可以将文章内容翻译成中文,广告屏蔽插件可能会导致该功能失效(如失效，请关闭广告屏蔽插件后再试):

问题:

Goal - To read csv file uploaded on google cloud storage bucket.

Environment - Run Jupyter notebook using SSH instance on Master node. Using python on Jupyter notebook trying to access a simple csv file uploaded onto google cloud storage bucket.

Approaches -

1st approach - Write a simple python program

Wrote following program

import csv
f = open('gs://python_test_hm/train.csv' , 'rb' ) 
csv_f = csv.reader(f)
for row in csv_f
     print row

Results - Error message "No such file or directory"

2nd Approach - Using gcloud Package tried to access the train.csv file. The sample code is shown below. Below code is not the actual code. The file on google Cloud storage in my version of code was referred to "gs:///Filename.csv" Results - Error message "No such file or directory"

Load data from CSV

import csv
from gcloud import bigquery
from gcloud.bigquery import SchemaField
client = bigquery.Client()
dataset = client.dataset('dataset_name')
dataset.create()  # API request

SCHEMA = [
    SchemaField('full_name', 'STRING', mode='required'),
    SchemaField('age', 'INTEGER', mode='required'),
 ]
table = dataset.table('table_name', SCHEMA)
table.create()

with open('csv_file', 'rb') as readable:
    table.upload_from_file(
        readable, source_format='CSV', skip_leading_rows=1)

3rd Approach -

import csv
import urllib

url = 'https://storage.cloud.google.com/<bucket>/train.csv'


response = urllib.urlopen(url)
cr = csv.reader(response)
print cr

for row in cr:
    print row

Results - Above code doesn't result in any error but it displays the XML content of the google page as shown below. I am interested in viewing the data of the train csv file.

['<!DOCTYPE html>']
['<html lang="en">']
['  <head>']
['  <meta charset="utf-8">']
['  <meta content="width=300', ' initial-scale=1" name="viewport">']
['  <meta name="google-site-verification" content="LrdTUW9psUAMbh4Ia074-   BPEVmcpBxF6Gwf0MSgQXZs">']
['  <title>Sign in - Google Accounts</title>']

Can someone throw some light on what could be possibly wrong here and how do I achieve my goal? Your help is highly appreciated.

Thanks so much for your help!

回答1:

I assume you are using Jupyter notebook running on a machine in Google Cloud Platform (GCP)? If that's the case, you will already have the Google Cloud SDK running on that machine (by default).

With this setup you have 2 easy options to work with Google Cloud Storage (GCS):

Use the gcloud/gsutil commands in Jupyter

Writing to GCS: gsutil cp train.csv gs://python_test_hm/train.csv

Reading from GCS: gsutil cp gs://python_test_hm/train.csv train.csv
Use google-cloud python library

Writing to GCS:

from google.cloud import storage
client = storage.Client()
bucket = client.get_bucket('python_test_hm')
blob = bucket.blob('train.csv')
blob.upload_from_string('this is test content!')

Reading from GCS:

from google.cloud import storage
client = storage.Client()
bucket = client.get_bucket('python_test_hm')
blob = storage.Blob('train.csv', bucket)
content = blob.download_as_string()

回答2:

The sign in page your app fetches isn't actually the object - it's an auth redirect page that, if interacted-with, would proceed to serve the object. You should check out the documentation on Cloud Storage to see about how auth works, and look up the auth details for whichever library or means you use to access the bucket / object.