How to open dataframe stored in pickle in S3 on AWS? - python

I'm trying to load a dataframe file stored in pickle, and edit the dataframe. I'm opening in linux ubuntu AWS server, loading from S3. I could open by using pd.read_pickle in my environment, but that doesn't seem possibly when loading from S3
I can load and save from and to CSV, and I can also save to pickle, but I can't find a solution to load a pickle file. I tried copying almost every advice i could get on stackoverflow, but none seemed to work.
First, I tried
import pandas as pd
import boto3
import io
s3=boto3.client('s3')
session = boto3.session.Session()
obj = s3.get_object(Bucket = 'mybucket', Key = 'inputfile.pkl')
response = s3.get_object(Bucket=bucket, Key= inputfile)
body_string = response['Body'].read()
pickled = pickle.dumps(body_string)
df=pd.read_pickle(pickled) #embedded null type error
I also tried
import pickle
import boto3
s3 = boto3.resource('s3')
my_pickle = pickle.loads(s3.Bucket('bucket').Object('inputfile').get()['Body'].read())
df = pd.read_pickle(my_pickle) # with open(path, 'rb') as fh:
# return pkl.load(fh) error
Below is the method i used to save to pickle file
s3_resource = boto3.resource("s3")
s3=boto3.client('s3')
obj = s3.get_object(Bucket = bucket, Key = inputfile)
df = pd.read_csv(obj['Body'],index_col = False, low_memory = False)
pickle_buffer = io.BytesIO()
df.to_pickle(outputfile)
s3_resource.Object(bucket, outputfile).put(Body = open(outputfile, 'rb'))
I expect to get a df that i can work on, as I get from df = pd.read_pickle() or df = pd.read_csv()

Related

How to write pyarrow table as csv to s3 directly? [duplicate]

In boto 2, you can write to an S3 object using these methods:
Key.set_contents_from_string()
Key.set_contents_from_file()
Key.set_contents_from_filename()
Key.set_contents_from_stream()
Is there a boto 3 equivalent? What is the boto3 method for saving data to an object stored on S3?
In boto 3, the 'Key.set_contents_from_' methods were replaced by
Object.put()
Client.put_object()
For example:
import boto3
some_binary_data = b'Here we have some data'
more_binary_data = b'Here we have some more data'
# Method 1: Object.put()
s3 = boto3.resource('s3')
object = s3.Object('my_bucket_name', 'my/key/including/filename.txt')
object.put(Body=some_binary_data)
# Method 2: Client.put_object()
client = boto3.client('s3')
client.put_object(Body=more_binary_data, Bucket='my_bucket_name', Key='my/key/including/anotherfilename.txt')
Alternatively, the binary data can come from reading a file, as described in the official docs comparing boto 2 and boto 3:
Storing Data
Storing data from a file, stream, or string is easy:
# Boto 2.x
from boto.s3.key import Key
key = Key('hello.txt')
key.set_contents_from_file('/tmp/hello.txt')
# Boto 3
s3.Object('mybucket', 'hello.txt').put(Body=open('/tmp/hello.txt', 'rb'))
boto3 also has a method for uploading a file directly:
s3 = boto3.resource('s3')
s3.Bucket('bucketname').upload_file('/local/file/here.txt','folder/sub/path/to/s3key')
http://boto3.readthedocs.io/en/latest/reference/services/s3.html#S3.Bucket.upload_file
You no longer have to convert the contents to binary before writing to the file in S3. The following example creates a new text file (called newfile.txt) in an S3 bucket with string contents:
import boto3
s3 = boto3.resource(
's3',
region_name='us-east-1',
aws_access_key_id=KEY_ID,
aws_secret_access_key=ACCESS_KEY
)
content="String content to write to a new S3 file"
s3.Object('my-bucket-name', 'newfile.txt').put(Body=content)
Here's a nice trick to read JSON from s3:
import json, boto3
s3 = boto3.resource("s3").Bucket("bucket")
json.load_s3 = lambda f: json.load(s3.Object(key=f).get()["Body"])
json.dump_s3 = lambda obj, f: s3.Object(key=f).put(Body=json.dumps(obj))
Now you can use json.load_s3 and json.dump_s3 with the same API as load and dump
data = {"test":0}
json.dump_s3(data, "key") # saves json to s3://bucket/key
data = json.load_s3("key") # read json from s3://bucket/key
A cleaner and concise version which I use to upload files on the fly to a given S3 bucket and sub-folder-
import boto3
BUCKET_NAME = 'sample_bucket_name'
PREFIX = 'sub-folder/'
s3 = boto3.resource('s3')
# Creating an empty file called "_DONE" and putting it in the S3 bucket
s3.Object(BUCKET_NAME, PREFIX + '_DONE').put(Body="")
Note: You should ALWAYS put your AWS credentials (aws_access_key_id and aws_secret_access_key) in a separate file, for example- ~/.aws/credentials
After some research, I found this. It can be achieved using a simple csv writer. It is to write a dictionary to CSV directly to S3 bucket.
eg: data_dict = [{"Key1": "value1", "Key2": "value2"}, {"Key1": "value4", "Key2": "value3"}]
assuming that the keys in all the dictionary are uniform.
import csv
import boto3
# Sample input dictionary
data_dict = [{"Key1": "value1", "Key2": "value2"}, {"Key1": "value4", "Key2": "value3"}]
data_dict_keys = data_dict[0].keys()
# creating a file buffer
file_buff = StringIO()
# writing csv data to file buffer
writer = csv.DictWriter(file_buff, fieldnames=data_dict_keys)
writer.writeheader()
for data in data_dict:
writer.writerow(data)
# creating s3 client connection
client = boto3.client('s3')
# placing file to S3, file_buff.getvalue() is the CSV body for the file
client.put_object(Body=file_buff.getvalue(), Bucket='my_bucket_name', Key='my/key/including/anotherfilename.txt')
it is worth mentioning smart-open that uses boto3 as a back-end.
smart-open is a drop-in replacement for python's open that can open files from s3, as well as ftp, http and many other protocols.
for example
from smart_open import open
import json
with open("s3://your_bucket/your_key.json", 'r') as f:
data = json.load(f)
The aws credentials are loaded via boto3 credentials, usually a file in the ~/.aws/ dir or an environment variable.
You may use the below code to write, for example an image to S3 in 2019. To be able to connect to S3 you will have to install AWS CLI using command pip install awscli, then enter few credentials using command aws configure:
import urllib3
import uuid
from pathlib import Path
from io import BytesIO
from errors import custom_exceptions as cex
BUCKET_NAME = "xxx.yyy.zzz"
POSTERS_BASE_PATH = "assets/wallcontent"
CLOUDFRONT_BASE_URL = "https://xxx.cloudfront.net/"
class S3(object):
def __init__(self):
self.client = boto3.client('s3')
self.bucket_name = BUCKET_NAME
self.posters_base_path = POSTERS_BASE_PATH
def __download_image(self, url):
manager = urllib3.PoolManager()
try:
res = manager.request('GET', url)
except Exception:
print("Could not download the image from URL: ", url)
raise cex.ImageDownloadFailed
return BytesIO(res.data) # any file-like object that implements read()
def upload_image(self, url):
try:
image_file = self.__download_image(url)
except cex.ImageDownloadFailed:
raise cex.ImageUploadFailed
extension = Path(url).suffix
id = uuid.uuid1().hex + extension
final_path = self.posters_base_path + "/" + id
try:
self.client.upload_fileobj(image_file,
self.bucket_name,
final_path
)
except Exception:
print("Image Upload Error for URL: ", url)
raise cex.ImageUploadFailed
return CLOUDFRONT_BASE_URL + id

Python: Read CSV file without Pandas from S3 bucket [duplicate]

I have uploaded an excel file to AWS S3 bucket and now I want to read it in python. Any help would be appreciated. Here is what I have achieved so far,
import boto3
import os
aws_id = 'aws_id'
aws_secret = 'aws_secret_key'
client = boto3.client('s3', aws_access_key_id=aws_id, aws_secret_access_key=aws_secret)
bucket_name = 'my_bucket'
object_key = 'my_excel_file.xlsm'
object_file = client.get_object(Bucket=bucket_name, Key=object_key)
body = object_file['Body']
data = body.read()
What do I need to do next in order to read this data and work on it?
Spent quite some time on it and here's how I got it working,
import boto3
import io
import pandas as pd
import json
aws_id = ''
aws_secret = ''
bucket_name = ''
object_key = ''
s3 = boto3.client('s3', aws_access_key_id=aws_id, aws_secret_access_key=aws_secret)
obj = s3.get_object(Bucket=bucket_name, Key=object_key)
data = obj['Body'].read()
df = pd.read_excel(io.BytesIO(data), encoding='utf-8')
You can directly read xls file from S3 without having to download or save it locally. xlrd module has a provision to provide raw data to create workbook object.
Following is the code snippet.
from boto3 import Session
from xlrd.book import open_workbook_xls
aws_id = ''
aws_secret = ''
bucket_name = ''
object_key = ''
s3_session = Session(aws_access_key_id=aws_id, aws_secret_access_key=aws_secret)
bucket_object = s3_session.resource('s3').Bucket(bucket_name).Object(object_key)
content = bucket_object.get()['Body'].read()
workbook = open_workbook_xls(file_contents=content)
You can directly read excel files using awswrangler.s3.read_excel. Note that you can pass any pandas.read_excel() arguments (sheet name, etc) to this.
import awswrangler as wr
df = wr.s3.read_excel(path=s3_uri)
Python doesn't support excel files natively. You could use the pandas library pandas library read_excel functionality

upload a dataframe as a zipped csv directly to s3 without saving it on the local machine

How can I upload a data frame as a zipped csv into S3 bucket without saving it on my local machine first?
I have the connection to that bucket already running using:
self.s3_output = S3(bucket_name='test-bucket', bucket_subfolder='')
We can make a file-like object with BytesIO and zipfile from the standard library.
# 3.7
from io import BytesIO
import zipfile
# .to_csv returns a string when called with no args
s = df.to_csv()
with zipfile.ZipFile(BytesIO(), mode="w",) as z:
z.writestr("df.csv", s)
# upload file here
You'll want to refer to upload_fileobj in order to customize how the upload behaves.
yourclass.s3_output.upload_fileobj(z, ...)
This works equally well for zip and gz:
import boto3
import gzip
import pandas as pd
from io import BytesIO, TextIOWrapper
s3_client = boto3.client(
service_name = "s3",
endpoint_url = your_endpoint_url,
aws_access_key_id = your_access_key,
aws_secret_access_key = your_secret_key
# Your file name inside zip
your_filename = "test.csv"
s3_path = f"path/to/your/s3/compressed/file/test.zip"
bucket = "your_bucket"
df = your_df
gz_buffer = BytesIO()
with gzip.GzipFile(
filename = your_filename,
mode = 'w',
fileobj = gz_buffer ) as gz_file:
df.to_csv(TextIOWrapper(gz_file, 'utf8'), index=False)
s3.put_object(
Bucket=bucket, Key=s3_path, Body=gz_buffer.getvalue()
)

How to load a pickle file from S3 to use in AWS Lambda?

I am currently trying to load a pickled file from S3 into AWS lambda and store it to a list (the pickle is a list).
Here is my code:
import pickle
import boto3
s3 = boto3.resource('s3')
with open('oldscreenurls.pkl', 'rb') as data:
old_list = s3.Bucket("pythonpickles").download_fileobj("oldscreenurls.pkl", data)
I get the following error even though the file exists:
FileNotFoundError: [Errno 2] No such file or directory: 'oldscreenurls.pkl'
Any ideas?
Super simple solution
import pickle
import boto3
s3 = boto3.resource('s3')
my_pickle = pickle.loads(s3.Bucket("bucket_name").Object("key_to_pickle.pickle").get()['Body'].read())
As shown in the documentation for download_fileobj, you need to open the file in binary write mode and save to the file first. Once the file is downloaded, you can open it for reading and unpickle.
import pickle
import boto3
s3 = boto3.resource('s3')
with open('oldscreenurls.pkl', 'wb') as data:
s3.Bucket("pythonpickles").download_fileobj("oldscreenurls.pkl", data)
with open('oldscreenurls.pkl', 'rb') as data:
old_list = pickle.load(data)
download_fileobj takes the name of an object in S3 plus a handle to a local file, and saves the contents of that object to the file. There is also a version of this function called download_file that takes a filename instead of an open file handle and handles opening it for you.
In this case it would probably be better to use S3Client.get_object though, to avoid having to write and then immediately read a file. You could also write to an in-memory BytesIO object, which acts like a file but doesn't actually touch a disk. That would look something like this:
import pickle
import boto3
from io import BytesIO
s3 = boto3.resource('s3')
with BytesIO() as data:
s3.Bucket("pythonpickles").download_fileobj("oldscreenurls.pkl", data)
data.seek(0) # move back to the beginning after writing
old_list = pickle.load(data)
This is the easiest solution. You can load the data without even downloading the file locally using S3FileSystem
from s3fs.core import S3FileSystem
s3_file = S3FileSystem()
data = pickle.load(s3_file.open('{}/{}'.format(bucket_name, file_path)))
According to my implementation, S3 file path read with pickle.
import pickle
import boto3
name = img_url.split('/')[::-1][0]
folder = 'media'
file_name = f'{folder}/{name}'
bucket_name = bucket_name
s3 = boto3.client('s3', aws_access_key_id=aws_access_key_id,aws_secret_access_key=aws_secret_access_key)
response = s3.get_object(Bucket=bucket_name, Key=file_name)
body = response['Body'].read()
data = pickle.loads(body)

Save Dataframe to csv directly to s3 Python

I have a pandas DataFrame that I want to upload to a new CSV file. The problem is that I don't want to save the file locally before transferring it to s3. Is there any method like to_csv for writing the dataframe to s3 directly? I am using boto3.
Here is what I have so far:
import boto3
s3 = boto3.client('s3', aws_access_key_id='key', aws_secret_access_key='secret_key')
read_file = s3.get_object(Bucket, Key)
df = pd.read_csv(read_file['Body'])
# Make alterations to DataFrame
# Then export DataFrame to CSV through direct transfer to s3
You can use:
from io import StringIO # python3; python2: BytesIO
import boto3
bucket = 'my_bucket_name' # already created on S3
csv_buffer = StringIO()
df.to_csv(csv_buffer)
s3_resource = boto3.resource('s3')
s3_resource.Object(bucket, 'df.csv').put(Body=csv_buffer.getvalue())
You can directly use the S3 path. I am using Pandas 0.24.1
In [1]: import pandas as pd
In [2]: df = pd.DataFrame( [ [1, 1, 1], [2, 2, 2] ], columns=['a', 'b', 'c'])
In [3]: df
Out[3]:
a b c
0 1 1 1
1 2 2 2
In [4]: df.to_csv('s3://experimental/playground/temp_csv/dummy.csv', index=False)
In [5]: pd.__version__
Out[5]: '0.24.1'
In [6]: new_df = pd.read_csv('s3://experimental/playground/temp_csv/dummy.csv')
In [7]: new_df
Out[7]:
a b c
0 1 1 1
1 2 2 2
Release Note:
S3 File Handling
pandas now uses s3fs for handling S3 connections. This shouldn’t break any code. However, since s3fs is not a required dependency, you will need to install it separately, like boto in prior versions of pandas. GH11915.
I like s3fs which lets you use s3 (almost) like a local filesystem.
You can do this:
import s3fs
bytes_to_write = df.to_csv(None).encode()
fs = s3fs.S3FileSystem(key=key, secret=secret)
with fs.open('s3://bucket/path/to/file.csv', 'wb') as f:
f.write(bytes_to_write)
s3fs supports only rb and wb modes of opening the file, that's why I did this bytes_to_write stuff.
This is a more up to date answer:
import s3fs
s3 = s3fs.S3FileSystem(anon=False)
# Use 'w' for py3, 'wb' for py2
with s3.open('<bucket-name>/<filename>.csv','w') as f:
df.to_csv(f)
The problem with StringIO is that it will eat away at your memory. With this method, you are streaming the file to s3, rather than converting it to string, then writing it into s3. Holding the pandas dataframe and its string copy in memory seems very inefficient.
If you are working in an ec2 instant, you can give it an IAM role to enable writing it to s3, thus you dont need to pass in credentials directly. However, you can also connect to a bucket by passing credentials to the S3FileSystem() function. See documention:https://s3fs.readthedocs.io/en/latest/
If you pass None as the first argument to to_csv() the data will be returned as a string. From there it's an easy step to upload that to S3 in one go.
It should also be possible to pass a StringIO object to to_csv(), but using a string will be easier.
You can also use the AWS Data Wrangler:
import awswrangler as wr
wr.s3.to_csv(
df=df,
path="s3://...",
)
Note that it will handle multipart upload for you to make the upload faster.
I found this can be done using client also and not just resource.
from io import StringIO
import boto3
s3 = boto3.client("s3",\
region_name=region_name,\
aws_access_key_id=aws_access_key_id,\
aws_secret_access_key=aws_secret_access_key)
csv_buf = StringIO()
df.to_csv(csv_buf, header=True, index=False)
csv_buf.seek(0)
s3.put_object(Bucket=bucket, Body=csv_buf.getvalue(), Key='path/test.csv')
I use AWS Data Wrangler. For example:
import awswrangler as wr
import pandas as pd
# read a local dataframe
df = pd.read_parquet('my_local_file.gz')
# upload to S3 bucket
wr.s3.to_parquet(df=df, path='s3://mys3bucket/file_name.gz')
The same applies to csv files. Instead of read_parquet and to_parquet, use read_csv and to_csv with the proper file extension.
since you are using boto3.client(), try:
import boto3
from io import StringIO #python3
s3 = boto3.client('s3', aws_access_key_id='key', aws_secret_access_key='secret_key')
def copy_to_s3(client, df, bucket, filepath):
csv_buf = StringIO()
df.to_csv(csv_buf, header=True, index=False)
csv_buf.seek(0)
client.put_object(Bucket=bucket, Body=csv_buf.getvalue(), Key=filepath)
print(f'Copy {df.shape[0]} rows to S3 Bucket {bucket} at {filepath}, Done!')
copy_to_s3(client=s3, df=df_to_upload, bucket='abc', filepath='def/test.csv')
You can use
pandas
boto3
s3fs (version ≤0.4)
I use to_csv with s3:// in path and storage_options
key = "folder/file.csv"
df.to_csv(
f"s3://{YOUR_S3_BUCKET}/{key}",
index=False,
storage_options={
"key": AWS_ACCESS_KEY_ID,
"secret": AWS_SECRET_ACCESS_KEY,
"token": AWS_SESSION_TOKEN,
},
To handle large files efficiently you can also use an open-source S3-compatible MinIO, with its minio python client package, like in this function of mine:
import minio
import os
import pandas as pd
minio_client = minio.Minio(..)
def write_df_to_minio(df,
minio_client,
bucket_name,
file_name="new-file.csv",
local_temp_folder="/tmp/",
content_type="application/csv",
sep=",",
save_row_index=False):
df.to_csv(os.path.join(local_temp_folder, file_name), sep=sep, index=save_row_index)
minio_results = minio_client.fput_object(bucket_name=bucket_name,
object_name=file_name,
file_path=os.path.join(local_temp_folder, file_name),
content_type=content_type)
assert minio_results.object_name == file_name
Another option is to do this with cloudpathlib, which supports S3 and also Google Cloud Storage and Azure Blob Storage. See example below.
import pandas as pd
from cloudpathlib import CloudPath
# read data from S3
df = pd.read_csv(CloudPath("s3://covid19-lake/rearc-covid-19-testing-data/csv/states_daily/states_daily.csv"))
# look at some of the data
df.head(1).T.iloc[:10]
#> 0
#> date 20210307
#> state AK
#> positive 56886.0
#> probableCases NaN
#> negative NaN
#> pending NaN
#> totalTestResultsSource totalTestsViral
#> totalTestResults 1731628.0
#> hospitalizedCurrently 33.0
#> hospitalizedCumulative 1293.0
# writing to S3
with CloudPath("s3://bucket-you-can-write-to/data.csv").open("w") as f:
df.to_csv(f)
CloudPath("s3://bucket-you-can-write-to/data.csv").exists()
#> True
Note, that you can't call df.to_csv(CloudPath("s3://drivendata-public-assets/test-asdf2.csv")) directly because of the way pandas handles paths/handles passed to it. Instead you need to open the file for writing and pass that handle directly to to_csv.
This comes with a few added benefits in terms of setting particular options or different authentication mechanisms or keeping a persistent cache so you don't always need to redownload from S3.
from io import StringIO
import boto3
#Creating Session With Boto3.
session = boto3.Session(
aws_access_key_id='<your_access_key_id>',
aws_secret_access_key='<your_secret_access_key>'
)
#Creating S3 Resource From the Session.
s3_res = session.resource('s3')
csv_buffer = StringIO()
df.to_csv(csv_buffer)
bucket_name = 'stackvidhya'
s3_object_name = 'df.csv'
s3_res.Object(bucket_name, s3_object_name).put(Body=csv_buffer.getvalue())
print("Dataframe is saved as CSV in S3 bucket.")
For those who might have problems with S3FS or fsspec using Lambda:
You have to create a layer for each libary and insert them in your Lambda.
You can find how to crate a layer here.
I read a csv with two columns from bucket s3, and the content of the file csv i put in pandas dataframe.
Example:
config.json
{
"credential": {
"access_key":"xxxxxx",
"secret_key":"xxxxxx"
}
,
"s3":{
"bucket":"mybucket",
"key":"csv/user.csv"
}
}
cls_config.json
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import os
import json
class cls_config(object):
def __init__(self,filename):
self.filename = filename
def getConfig(self):
fileName = os.path.join(os.path.dirname(__file__), self.filename)
with open(fileName) as f:
config = json.load(f)
return config
cls_pandas.py
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import pandas as pd
import io
class cls_pandas(object):
def __init__(self):
pass
def read(self,stream):
df = pd.read_csv(io.StringIO(stream), sep = ",")
return df
cls_s3.py
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import boto3
import json
class cls_s3(object):
def __init__(self,access_key,secret_key):
self.s3 = boto3.client('s3', aws_access_key_id=access_key, aws_secret_access_key=secret_key)
def getObject(self,bucket,key):
read_file = self.s3.get_object(Bucket=bucket, Key=key)
body = read_file['Body'].read().decode('utf-8')
return body
test.py
#!/usr/bin/env python
# -*- coding: utf-8 -*-
from cls_config import *
from cls_s3 import *
from cls_pandas import *
class test(object):
def __init__(self):
self.conf = cls_config('config.json')
def process(self):
conf = self.conf.getConfig()
bucket = conf['s3']['bucket']
key = conf['s3']['key']
access_key = conf['credential']['access_key']
secret_key = conf['credential']['secret_key']
s3 = cls_s3(access_key,secret_key)
ob = s3.getObject(bucket,key)
pa = cls_pandas()
df = pa.read(ob)
print df
if __name__ == '__main__':
test = test()
test.process()

Categories

Resources