datax-admin/app/services/jm_job.py

from asyncio import current_task
import json
import time
from turtle import update
from app import crud, models, schemas
from app.common import minio
from app.core.datasource.datasource import DataSourceBase
from app.crud.jm_homework_datasource_relation import get_jm_relations
from app.utils.send_util import *
from sqlalchemy.orm import Session
from app.common.minio import minio_client
from configs.settings import DefaultOption, config
DATABASE_NAME = config.get('HIVE', 'DATABASE_NAME')

type_dict = {
    "Java": "java",
    "Python": "python",
    "Dag": "sparks"
}

def jm_job_create_task(jm_homework: models.JmHomework, relation_list, db: Session):
    content = ''
    envs = {}
    if jm_homework.type == "Dag":
        content = content = red_dag_and_format(jm_homework, relation_list, db)
        requirements_relation = crud.get_requirements_relation(db, jm_homework.dag_uuid)
        if requirements_relation:
            envs.update({'requirement_package_path': f'jpt/requirements/dag_{jm_homework.dag_uuid.lower()})'})
    elif jm_homework.type == "Python":
        content = red_python_and_format(jm_homework)
    af_task = {
        "name": jm_homework.name,
        "file_urls": [] if jm_homework.type != "Java" else [jm_homework.script_file],
        "script": content if jm_homework.type != "Java" else "",
        "cmd": jm_homework.execute_command if jm_homework.type != "Dag" else "",
        "cmd_parameters": "",
        "envs": envs,
        "run_image": jm_homework.image_url if jm_homework.type != "Dag" else "",
        "task_type": type_dict[jm_homework.type],
        "user_id": 0,
    }
    res = send_post('/af/af_task', af_task)
    af_task = res['data']
    return af_task

def jm_job_update_task(jm_homework: models.JmHomework, relation_list, db: Session):
    relation = crud.get_af_id(db, jm_homework.id, 'task')
    content = ''
    envs = {}
    if jm_homework.type == "Dag":
        content = content = red_dag_and_format(jm_homework, relation_list, db)
        requirements_relation = crud.get_requirements_relation(db, jm_homework.dag_uuid)
        if requirements_relation:
            envs.update({'requirement_package_path': f'jpt/requirements/dag_{jm_homework.dag_uuid.lower()})'})
    elif jm_homework.type == "Python":
        content = red_python_and_format(jm_homework)
    af_task = {
        "name": jm_homework.name,
        "file_urls": [] if jm_homework.type != "Java" else [jm_homework.script_file],
        "script": content if jm_homework.type != "Java" else "",
        "cmd": jm_homework.execute_command if jm_homework.type != "Dag" else "",
        "cmd_parameters": "",
        "envs": envs,
        "run_image": jm_homework.image_url if jm_homework.type != "Dag" else "",
        "task_type": type_dict[jm_homework.type],
        "user_id": 0,
    }
    res = send_put('/af/af_task', relation.af_id, af_task)
    af_task = res['data']
    return af_task

def jm_job_create_job(jm_job_info: models.JmJobInfo, nodes, edges, db: Session):
    homework_ids = [node['homework_id'] for node in nodes]
    node_uuid_to_h_id = {node['id']:node['homework_id'] for node in nodes}
    relations = crud.get_af_ids(db,homework_ids, 'task')
    se_id_to_af_id_dict = { relation.se_id:relation.af_id for relation in relations}
    tasks = [ send_get("/af/af_task/getOnce",id)['data'] for id in se_id_to_af_id_dict.values()]
    dependence = [[se_id_to_af_id_dict[node_uuid_to_h_id[str(edge['source'])]],se_id_to_af_id_dict[node_uuid_to_h_id[str(edge['target'])]]] for edge in edges]
    cron = jm_job_info.cron_expression if jm_job_info.cron_type == 2 else '@once'
    cron = cron.replace('?','*')
    af_job = {
        "tasks": tasks,
        "name": jm_job_info.name,
        "dependence": dependence,
        "cron": cron,
        "desc": jm_job_info.name,
        "route_strategy": "",
        "block_strategy": "",
        "executor_timeout": 0,
        "executor_fail_retry_count": 0,
        "trigger_status": jm_job_info.status,
        "job_mode":1,
        "job_type": 0,
        "user_id": 0,
    }
    res = send_post('/af/af_job', af_job)
    af_job = res['data']
    send_submit(af_job['id'])
    return af_job


def jm_job_update_job(jm_job_info: models.JmJobInfo, nodes, edges, db: Session):
    homework_ids = [node['homework_id'] for node in nodes]
    node_uuid_to_h_id = {node['id']:node['homework_id'] for node in nodes}
    relations = crud.get_af_ids(db,homework_ids, 'task')
    se_id_to_af_id_dict = { relation.se_id:relation.af_id for relation in relations}
    tasks = [ send_get("/af/af_task/getOnce",id)['data'] for id in se_id_to_af_id_dict.values()]
    dependence = [[se_id_to_af_id_dict[node_uuid_to_h_id[str(edge['source'])]],se_id_to_af_id_dict[node_uuid_to_h_id[str(edge['target'])]]] for edge in edges]
    cron = jm_job_info.cron_expression if jm_job_info.cron_type == 2 else '@once'
    cron = cron.replace('?','*')
    af_job = {
        "tasks": tasks,
        "name": jm_job_info.name,
        "dependence": dependence,
        "cron": cron,
        "desc": jm_job_info.name,
        "route_strategy": "",
        "block_strategy": "",
        "executor_timeout": 0,
        "executor_fail_retry_count": 0,
        "trigger_status": jm_job_info.status,
    }
    job_relation = crud.get_af_id(db,jm_job_info.id,'job')
    res = send_put('/af/af_job', job_relation.af_id, af_job)
    af_job = res['data']
    send_submit(af_job['id'])

def red_dag_and_format(jm_homework: models.JmHomework, relation_list, db: Session):
    node_relation_dict = { relation['node_uuid']:relation for relation in relation_list}
    file = minio_client.get_file(jm_homework.dag_url)
    result = json.loads(file)
    edges = result['edges']
    # t_s = {}
    # input_num = {}
    # for edge in edges:
    #     if edge['target'] in t_s.keys():
    #         t_s[edge['target']].append(edge['source'])
    #     else:
    #         t_s.update({edge['target']:[edge['source']]})
    nodes = result['nodes']
    sub_nodes = []
    for node in nodes:
        if node['op'] == 'datasource':
            fileds = node['data']['input_source']
            script = 'select '
            for filed in fileds:
                script += filed['dataField'] + ','
            script = script.strip(',')
            ds_id = node_relation_dict[node['id']]['datasource_id']
            database_name = ""
            if ds_id  == -1:
                database_name = DATABASE_NAME
            else:
                data_source = crud.get_job_jdbc_datasource(db,ds_id)
                database_name = data_source.database_name
            script += ' from ' + database_name + '.'+node_relation_dict[node['id']]['table']
            sub_node = {
                "id": node['id'],
                "name": node['name'],
                "op": 'sql',
                "script":script
            }
            sub_nodes.append(sub_node)
        elif node['op'] == 'outputsource':
            fileds = node['data']['output_source']
            ds_id = node_relation_dict[node['id']]['datasource_id']
            database_name = ""
            if ds_id  == -1:
                database_name = DATABASE_NAME
            else:
                data_source = crud.get_job_jdbc_datasource(db,ds_id)
                database_name = data_source.database_name
            script = '''def main_func (input0, spark,sc):
    input0.write.mode("overwrite").saveAsTable("''' + database_name + '.'+node_relation_dict[node['id']]['table']+'''")'''
            inputs = node['data']['inputs'] if 'inputs' in node['data'].keys() else {}
            # index = 0
            # input_list = t_s[node['id']] if node['id'] in t_s.keys() else []
            # for input in input_list:
            #     if input in input_num.keys():
            #         input_num[input]+=1
            #     else:
            #         input_num.update({input:0})
            #     inputs.update({'input'+str(index):[input,input_num[input]]})
            #     index+=1
            sub_node = {
                "id": node['id'],
                "name": node['name'],
                "op": 'pyspark',
                "inputs": inputs,
                "script":script
            }
            sub_nodes.append(sub_node)
        else:
            inputs = node['data']['inputs'] if 'inputs' in node['data'].keys() else {}
            # index = 0
            # input_list = t_s[node['id']] if node['id'] in t_s.keys() else []
            # for input in input_list:
            #     if input in input_num.keys():
            #         input_num[input]+=1
            #     else:
            #         input_num.update({input:0})
            #     inputs.update({'input'+str(index):[input,input_num[input]]})
            #     index+=1
            script = node['data']['script']
            if node['op'] == 'sql':
                script = script.replace('\n', ' ')
            sub_node = {
                "id": node['id'],
                "name": node['name'],
                "op": node['op'],
                "inputs": inputs,
                "script": script,
            }
            sub_nodes.append(sub_node)
    res = {
        'sub_nodes': sub_nodes,
        'edges': [(edge['source'],edge['target']) for edge in edges]
    }
    return json.dumps(res)

def red_python_and_format(jm_homework):
    file = minio_client.get_file(jm_homework.script_file if jm_homework.script_file else "/python/test.py")
    return file.decode("utf-8")

def on_off_control(af_job_id: int,status: int):
    for i in range(0,11):
        parsed_res = get_job_last_parsed_time(af_job_id)
        last_parsed_time = parsed_res['data']['last_parsed_time']
        if last_parsed_time:
            send_pause(af_job_id,status)
            print(f"{af_job_id}<==状态修改成功==>{last_parsed_time}")
            break
        if i >= 10:
            raise Exception(f"{af_job_id}==>状态修改失败")
        time.sleep(2)

def execute_job(af_job_id: int):
    res = get_job_last_parsed_time(af_job_id)
    current_time = res['data']['last_parsed_time'] if 'last_parsed_time' in res['data'].keys() else None
    send_submit(af_job_id)
    for i in range(0,21):
        parsed_res = get_job_last_parsed_time(af_job_id)
        last_parsed_time = parsed_res['data']['last_parsed_time']
        if last_parsed_time and last_parsed_time != current_time:
            res = send_execute(af_job_id)
            print(f"{af_job_id}<==任务执行成功==>{last_parsed_time}")
            return res
        if i >= 20:
            raise Exception(f"{af_job_id}==>文件正在转化中")
        time.sleep(2)