parser/healthcareio/server/proxy.py

"""
    This file serves as proxy to healthcare-io, it will be embedded into the API
"""
import os
import transport
import numpy as np
from healthcareio import x12
import pandas as pd
import smart
from healthcareio.analytics import Apex
import time
class get :
    PROCS = []
    PATH = os.sep.join([os.environ['HOME'],'.healthcareio','config.json'])
    @staticmethod
    def resume (files,args):
        """
        This function will determine the appropriate files to be processed by performing a simple complementary set operation against the logs
        @TODO: Support data-stores other than mongodb
        :param files   list of files within a folder
        :param _args    configuration
        """
        _args = args['store'].copy()
        if 'mongo' in _args['type'] :
            _args['type'] = 'mongo.MongoReader'
            reader = transport.factory.instance(**_args)
        _files = []
        try:
            pipeline = [{"$match":{"completed":{"$eq":True}}},{"$group":{"_id":"$name"}},{"$project":{"name":"$_id","_id":0}}]
            _args = {"aggregate":"logs","cursor":{},"allowDiskUse":True,"pipeline":pipeline}
            _files = reader.read(mongo = _args)
            _files = [item['name'] for item in _files]
        except Exception as e :
            pass
        print ( [len(list(set(files) - set(_files))),' files to be processed'])
        return list(set(files) - set(_files))

    @staticmethod
    def processes(_args):
        APP_NAME ='healthcare-io'
        _info = smart.top.read(name=APP_NAME) #pd.DataFrame(smart.top.read(name='healthcare-io'))[['name','cpu','mem']]
        
        
        if _info.shape[0] == 0 :
            _info = pd.DataFrame({"name":[APP_NAME],"cpu":[0],"mem":[0]})
        # _info = pd.DataFrame(_info.groupby(['name']).sum())
        # _info['name'] = ['healthcare-io.py']
        m = {'cpu':'CPU','mem':'RAM','name':'name'}
        _info  = _info.rename(columns=m)
        # _info.columns = [m[name] for name in _info.columns.tolist() if name in m]
        _info.index = np.arange(_info.shape[0])

        charts = []
        for label in ['CPU','RAM'] :
            value = _info[label].sum()
            df = pd.DataFrame({"name":[label],label:[value]})
            charts.append (
                Apex.apply(
                    {"data":df, "chart":{"type":"radial","axis":{"x":label,"y":"name"}}}
                    )['apex']
                )
        
        return {"process":{"chart":charts,"counts":_info.shape[0]}}
    @staticmethod
    def files (_args):
        folder = _args['args']['folder']
        _info = smart.folder.read(path=folder)
        
        N = _info.files.tolist()[0]
        store_args = _args['store'].copy()
        store_args['context'] = 'read'
        
        # if 'mongo' in _args['store']['type'] :
        if _args['store']['provider'] in ['mongo', 'mongodb']:
            # store_args = dict(_args['store'].copy(),**{"type":"mongo.MongoReader"})
            # reader = transport.factory.instance(**_args)
            
            pipeline = [{"$group":{"_id":"$name","count":{"$sum":{"$cond":[{"$eq":["$completed",True]},1,0]}} }},{"$group":{"_id":None,"count":{"$sum":"$count"}}},{"$project":{"_id":0,"status":"completed","count":1}}]
            query = {"mongo":{"aggregate":"logs","allowDiskUse":True,"cursor":{},"pipeline":pipeline}}
            # _info = pd.DataFrame(reader.read(mongo={"aggregate":"logs","allowDiskUse":True,"cursor":{},"pipeline":pipeline}))
            pipeline = [{"$group":{"_id":"$parse","claims":{"$addToSet":"$name"}}},{"$project":{"_id":0,"type":"$_id","count":{"$size":"$claims"}}}]
            _query = {"mongo":{"aggregate":"logs","cursor":{},"allowDiskUse":True,"pipeline":pipeline}} #-- distribution claims/remits


        else:
            # store_args = dict(_args['store'].copy(),**{"type":"disk.SQLiteReader"})
            
            # store_args['args']['table'] = 'logs'
            store_args['table'] = 'logs'
            query= {"sql":"select count(distinct json_extract(data,'$.name')) as count, 'completed' status from logs where json_extract(data,'$.completed') = true"}
            _query={"sql":"select json_extract(data,'$.parse') as type,count(distinct json_extract(data,'$.name')) as count from logs group by type"} #-- distribution claim/remits
        reader = transport.factory.instance(**store_args)
        _info = pd.DataFrame(reader.read(**query))
        
        if not _info.shape[0] :
            _info = pd.DataFrame({"status":["completed"],"count":[0]})
        _info['count'] = np.round( (_info['count'] * 100 )/N,2)
        
        charts = [Apex.apply({"data":_info,"chart":{"type":"radial","axis":{"y":"status","x":"count"}}})['apex']]
        #
        # Let us classify the files now i.e claims / remits
        #
        r = pd.DataFrame(reader.read(**_query)) #-- distribution claims/remits
        r = Apex.apply({"chart":{"type":"donut","axis":{"x":"count","y":"type"}},"data":r})['apex']
        r['chart']['height'] = '100%'
        r['legend']['position'] = 'bottom'

        charts += [r]

        
        return {"files":{"counts":N,"chart":charts}}

        pass
#
# Process handling ....


def run (_args) :
    """
    This function will run the jobs and insure as processes (as daemons).
    :param _args    system configuration
    """
    FILES = []
    BATCH = int(_args['args']['batch']) #-- number of processes (poorly named variable)

    for root,_dir,f in os.walk(_args['args']['folder']) :
        if f :
            FILES += [os.sep.join([root,name]) for name in f]
    FILES = get.resume(FILES,_args)
    FILES = np.array_split(FILES,BATCH)
    
    for FILE_GROUP in FILES :
        
        FILE_GROUP = FILE_GROUP.tolist()
        # logger.write({"process":index,"parse":_args['parse'],"file_count":len(row)})
        # proc = Process(target=apply,args=(row,info['store'],_info,))
        parser = x12.Parser(get.PATH) #os.sep.join([PATH,'config.json']))
        parser.set.files(FILE_GROUP)   
        parser.daemon = True
        parser.start()
        get.PROCS.append(parser)     
        time.sleep(3)
    #
    # @TODO:consider submitting an update to clients via publish/subscribe framework
    #
    return get.PROCS
def stop(_args):
    for job in get.PROCS :
        if job.is_alive() :
            job.terminate()
    get.PROCS = []
    #
    # @TODO: consider submitting an update to clients via publish/subscribe framework
    pass
def write(src_args,dest_args,files) :
    #
    # @TODO: Support for SQLite
    pass
def publish (src_args,dest_args,folder="/data"):
    FILES = []
    for root,_dir,f in os.walk(folder) :
        if f :
            FILES += [os.sep.join([root,name]) for name in f]
    #
    # @TODO: Add support for SQLite ....
    
    FILES = np.array_split(FILES,4)