1

在 python 3.6.2 上使用 ibm-cos-sdk 时，我得到一个 Unexpected keyword argument 'ibm_api_key_id' 错误参数。

我已经使用以下步骤在干净的虚拟环境中安装了该库：

virtualenv --python=python3.5 boto-test
source boto-test/bin/activate
pip install ibm-cos-sdk

然后，我尝试从这里运行示例：

import boto3
import json
import requests
import random
from botocore.client import Config
from pprint import pprint

with open('./credentials.json') as data_file:
    credentials = json.load(data_file)

print("Service credential:")
print(json.dumps(credentials, indent=2))
print("")
print("Connecting to COS...")

# Rquest detailed enpoint list
endpoints = requests.get(credentials.get('endpoints')).json()
#import pdb; pdb.set_trace()

# Obtain iam and cos host from the the detailed endpoints
iam_host = (endpoints['identity-endpoints']['iam-token'])
cos_host = (endpoints['service-endpoints']['cross-region']['us']['public']['us-geo'])

api_key = credentials.get('apikey')
service_instance_id = credentials.get('resource_instance_id')

# Constrict auth and cos endpoint
auth_endpoint = "https://" + iam_host + "/oidc/token"
service_endpoint = "https://" + cos_host

print("Creating client...")
# Get bucket list
cos = boto3.client('s3',
                    ibm_api_key_id=api_key,
                    ibm_service_instance_id=service_instance_id,
                    ibm_auth_endpoint=auth_endpoint,
                    config=Config(signature_version='oauth'),
                    endpoint_url=service_endpoint)


# Call S3 to list current buckets
response = cos.list_buckets()

# Get a list of all bucket names from the response
buckets = [bucket['Name'] for bucket in response['Buckets']]

# Print out the bucket list
print("Current Bucket List:")
print(json.dumps(buckets, indent=2))
print("---")
result = [bucket for bucket in buckets if 'cos-bucket-sample-' in bucket]

print("Creating a new bucket and uploading an object...")
if len(result) == 0 :
   bucket_name = 'cos-bucket-sample-' + str(random.randint(100,99999999));
   # Create a bucket
   cos.create_bucket(Bucket=bucket_name)
   # Upload a file
   cos.upload_file('./example.py', bucket_name, 'example-object')

   # Call S3 to list current buckets
   response = cos.list_buckets()

   # Get a list of all bucket names from the response
   buckets = [bucket['Name'] for bucket in response['Buckets']]

   # Print out the bucket list
   print("New Bucket List:")
   print(json.dumps(buckets, indent=2))
   print("---")
else :
   bucket_name = result[0];


# Call S3 to list current objects
response = cos.list_objects(Bucket=bucket_name)

# Get a list of all object names from the response
objects = [object['Key'] for object in response['Contents']]

# Print out the object list
print("Objects in %s:" % bucket_name)
print(json.dumps(objects, indent=2))

但是在运行时，我得到以下输出：

回溯（最后一次调用）：文件“boto3test.py”，第 1 行，导入 boto3 文件“/home/giovanni/Downloads/boto-test/lib/python3.5/site-packages/boto3/init .py ” ，第 16 行，从 boto3.session 导入会话文件“/home/giovanni/Downloads/boto-test/lib/python3.5/site-packages/boto3/session.py”，第 27 行，在 import botocore.session 文件中“/home/giovanni/Downloads/boto-test/lib/python3.5/site-packages/botocore/session.py”，第 37 行，在 import botocore.credentials 文件中“/home/giovanni/Downloads/boto-test/ lib/python3.5/site-packages/botocore/credentials.py"，第 27 行，在 import httplib ImportError: No module named 'httplib'

我在哪里做错了什么？我应该在我的 virtualenv 中安装 botocore 吗？

4

2 回答 2

2

编辑 botocore/credentials.py 并将 import httplib 更改为 import http.client as httplib

于 2017-10-16T14:24:59.597 回答
1

我发现了问题：IBM 的库ibm-cos-sdk-python-core是他们自己的 botocore 库版本，但是，在他们的 repo 中的credentials.py上，有一个对 Python 3 上已重命名的库的引用（httplib -> http.client）。

所以我的解决方法是在我的本地安装目录中替换第 27 行的credentials.py：

import httplib

至：

import http.client as httplib

案例（#1）有一个未解决的问题，但是，我没有看到，因为存储库彼此没有联系，我仍在学习 IBM 的库是如何工作的。

于 2017-10-16T14:47:19.083 回答