python︱大规模数据存储与读取、并行计算：Dask库简述

数据结构与pandas非常相似，比较容易理解。

• 原文文档：http://dask.pydata.org/en/latest/index.html

github：https://github.com/dask

dask的内容很多，挑一些我比较看好的内容着重点一下。
.

---

公众号“素质云笔记”定期更新博客内容：

---

一、数据读取与存储

先来看看dask能读入哪些内容：

1、csv

dask并不能读入excel，这个注意

# pandas
import pandas as pd                    
df = pd.read_csv('2015-01-01.csv')      
df.groupby(df.user_id).value.mean()     

#dask
 import dask.dataframe as dd
 df = dd.read_csv('2015-*-*.csv')
 df.groupby(df.user_id).value.mean().compute()

非常相似，除了.compute()
.

2、Dask Array读取hdf5

import numpy as np                       import dask.array as da
f = h5py.File('myfile.hdf5')             f = h5py.File('myfile.hdf5')
x = np.array(f['/small-data'])           x = da.from_array(f['/big-data'],
                                                           chunks=(1000, 1000))
x - x.mean(axis=1)                       x - x.mean(axis=1).compute()

左是Pandas，右边是dask
.

3、Dask Bag

import dask.bag as db
b = db.read_text('2015-*-*.json.gz').map(json.loads)
b.pluck('name').frequencies().topk(10, lambda pair: pair[1]).compute()

读取大规模json文件，几亿都很easy

>>> b = db.read_text('myfile.txt')
>>> b = db.read_text(['myfile.1.txt', 'myfile.2.txt', ...])
>>> b = db.read_text('myfile.*.txt')

读取txt

>>> import dask.bag as db
>>> b = db.from_sequence([{'name': 'Alice',   'balance': 100},
...                       {'name': 'Bob',     'balance': 200},
...                       {'name': 'Charlie', 'balance': 300}],
...                      npartitions=2)
>>> df = b.to_dataframe()

变为dataframe格式的内容
.

4、Dask Delayed 并行计算

from dask import delayed
L = []
for fn in filenames:                  # Use for loops to build up computation
    data = delayed(load)(fn)          # Delay execution of function
    L.append(delayed(process)(data))  # Build connections between variables

result = delayed(summarize)(L)
result.compute()

.

5、concurrent.futures自定义任务

from dask.distributed import Client
client = Client('scheduler:port')

futures = []
for fn in filenames:
    future = client.submit(load, fn)
    futures.append(future)

summary = client.submit(summarize, futures)
summary.result()

.

---

二、Delayed 并行计算模块

一个先行例子，本来的案例：

def inc(x):
    return x + 1

def double(x):
    return x + 2

def add(x, y):
    return x + y

data = [1, 2, 3, 4, 5]

output = []
for x in data:
    a = inc(x)
    b = double(x)
    c = add(a, b)
    output.append(c)

total = sum(output)

再来看看用delay加速的：

from dask import delayed

output = []
for x in data:
    a = delayed(inc)(x)
    b = delayed(double)(x)
    c = delayed(add)(a, b)
    output.append(c)

total = delayed(sum)(output)

还可以将计算流程可视化:

total.visualize()  # see image to the right

.

---

三、和SKLearn结合的并行算法

广义回归GLM：https://github.com/dask/dask-glm
tensorflow深度学习库：Dask-Tensorflow

以XGBoost为例，官方：https://github.com/dask/dask-xgboost
来看一个案例code
.

1、加载数据

import dask.dataframe as dd

# Subset of the columns to use
cols = ['Year', 'Month', 'DayOfWeek', 'Distance',
        'DepDelay', 'CRSDepTime', 'UniqueCarrier', 'Origin', 'Dest']

# Create the dataframe
df = dd.read_csv('s3://dask-data/airline-data/20*.csv', usecols=cols,
                  storage_options={'anon': True})

df = df.sample(frac=0.2) # we blow out ram otherwise

is_delayed = (df.DepDelay.fillna(16) > 15)

df['CRSDepTime'] = df['CRSDepTime'].clip(upper=2399)
del df['DepDelay']

df, is_delayed = persist(df, is_delayed)
progress(df, is_delayed)

2、One hot encode编码

df2 = dd.get_dummies(df.categorize()).persist()

.

3、准备训练集和测试集 + 训练

data_train, data_test = df2.random_split([0.9, 0.1], 
                                         random_state=1234)
labels_train, labels_test = is_delayed.random_split([0.9, 0.1], 
                                                    random_state=1234)

训练

import dask_xgboost as dxgb

params = {'objective': 'binary:logistic', 'nround': 1000, 
          'max_depth': 16, 'eta': 0.01, 'subsample': 0.5, 
          'min_child_weight': 1}

bst = dxgb.train(client, params, data_train, labels_train)
bst

.

4、预测

# Use normal XGBoost model with normal Pandas
import xgboost as xgb
dtest = xgb.DMatrix(data_test.head())
bst.predict(dtest)

predictions = dxgb.predict(client, bst, data_test).persist()
predictions.head()

.

5、模型评估

from sklearn.metrics import roc_auc_score, roc_curve
print(roc_auc_score(labels_test.compute(), 
                    predictions.compute()))
import matplotlib.pyplot as plt
%matplotlib inline

fpr, tpr, _ = roc_curve(labels_test.compute(), predictions.compute())
# Taken from http://scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html#sphx-glr-auto-examples-model-selection-plot-roc-py
plt.figure(figsize=(8, 8))
lw = 2
plt.plot(fpr, tpr, color='darkorange', lw=lw, label='ROC curve')
plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('Receiver operating characteristic example')
plt.legend(loc="lower right")
plt.show()

.

---

四、计算流程可视化部分——Dask.array

来源：https://gist.github.com/mrocklin/b61f795004ec0a70e43de350e453e97e

import numpy as np
import dask.array as da
x = da.ones(15, chunks=(5,))
x.visualize('dask.svg')

(x + 1).sum().visualize('dask.svg')

来一个二维模块的：

x = da.ones((15, 15), chunks=(5, 5))
x.visualize('dask.svg')
(x.dot(x.T + 1) - x.mean(axis=0)).std().visualize('dask.svg')