def check():
changes = db.Prediction('changes').get()
if changes > cfg.numChanges:
if not db.Prediction.has_key('EnoughLabels'):
db.Prediction('EnoughLabels').put(False)
if db.Prediction('EnoughLabels').get():
return True
else:
if checkLabelRegions():
db.Prediction('EnoughLabels').put(True)
return True
return False
return False
def dropBadCols(df):
pd.set_option("display.max_rows", None, "display.max_columns", None)
noNegatives = df.replace(-np.Inf, np.nan)
output = noNegatives.dropna(axis=1)
# Take a note of what columns were dropped so that can be later used during prediction
# This line just compares the two column indices and finds the differences
badCols = list(set(df.columns) - set(output.columns))
db.Prediction('badCols').put(badCols)
return output
def doPrediction(data, problem, txn=None):
features = db.Features(data['user'], data['hub'], data['track'],
problem['chrom'], problem['chromStart']).get()
if features.empty:
return False
model = db.Prediction('model').get()
if not isinstance(model, dict):
return False
colsToDrop = db.Prediction('badCols').get()
featuresDropped = features.drop(labels=colsToDrop)
prediction = predictWithFeatures(featuresDropped, model)
if prediction is None:
return False
return prediction
def removeLabel(data):
toRemove = pd.Series({
'chrom': data['ref'],
'chromStart': data['start'],
'chromEnd': data['end']
})
txn = db.getTxn()
labels = db.Labels(data['user'], data['hub'], data['track'], data['ref'])
removed, after = labels.remove(toRemove, txn=txn)
db.Prediction('changes').increment(txn=txn)
Models.updateAllModelLabels(data, after)
txn.commit()
return removed.to_dict()
def updateLabel(data):
label = data['label']
updateLabel = pd.Series({
'chrom': data['ref'],
'chromStart': data['start'],
'chromEnd': data['end'],
'annotation': label
})
txn = db.getTxn()
labelDb = db.Labels(data['user'], data['hub'], data['track'], data['ref'])
item, labels = labelDb.add(updateLabel, txn=txn)
db.Prediction('changes').increment(txn=txn)
Models.updateAllModelLabels(data, labels)
txn.commit()
return item.to_dict()
def addLabel(data):
label = 'unknown'
for i in range(100):
print("#################################################")
print("THIS IS THE DATA", data)
for i in range(10):
print("#################################################")
# Duplicated because calls from updateLabel are causing freezing
newLabel = pd.Series({
'chrom': data['ref'],
'chromStart': data['start'],
'chromEnd': data['end'],
'annotation': label
})
txn = db.getTxn()
item, labels = db.Labels(data['user'], data['hub'], data['track'],
data['ref']).add(newLabel, txn=txn)
db.Prediction('changes').increment(txn=txn)
Models.updateAllModelLabels(data, labels)
txn.commit()
return data
def learn(X, Y):
cvfit = cvglmnet(x=X.to_numpy().copy(), y=Y.to_numpy().copy())
db.Prediction('model').put(cvfit)
def makePrediction(data):
model = db.Prediction('model').get()
print(model)