class WordsegMarkJobImportResource(Resource):
@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),
"task_type":
fields.String(required=True,
validate=lambda x: x in ['machine', 'manual']),
},
locations=('form', 'files'))
def post(self: Resource,
args: typing.Dict) -> typing.Tuple[typing.Dict, int]:
"""
上传已标注数据
"""
files = args['files']
# validate file extensions
for f in files:
if get_ext(f.filename) not in ["txt"]:
abort(400, message="上传已标注分词数据仅支持txt格式。")
result = MarkJobService().import_mark_job(files,
args,
nlp_task=NlpTaskEnum.wordseg)
return {"message": "创建成功", "result": result}, 201
class ClassifyMarkJobImportResource(Resource):
@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),
}, locations=('form', 'files'))
def post(
self: Resource,
args: typing.Dict
) -> typing.Tuple[typing.Dict, int]:
files = args['files']
for f in files:
if get_ext(f.filename) not in ["csv"]:
abort(400, message="已标注分类数据仅支持csv格式。")
try:
result = MarkJobService().import_mark_job(files, args, nlp_task=NlpTaskEnum.classify)
return {
"message": "创建成功",
"result": result
}, 201
except UnicodeDecodeError:
abort(400, message="文件编码错误 请上传utf-8编码文件")
except KeyError:
abort(400, message="文件格式不合规 请查看csv文件模版")
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 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 ExtractMarkJobImportResource(Resource):
@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),
}, locations=('form', 'files'))
def post(self: Resource, args: typing.Dict):
files = args['files']
args['task_type'] = 'manual'
# validate file extensions
for f in files:
if get_ext(f.filename) not in ["txt"]:
abort(400, message="导入已标注序列标注数据仅支持txt格式。")
result = MarkJobService().import_mark_job(files, args, nlp_task=NlpTaskEnum.extract)
return {
"message": "创建成功",
"result": result
}, 201
class ExtractJobListResource(Resource, CurrentUserMixin):
@parse({
"offset": fields.Integer(missing=0),
"limit": fields.Integer(missing=10),
"query": fields.String(missing=''),
"doc_type_id": fields.Integer(missing=0),
"order_by": fields.String(missing='-created_time'),
})
def get(self: Resource,
args: typing.Dict) -> typing.Tuple[typing.Dict, int]:
nlp_task_id = Common().get_nlp_task_id_by_route()
order_by = args["order_by"][1:]
order_by_desc = True if args["order_by"][0] == "-" else False
count, predict_job_list = PredictService(
).get_predict_job_list_by_nlp_task_id(
nlp_task_id=nlp_task_id,
doc_type_id=args['doc_type_id'],
search=args['query'],
order_by=order_by,
order_by_desc=order_by_desc,
offset=args['offset'],
limit=args['limit'],
current_user=self.get_current_user())
# get the serialized result
result = PredictJobSchema().dump(predict_job_list, many=True)
return {
"message": "请求成功",
"result": result,
"count": count,
}, 200
@parse(
{
"extract_job_name":
fields.String(required=True),
"extract_job_type":
fields.String(required=True),
"extract_job_desc":
fields.String(missing=""),
"doc_type_id":
fields.Integer(required=True),
"files":
fields.List(fields.File(), required=True),
"task_type":
fields.String(required=True,
validate=lambda x: x in ['machine', 'manual']),
"use_rule":
fields.Integer(missing=0)
},
locations=('form', 'files'))
def post(self: Resource,
args: typing.Dict) -> typing.Tuple[typing.Dict, int]:
predict_job = PredictService().create_predict_job_by_doc_type_id(
doc_type_id=args["doc_type_id"],
predict_job_name=args["extract_job_name"],
predict_job_desc=args["extract_job_desc"],
predict_job_type=args["extract_job_type"],
files=args["files"],
use_rule=args["use_rule"])
result = PredictJobSchema().dump(predict_job)
return {"message": "创建成功", "result": result}, 201