utils.py

#© 2020 By The Rector And Visitors Of The University Of Virginia

#Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
#The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

#THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
import os, io, stardog,json
import pandas as pd
from werkzeug.routing import PathConverter
import requests
import time
#session = requests.Session()
#session.auth = (SD_USERNAME,SD_PASSWORD)

SD_URL = os.environ.get('STARDOG_URL','http://stardog.uvadcos.io')
SD_USERNAME = os.environ.get('STARDOG_USERNAME')
SD_PASSWORD = os.environ.get('STARDOG_PASSWORD')

ORS_URL = os.environ.get("ORS_URL","ors.uvadco.io/")
EVI_PREFIX = 'evi:'
conn_details = {
        'endpoint': SD_URL,
        'username': SD_USERNAME,
        'password': SD_PASSWORD
    }
conn = stardog.Connection('ors', **conn_details)

class EverythingConverter(PathConverter):
    regex = '.*?'

def mint_eg_id(eg):
    '''
    Mints Id for newly created evidence graph
    '''

    r = requests.post(ORS_URL + "shoulder/ark:99999",data = json.dumps(eg))

    if 'created' in r.json():
        return r.json()['created']

def add_eg_to_og_id(ark,eg_id):

    r = requests.put(ORS_URL +  ark,
                data=json.dumps({EVI_PREFIX + 'hasEvidenceGraph':eg_id}))

def eg_exists(ark,token):
    '''
    Pings ors to see if ark is reconginzed
    '''

    r = requests.get(ORS_URL + ark,headers = {"Authorization": token})

    meta = r.json()



    if EVI_PREFIX + 'hasEvidenceGraph' in meta.keys():
        return True, meta[EVI_PREFIX + 'hasEvidenceGraph']

    elif 'error' in meta.keys():
        raise Exception

    return False,0

def existing_eg(eg_id):

    r = requests.get(ORS_URL + eg_id)
    eg = r.json()
    return eg


def query_stardog(ark,type = 'csv'):
    '''
    Performs the path query
    '''

    if type == 'json':
        results = conn.paths("PATHS ALL START ?x=<"+ ark + "> END ?y VIA ?p")
        return results
    else:
        results = conn.paths("PATHS ALL START ?x=<"+ ark + "> END ?y VIA ?p",content_type='text/csv')
    string_csv = io.StringIO(results.decode("utf-8"))

    df_eg = pd.read_csv(string_csv, sep=",")

    return df_eg

def is_id(string):
    if 'ark:' in string and len(string) == 46:
        return True
    if 'orchid:' in string:
        return True
    if 'https://orcid.org/' in string:
        return True

    return False

def parse_csv(df):
    '''
    Parses the ugly csv returned from stardog
    to build json evidence graph
    '''

    context = {'http://www.w3.org/1999/02/22-rdf-syntax-ns#':'@',
               'http://schema.org/_id':'@id',
              'http://schema.org/':'',
               'http://example.org/':'eg:',
               'http://example.org/':'evi:',
               "https://wf4ever.github.io/ro/2016-01-28/wfdesc/":'wfdesc:'
              }

    eg = {}
    current = eg
    for index, row in df.iterrows():

        #Nan row indicates next item restart at root node
        if pd.isna(row['x']):
            current = eg
            continue

        #cleans tags
        for key in context:
            if key in row['p']:
                row['p'] = row['p'].replace(key,context[key])
            if key in row['y']:
                row['y'] = row['y'].replace(key,context[key])

        #If key not already in eg add it
        #If its an id make it a dict not just an element
        if row['p'] not in current.keys():
            if is_id(row['y']):
                if row['p'] == '@id':
                    current['@id'] = row['y']
                else:
                    current[row['p']] = {'@id':row['y']}
                    current = current[row['p']]
            else:
                current[row['p']] = row['y']

        #If key already exists and its a new id its
        #not single dict so make it list of dicts
        elif isinstance(current[row['p']],dict):
            if current[row['p']]['@id'] == row['y']:
                current = current[row['p']]
            else:
                current[row['p']] = [current[row['p']],{'@id':row['y']}]
                current = current[row['p']][-1]

        #Look to see if we're moving into an element of the list
        #ie this is adding on to a known id
        #or if we're appending a new id to the list
        elif isinstance(current[row['p']],list):
            missing = True
            for dictionary in current[row['p']]:
                if dictionary['@id'] == row['y']:
                    current = dictionary
                    missing = False
                    break
            if missing:
                current[row['p']].append({'@id':row['y']})
                current = current[row['p']][-1]
    return eg

def build_evidence_graph(data,clean = True):
    '''
    Parses the ugly csv returned from stardog
    to build json evidence graph
    '''
    eg = {}

    context = {'http://www.w3.org/1999/02/22-rdf-syntax-ns#':'@',
          'http://schema.org/':'',
           'http://example.org/':'eg:',
           "https://wf4ever.github.io/ro/2016-01-28/wfdesc/":'wfdesc:'
          }
    trail = []

    for index, row in data.iterrows():
        if pd.isna(row['x']):
            trail = []
            continue
        if clean:
            for key in context:
                if key in row['p']:
                    row['p'] = row['p'].replace(key,context[key])
                print(row['y'])
                if key in row['y']:
                    row['y'] = row['y'].replace(key,context[key])

        if '@id' not in eg.keys():
            eg['@id'] = row['x']

        if trail == []:

            if row['p'] not in eg.keys():
                eg[row['p']] = row['y']

            else:
                trail.append(row['p'])
                if not isinstance(eg[row['p']],dict):
                    eg[row['p']] = {'@id':row['y']}

            continue

        current = eg
        for t in trail:
            current = current[t]

        if not isinstance(current,dict):
            continue

        if row['p'] not in current.keys():
            current[row['p']] = row['y']
        else:
            trail.append(row['p'])

            if not isinstance(current[row['p']],dict):
                current[row['p']] = {'@id':row['y']}

    return eg


def clean_eg(eg,eg_only = True,keep = []):
    '''
    Goes through json evidence graph removes keys unrelated to basic in or
    the evi ontology fixes minor formatting things
    '''

    for key in list(eg):

        if 'evi' not in key and 'eg' not in key and key != '@id' and \
                        key != 'author' and key != 'name' and key != '@type' \
                        and key not in keep:
            eg.pop(key, None)
            continue
        if 'evi:supports' == key:
            eg.pop(key, None)
            continue

        elif isinstance(eg[key],list):

            for dictionary in eg[key]:
                if isinstance(dictionary,dict):
                    dictionary = clean_eg(dictionary)

        elif isinstance(eg[key],dict):
            if len(eg[key]) == 1:
                #if dict only has one make it no longer a dict
                #just value instead
                eg[key] = eg[key][list(eg[key].keys())[0]]
            else:

                eg[key] = clean_eg(eg[key])

    return eg

def create_eg(ark,keep = []):
    '''
    Runs all functions required to create an evidence graph
    ie
        1.) Path query
        2.) Parse from csv -> json
        3.) Clean up json
    '''

    start = time.time()
    df_eg = query_stardog(ark)
    print('Querying stardog took: ' + str(time.time() - start))
    start = time.time()
    eg = parse_csv(df_eg)
    print('Parsing took: ' + str(time.time() - start))
    #eg = build_evidence_graph(df_eg)
    start = time.time()
    eg = clean_eg(eg,keep = keep)
    print('Cleaning took: ' + str(time.time() - start))

    eg["@context"] = {
    "@vocab": "http://schema.org/",
    "evi": "http://w3id.org/EVI#"
    }

    return eg

def parse_json(stardog_json):
    '''
    Parses the ugly csv returned from stardog
    to build json evidence graph
    '''

    context = {'http://www.w3.org/1999/02/22-rdf-syntax-ns#':'@',
               'http://schema.org/_id':'@id',
              'http://schema.org/':'',
               'http://example.org/':'eg:',
               'http://example.org/':'evi:',
               "https://wf4ever.github.io/ro/2016-01-28/wfdesc/":'wfdesc:'
              }

    eg = {}
    current = eg
    for bind in stardog_json['results']['bindings']:

        #Nan row indicates next item restart at root node
        if bind == {}:
            current = eg
            continue

        #cleans tags
        for key in context:
            if key in bind['p']['value']:
                bind['p']['value'] = bind['p']['value'].replace(key,context[key])
            if key in bind['y']['value']:
                bind['y']['value'] = bind['y']['value'].replace(key,context[key])

        #If key not already in eg add it
        #If its an id make it a dict not just an element
        if bind['p']['value'] not in current.keys():
            if is_id(bind['y']['value']):
                if bind['p']['value'] == '@id':
                    current['@id'] = bind['y']['value']
                else:
                    current[bind['p']['value']] = {'@id':bind['y']['value']}
                    current = current[bind['p']['value']]
            else:
                current[bind['p']['value']] = bind['y']['value']

        #If key already exists and its a new id its
        #not single dict so make it list of dicts
        elif isinstance(current[bind['p']['value']],dict):
            if current[bind['p']['value']]['@id'] == bind['y']['value']:
                current = current[bind['p']['value']]
            else:
                current[bind['p']['value']] = [current[bind['p']['value']],{'@id':bind['y']['value']}]
                current = current[bind['p']['value']][-1]

        #Look to see if we're moving into an element of the list
        #ie this is adding on to a known id
        #or if we're appending a new id to the list
        elif isinstance(current[bind['p']['value']],list):
            missing = True
            for dictionary in current[bind['p']['value']]:
                if dictionary['@id'] == bind['y']['value']:
                    current = dictionary
                    missing = False
                    break
            if missing:
                current[bind['p']['value']].append({'@id':bind['y']['value']})
                current = current[bind['p']['value']][-1]
    return eg

def create_eg_json(ark,keep = []):
    '''
    Runs all functions required to create an evidence graph
    ie
        1.) Path query
        2.) Parse from csv -> json
        3.) Clean up json
    '''

    start = time.time()
    stardog_json = query_stardog(ark,type = 'json')
    print('Querying stardog took: ' + str(time.time() - start))
    start = time.time()
    eg = parse_json(stardog_json)
    print('Parsing took: ' + str(time.time() - start))
    #eg = build_evidence_graph(df_eg)
    start = time.time()
    eg = clean_eg(eg,keep = keep)
    print('Cleaning took: ' + str(time.time() - start))

    eg["@context"] = {
    "@vocab": "http://schema.org/",
    "evi": "http://w3id.org/EVI#"
    }

    return eg