fix test predictions depending on model output specification

This commit is contained in:
2017-08-03 07:51:58 +02:00
parent 8ac195ba6f
commit 787f43b328
3 changed files with 38 additions and 19 deletions

View File

@@ -1,17 +1,26 @@
#!/usr/bin/python2
import sys
import joblib
import numpy as np
import pandas as pd
df = joblib.load("/mnt/projekte/pmlcluster/cisco/trainData/multipleTaskLearning/currentData.joblib")
fn = sys.argv[1]
df = joblib.load("/mnt/projekte/pmlcluster/cisco/trainData/multipleTaskLearning/{}.joblib".format(fn))
df = pd.concat(df["data"])
df.reset_index(inplace=True)
df.dropna(axis=0, how="any", inplace=True)
df[["duration", "bytes_down", "bytes_up"]] = df[["duration", "bytes_down", "bytes_up"]].astype(np.int)
df[["domain", "server_ip"]] = df[["domain", "server_ip"]].astype(str)
df.serverLabel = pd.to_numeric(df.serverLabel, errors='coerce')
df.duration = pd.to_numeric(df.duration, errors='coerce')
df.bytes_down = pd.to_numeric(df.bytes_down, errors='coerce')
df.bytes_up = pd.to_numeric(df.bytes_up, errors='coerce')
df.http_method = df.http_method.astype("category")
df.serverLabel = df.serverLabel.astype(np.bool)
df.virusTotalHits = df.virusTotalHits.astype(np.int8)
df.trustedHits = df.trustedHits.astype(np.int8)
df.to_csv("/tmp/rk/full_future_dataset.csv.gz", compression="gzip")
df.to_csv("/tmp/rk/{}.csv".format(fn), encoding="utf-8")