def insert(file, table):
count=0
cursor=Config.connect()
row={}
for line in file:
tmp_array=line.split("\t")
row['UniprotAC']=tmp_array[0]
row['GI']=tmp_array[4]
row['RefSeq']=tmp_array[3]
row['EntrezGene']=tmp_array[2]
row['GoId']=tmp_array[7]
if tmp_array[7]=="":
pass
GI_data=tmp_array[4]
refseq_data=tmp_array[3]
if len(GI_data.split(";"))>1:
tmp_GI=GI_data.split(";")
for AC_GI in tmp_GI:
row['GI']=AC_GI.replace(" ","")
query=QueryGenerator.gen_insert(table, row)
cursor.execute(query)
count=count+1
if len(refseq_data.split(";"))>1:
tmp_refseq=refseq_data.split(";")
for AC_RefSeq in tmp_refseq:
row['RefSeq']=AC_RefSeq.replace(" ","")
query=QueryGenerator.gen_insert(table, row)
cursor.execute(query)
count=count+1
else:
query=QueryGenerator.gen_insert(table, row)
cursor.execute(query)
count=count+1
return count
def update_results():
# get user choices if they exist (batch required)
bSelected = request.args.get('bSelected', type=str)
if len(bSelected) == 0:
# TODO: Figure out a better way to catch this error, probably
# through html solution that requires user to choose batch before
# using other selects
return jsonify(resultstable="MUST SELECT A BATCH")
cSelected = request.args.getlist('cSelected[]')
mSelected = request.args.getlist('mSelected[]')
#query narf results by batch, store in dataframe
results = qg.QueryGenerator(dbc,tablename='NarfResults',\
batchnum=bSelected).send_query()
# TODO: Figure out why models parsed from analysis selection aren't
# matching up with cells filtered by batch selection.
#filter dataframe by selected cells, if any
if not (len(cSelected) == 0):
results = results[results.cellid.isin(cSelected)]
#same for selected models, if any
if not (len(mSelected) == 0):
results = results[results.modelname.isin(mSelected)]
return jsonify(resultstable=results.to_html(classes='table-hover\
table-condensed'))
def main_view():
tablelist = 'NarfResults'
plottypelist = 'Scatter'
measurelist = 'r_test'
resultstable = ''
# TODO: Figure out how to get initial lists to load faster - added distinct
# and field selections to shrink query, but initial page load is still very slow
# if selection limit is lifted (so far tested up to 200k, takes ~20 seconds to load)
# possibility: organize as analysis, batch, model objects and load a global-scope
# dataframe with the relevant data on app launch? then just have view functions index
# into dataframe as needed instead of generating new queries each time
analyses = qg.QueryGenerator(dbc, column='name',
tablename='NarfAnalysis').send_query()
analyses = analyses.iloc[:, 0] #convert from 1-dim df to series
analysislist = analyses.tolist()
batches = qg.QueryGenerator(dbc,distinct=True,column='batch',tablename=\
'NarfBatches').send_query()
batches = batches.iloc[:, 0]
batchlist = batches.tolist()
# Leave this out for now to improve page load time
"""
## Maybe don't need this one with new setup? Cells can just populate after
## analysis is selected. Unless want to be able to sort by a specific cell first.
cells = qg.QueryGenerator(dbc,distinct=True,column='cellid',tablename='NarfBatches').send_query()
cells = cells.iloc[:,0]
# get unique list since there are duplicates
celllist = cells.tolist()
models = qg.QueryGenerator(dbc,distinct=True,column='modelname',tablename='NarfResults').send_query()
models = models.iloc[:,0]
modellist = models.tolist()
"""
return render_template('main.html', analysislist = analysislist,\
tablelist = tablelist,\
batchlist = batchlist,\
#celllist = celllist,\
#modellist = modellist,\
plottypelist = plottypelist,
measurelist = measurelist,
)
def print_query_generator(queries_file):
if opts.verbose:
gen = QueryGenerator.parse(queries_file)
for op_name, values in gen.ops.iteritems():
print op_name, values
print '\'', gen.base, '\''
print gen.idents
print ''
def view_database():
tablelist = 'NarfResults'
batches = qg.QueryGenerator(dbc, column='batch',
tablename='NarfBatches').send_query()
batches = batches.iloc[:, 0]
# get unique list since there are duplicates
batchlist = list(set(batches.tolist()))
models = qg.QueryGenerator(dbc,
column='modelname',
tablename='NarfResults').send_query()
models = models.iloc[:, 0]
modellist = list(set(models.tolist()))
return render_template('database.html', tablelist=tablelist, batchlist=batchlist,\
modellist=modellist)
def print_query_generator(queries_file): if opts.verbose: gen = QueryGenerator.parse(queries_file) for op_name, values in gen.ops.iteritems(): print op_name, values print '\'', gen.base, '\'' print gen.idents print ''
def update_cells():
bSelected = request.args.get('bSelected', 'no selection', type=str)
celllist = qg.QueryGenerator(dbc,column='cellid',tablename='NarfBatches',\
batchnum=bSelected).send_query()
celllist = celllist.iloc[:, 0].tolist()
return jsonify(celllist=celllist)
def update_batch():
aSelected = request.args.get('aSelected', 'no selection', type=str)
batch = qg.QueryGenerator(dbc,tablename='NarfAnalysis',\
analysis=aSelected).send_query()
# get string of first 3 characters, rest is description of batch
batchnum = batch['batch'].iloc[0][:3]
#return batchnum for selected analysis in jQuery-friendly format
return jsonify(batchnum=batchnum)
def req_query(tablename, batchnum, modelname):
# parse variables from URL and pass to QueryGenerator object as attributes
query = qg.QueryGenerator(tablename, batchnum, modelname)
# populate dataframe by calling send_query() on qg object
data = query.send_query()
# TODO: Figure out best practice for when and where to close database connection.
# For now, will continue putting it in any time the conneciton is no longer needed
# for current operation.
query.close_connection()
# generate descriptive table title from variables
tabletitle = ("%s, filter by: batch=%s, model=%s" %
(tablename, batchnum, modelname))
# generage html page via table.html template, pass in html export of dataframe
# and table title as variables
return render_template('table.html', table=data.to_html(classes='Table'),\
title=tabletitle)
def req_query():
# add if statements to check if request data exists before pulling
tablename = request.form['tablename']
batchnum = request.form['batchnum']
modelname = request.form['modelname']
query = qg.QueryGenerator(dbc,tablename=tablename,batchnum=batchnum,\
modelname=modelname)
# populate dataframe by calling send_query() on qg object
data = query.send_query()
tabletitle = ("%s, filter by: batch=%s, model=%s" %
(tablename, batchnum, modelname))
# generage html page via table.html template, pass in html export of dataframe
# and table title as variables
return render_template('table.html', table=data.to_html(classes='table'),\
title=tabletitle)
def update_models():
aSelected = request.args.get('aSelected', 'no selection', type=str)
#currently disabled until modelfinder methods are fixed - combo array
#recursion crashing website
analysis = qg.QueryGenerator(dbc,column='modeltree',tablename='NarfAnalysis',\
analysis=aSelected).send_query()
# pull modeltree text from NarfAnalysis
# and convert to string rep
modeltree = analysis.iloc[0, 0]
modelFinder = mf.ModelFinder(modeltree)
modellist = modelFinder.modellist
#modellist = ['testing','jquery','code','for','analysis','update',aSelected]
return jsonify(modellist=modellist)
def loadData(balanceDatasets):
# Read from files
legitStrings, maliciousStrings = qg.load_csvFiles(balanceDatasets)
# Create SQL Injection queries
legitSet, maliciousSet = qg.generate_dataset(legitStrings, maliciousStrings)
return legitSet, maliciousSet
def check_database(queries_file,
exec_sql,
exec_aql,
stop_on_failure=False,
verbose=False):
queries_log = ''
if queries_file is None:
raise Exception('you need to provide a generator queries file')
sum_rows = 0
nb_checked, nb_error = 0, 0
sql_sum_time, aql_sum_time = 0, 0
for aql_query in QueryGenerator.iterate(queries_file):
ss = AQLParser.Statements()
ss.parse(aql_query)
sql_query = ss.to_sql(separator='\n')
# aql_query = AQLParser.kjeq_parse(aql_query) # FIXME
if verbose:
print ''
print '==='
print ''
print aql_query
print ''
print sql_query
##################################################
# skip full outer because not supported by mysql #
if exec_sql.sgbd == 'MySQL' and sql_query.lower().find(
'full outer') != -1:
if verbose:
print "full outer not supported by MySQL, skip query"
continue
##################################################
nb_checked += 1
rc, sql_time, sql_rows = exec_sql.execute_and_fetch(
sql_query.replace(';', ''))
rc, aql_time, aql_rows = exec_aql.execute(aql_query)
if rc == 0:
if ('ORDER' in aql_query) and (not util.row_in(
sql_rows, aql_rows, True)):
rc = 1
elif (not util.row_in(sql_rows, aql_rows)) or (not util.row_in(
aql_rows, sql_rows)):
rc = 1
if rc != 0:
nb_error += 1
if verbose:
print_query(sys.stderr, 'ERROR: query failed', aql_query,
sql_query, aql_rows, sql_rows)
elif rc == 0:
sys.stdout.write('d')
elif rc != 0:
sys.stdout.write('e')
queries_log += '<Query id="' + str(
nb_checked) + '" status="error" type="' + (
'data' if rc == 0 else 'aq_engine') + '">\n'
queries_log += '<AQL>\n'
queries_log += aql_query + '\n'
queries_log += '</AQL>\n'
queries_log += '<SQL>\n'
queries_log += sql_query + '\n'
queries_log += '</SQL>\n'
queries_log += '<Results nb="' + str(len(aql_rows)) + '">\n'
for row in aql_rows:
queries_log += '<row>' + ','.join(str(v)
for v in row) + '</row>\n'
queries_log += '\n'
queries_log += '</Results>\n'
queries_log += '<Expected nb="' + str(len(sql_rows)) + '">\n'
for row in sql_rows:
queries_log += '<row>' + str(row) + '</row>\n'
queries_log += '\n'
queries_log += '</Expected>\n'
queries_log += '</Query>\n'
if stop_on_failure:
return (nb_checked, nb_error)
else:
sum_rows += len(sql_rows)
sql_sum_time += sql_time
aql_sum_time += aql_time
if verbose:
print_query(sys.stdout, 'query successfully checked',
aql_query, sql_query, aql_rows, sql_rows)
else:
sys.stdout.write('.')
# queries_log += '<query status="successful">\n'
# queries_log += aql_query
# queries_log += '</query>\n'
return (nb_checked, nb_error, queries_log, sql_sum_time, aql_sum_time,
sum_rows)
__author__ = 'Ivan Dortulov'
from QueryGenerator import *
gen = QueryGenerator('localhost', 'WorkDatabase', 'postgres', 'postgres')
graph = gen.build_table_graph()
print(gen.generate_query(graph, 'blue'))
def check_database(queries_file, exec_sql, exec_aql, stop_on_failure=False, verbose=False):
queries_log = ''
if queries_file is None:
raise Exception('you need to provide a generator queries file')
sum_rows = 0
nb_checked, nb_error = 0, 0
sql_sum_time, aql_sum_time = 0, 0
for aql_query in QueryGenerator.iterate(queries_file):
ss = AQLParser.Statements()
ss.parse(aql_query)
sql_query = ss.to_sql(separator = '\n')
# aql_query = AQLParser.kjeq_parse(aql_query) # FIXME
if verbose:
print ''
print '==='
print ''
print aql_query
print ''
print sql_query
##################################################
# skip full outer because not supported by mysql #
if exec_sql.sgbd == 'MySQL' and sql_query.lower().find('full outer') != -1:
if verbose:
print "full outer not supported by MySQL, skip query"
continue
##################################################
nb_checked += 1
rc, sql_time, sql_rows = exec_sql.execute_and_fetch(sql_query.replace(';', ''))
rc, aql_time, aql_rows = exec_aql.execute(aql_query)
if rc == 0:
if ('ORDER' in aql_query) and (not util.row_in(sql_rows, aql_rows, True)):
rc = 1
elif (not util.row_in(sql_rows, aql_rows)) or (not util.row_in(aql_rows, sql_rows)):
rc = 1
if rc != 0:
nb_error += 1
if verbose:
print_query(sys.stderr, 'ERROR: query failed', aql_query, sql_query, aql_rows, sql_rows)
elif rc == 0:
sys.stdout.write('d')
elif rc != 0:
sys.stdout.write('e')
queries_log += '<Query id="' + str(nb_checked) + '" status="error" type="' + ('data' if rc == 0 else 'aq_engine') + '">\n'
queries_log += '<AQL>\n'
queries_log += aql_query + '\n'
queries_log += '</AQL>\n'
queries_log += '<SQL>\n'
queries_log += sql_query + '\n'
queries_log += '</SQL>\n'
queries_log += '<Results nb="' + str(len(aql_rows)) + '">\n'
for row in aql_rows:
queries_log += '<row>' + ','.join(str(v) for v in row) + '</row>\n'
queries_log += '\n'
queries_log += '</Results>\n'
queries_log += '<Expected nb="' + str(len(sql_rows)) + '">\n'
for row in sql_rows:
queries_log += '<row>' + str(row) + '</row>\n'
queries_log += '\n'
queries_log += '</Expected>\n'
queries_log += '</Query>\n'
if stop_on_failure:
return (nb_checked, nb_error)
else:
sum_rows += len(sql_rows)
sql_sum_time += sql_time
aql_sum_time += aql_time
if verbose:
print_query(sys.stdout, 'query successfully checked', aql_query, sql_query, aql_rows, sql_rows)
else:
sys.stdout.write('.')
# queries_log += '<query status="successful">\n'
# queries_log += aql_query
# queries_log += '</query>\n'
return (nb_checked, nb_error, queries_log, sql_sum_time, aql_sum_time, sum_rows)
- Used to parse modelstring from NarfAnalysis
- into a list of model names that can be passed back
- to model selector
"""
import ast
import re
#for testing w/ actual model string
import QueryGenerator as qg
import DB_Connection as dbcon
db = dbcon.DB_Connection()
dbc = db.connection
analysis = qg.QueryGenerator(dbc,
tablename='NarfAnalysis',
analysis='Jake Test').send_query()
modstring = analysis['modeltree'][0]
"""
# for testing with simpler nested list that (hopefully) won't crash the universe
modstring = "{'a','b', {{'c','d'},{ 'e','f' }}, 'g'}"
# list of combos should end up as:
# ['abceg','abcfg','abdeg','abdfg']
"""
class ModelFinder():
def __init__(self, modelstring=''):
self.modelstring = modelstring
# as soon as modelstring is passed, go ahead and
# parse into array then to list so that attribs can be retrieved