int]:
"""
获取单条模型训练记录
"""
# get single task task by id
train_task = ModelTrainService().get_train_task_by_id(train_task_id=model_train_id)
result = TrainTaskSchema().dump(train_task)
# add extra algorithm information for extract and relation.
if args["model_type"] in ["extract", "relation"]:
result = ModelTrainService().add_algo_dict_for_extract_relation(result)
return {
"message": "请求成功",
"result": result,
}, 200
@parse({
"model_train_state": fields.String(),
"model_train_result": fields.Dict(),
"check_train_terms": fields.Boolean(missing=False),
"model_type": fields.String(required=True, validate=lambda x: x in ("extract", "classify", "relation", "wordseg"))
})
def patch(
self: Resource,
args: typing.Dict,
model_id: int,
model_train_id: int
) -> typing.Tuple[typing.Dict, int]:
"""
修改模型的状态和结果
"""
update_params = {}
if args.get("model_train_state"): # 这里不考虑model_train_result因为新的表结构里没有这个列了
update_params.update(train_status=status_str2int_mapper()[args["model_train_state"]])
train_task = ModelTrainService().update_train_task_by_id(train_job_id=model_id, train_task_id=model_train_id,
is_check_train_terms=args["check_train_terms"], model_type = args["model_type"],
args=update_params)
result = TrainTaskSchema().dump(train_task)
return {
"message": "更新成功",
"result": result,
}, 200
@parse({
"model_type": fields.String(required=True, validate=lambda x: x in ("extract", "classify"))
})
def delete(
self: Resource,
args: typing.Dict,
model_id: int,
model_train_id: int
) -> typing.Tuple[typing.Dict, int]:
"""
删除模型
"""
ModelTrainService().delete_train_task_by_id(train_task_id=model_train_id)
return {
"message": "删除成功",
}, 200