IBM Cloudant Python SDK is a client library that interacts with the IBM Cloudant APIs.
Disclaimer: This library is still a 0.x release. We do consider this library production-ready and capable, but there are still some limitations we’re working to resolve, and refinements we want to deliver. We are working really hard to minimise the disruption from now until the 1.0 release, but there may still be some changes that impact applications using this SDK. For now, be sure to pin versions to avoid surprises.
The IBM Cloudant Python SDK allows developers to programmatically
interact with IBM Cloudant
with the help of the ibmcloudant package.
The purpose of this Python SDK is to wrap most of the HTTP request APIs provided by Cloudant and supply other functions to ease the usage of Cloudant. This SDK should make life easier for programmers to do what’s really important to them: developing software.
Reasons why you should consider using Cloudant Python SDK in your project:
To install, use pip or easy_install:
pip install --upgrade "ibmcloudant>=0.9.3"
or
easy_install --upgrade "ibmcloudant>=0.9.3"
For fundamental SDK usage information and config options, please see the common IBM Cloud SDK documentation.
This library requires configuration with a service URL and Cloudant service credentials to authenticate with your account.
There are several ways to set these authentication properties:
The following section describes the different authentication types and provides environment variable examples. Examples for other configuration methods are available by following the provided links.
This library requires credentials to authenticate with IBM Cloudant. These credentials may be:
CONTAINER, VPC and IAM)
COUCHDB_SESSION and BASIC)
IAM only.For other compatible APIs that are not Cloudant accounts (e.g. Apache CouchDB) non-IAM based authentication types must be used.
This table summarizes the available authentication types. The authentication types are listed in order of recommendation, preferably use the authentication type from the first row in the table that is compatible with your environment.
| Authentication type | Recommended for | AUTH_TYPE | Description |
|---|---|---|---|
| IAM Trusted Profiles compute resource (container) | Cloudant (SDK running in IBM Cloud IKS) | CONTAINER | Obtains a compute resource (CR) token from the container. Exchanges the CR token for an IAM access_token.Adds an Authorization: Bearer <access_token> header to each HTTP request.Automatically renews the access token when needed. |
| IAM Trusted Profiles compute resource (VPC) | Cloudant (SDK running in IBM Cloud VPC) | VPC | Obtains an identity token from the VPC instance metadata. Exchanges the identity token for an IAM access_token.Adds an Authorization: Bearer <access_token> header to each HTTP request.Automatically renews the access token when needed. |
| IAM API key | Cloudant | IAM | Exchanges an IAM API key for an IAM access_token.Adds an Authorization: Bearer <access_token> header to each HTTP request.Automatically renews the access token when needed. |
| Session cookie | Cloudant (legacy credentials & instances without IAM) Apache CouchDB | COUCHDB_SESSION | Exchanges credentials with /_session endpoint to retrieve a cookie.Adds Cookie header and content to each HTTP request.Automatically renews session when needed. |
| Bearer token | Apache CouchDB (using JWT authentication) | BEARERTOKEN | Adds an Authorization: Bearer <token> to each HTTP request.No token management or renewal. Also compatible with IAM access tokens managed independently of the SDK. |
| Basic | Apache CouchDB (if cookies are not enabled) | BASIC | Adds an Authorization: Basic <encoded username and password> header to each HTTP request. |
| None | - | NOAUTH | Note that this authentication type only works for operations against a database allowing access for unauthenticated users. |
The default authentication type for the SDK is CONTAINER unless APIKEY configuration is supplied, which changes the default authentication type to IAM.
The default service name is CLOUDANT so CLOUDANT_ prefixed names are used in these examples.
Any custom service name prefix can be used as long as the matching name is used to instantiate the SDK client and the same prefix is used for all configuration options.
For Cloudant IAM API key authentication, set the following environmental variables by
replacing the <url> and <apikey> with your proper
service credentials. There is no need to set
CLOUDANT_AUTH_TYPE to IAM because it is the default when an APIKEY is set.
CLOUDANT_URL=<url>
CLOUDANT_APIKEY=<apikey>
For Cloudant IAM Trusted profile compute resource container authentication, set the following environmental variables by
replacing the <url> and <id> with your values. There is no need to set
CLOUDANT_AUTH_TYPE to CONTAINER because it is the default.
CLOUDANT_URL=<url>
CLOUDANT_IAM_PROFILE_ID=<id>
Alternatively a profile name may be used instead of an ID by replacing CLOUDANT_IAM_PROFILE_ID with CLOUDANT_IAM_PROFILE_NAME.
For Cloudant IAM Trusted profile compute resource vpc authentication, set the following environmental variables by
replacing the <url> and <id> with your values.
CLOUDANT_AUTH_TYPE=VPC
CLOUDANT_URL=<url>
CLOUDANT_IAM_PROFILE_ID=<id>
Alternatively a profile CRN may be used instead of an ID by replacing CLOUDANT_IAM_PROFILE_ID with CLOUDANT_IAM_PROFILE_CRN.
For COUCHDB_SESSION authentication, set the following environmental variables
by replacing the <url>, <username> and <password> with your proper
service credentials.
CLOUDANT_AUTH_TYPE=COUCHDB_SESSION
CLOUDANT_URL=<url>
CLOUDANT_USERNAME=<username>
CLOUDANT_PASSWORD=<password>
To use an external configuration file, the Cloudant API docs, or the general SDK usage information will guide you.
To learn more about how to use programmatic authentication, see the related documentation in the Cloudant API docs or in the Python SDK Core document about authentication.
The SDK supports a generalized retry feature that can automatically retry on common errors.
The automatic retries section has details on how to enable the retries with default values and customize the retries programmatically or with external configuration.
No request timeout is defined, but a 2.5m read and a 60s connect timeout are set by default. Be sure to set a request timeout appropriate to your application usage and environment. The request timeout section contains details on how to change the value.
Note: System settings may take precedence over configured timeout values.
The following code examples authenticate with the environment variables.
Note: This example code assumes that orders database does not exist in your account.
This example code creates orders database and adds a new document "example"
into it. To connect, you must set your environment variables with
the service url, authentication type and authentication credentials
of your Cloudant service.
Cloudant environment variable naming starts with a service name prefix that identifies your service.
By default, this is CLOUDANT, see the settings in the
authentication with environment variables section.
If you would like to rename your Cloudant service from CLOUDANT,
you must use your defined service name as the prefix for all Cloudant related environment variables.
Once the environment variables are set, you can try out the code examples.
from ibm_cloud_sdk_core import ApiException
from ibmcloudant.cloudant_v1 import CloudantV1, Document
# 1. Create a client with `CLOUDANT` default service name =============
client = CloudantV1.new_instance()
# 2. Create a database ================================================
example_db_name = "orders"
# Try to create database if it doesn't exist
try:
put_database_result = client.put_database(
db=example_db_name
).get_result()
if put_database_result["ok"]:
print(f'"{example_db_name}" database created.')
except ApiException as ae:
if ae.status_code == 412:
print(f'Cannot create "{example_db_name}" database, ' +
'it already exists.')
# 3. Create a document ================================================
# Create a document object with "example" id
example_doc_id = "example"
# Setting `id` for the document is optional when "post_document"
# function is used for CREATE. When `id` is not provided the server
# will generate one for your document.
example_document: Document = Document(id=example_doc_id)
# Add "name" and "joined" fields to the document
example_document.name = "Bob Smith"
example_document.joined = "2019-01-24T10:42:59.000Z"
# Save the document in the database with "post_document" function
create_document_response = client.post_document(
db=example_db_name,
document=example_document
).get_result()
# =====================================================================
# Note: saving the document can also be done with the "put_document"
# function. In this case `doc_id` is required for a CREATE operation:
"""
create_document_response = client.put_document(
db=example_db_name,
doc_id=example_doc_id,
document=example_document
).get_result()
"""
# =====================================================================
# Keeping track of the revision number of the document object
# is necessary for further UPDATE/DELETE operations:
example_document.rev = create_document_response["rev"]
print(f'You have created the document:\n{example_document}')
When you run the code, you see a result similar to the following output.
"orders" database created.
You have created the document:
{
"_id": "example",
"_rev": "1-1b403633540686aa32d013fda9041a5d",
"name": "Bob Smith",
"joined": "2019-01-24T10:42:99.000Z"
}
Note: This example code assumes that you have created both the orders
database and the example document by
running the previous example code
successfully. Otherwise, the following error message occurs, "Cannot delete document because either 'orders'
database or 'example' document was not found."
import json
from ibmcloudant.cloudant_v1 import CloudantV1
# 1. Create a client with `CLOUDANT` default service name ============
client = CloudantV1.new_instance()
# 2. Get server information ===========================================
server_information = client.get_server_information(
).get_result()
print(f'Server Version: {server_information["version"]}')
# 3. Get database information for "orders" ==========================
db_name = "orders"
db_information = client.get_database_information(
db=db_name
).get_result()
# 4. Show document count in database ==================================
document_count = db_information["doc_count"]
print(f'Document count in \"{db_information["db_name"]}\" '
f'database is {document_count}.')
# 5. Get "example" document out of the database by document id ============
document_example = client.get_document(
db=db_name,
doc_id="example"
).get_result()
print(f'Document retrieved from database:\n'
f'{json.dumps(document_example, indent=2)}')
Server Version: 2.1.1
Document count in "orders" database is 1.
Document retrieved from database:
{
"_id": "example",
"_rev": "1-1b403633540686aa32d013fda9041a5d",
"name": "Bob Smith",
"joined": "2019-01-24T10:42:99.000Z"
}
Note: This example code assumes that you have created both the orders
database and the example document by
running the previous example code
successfully. Otherwise, the following error message occurs, "Cannot update document because either 'orders'
database or 'example' document was not found."
import json
from ibm_cloud_sdk_core import ApiException
from ibmcloudant.cloudant_v1 import CloudantV1
# 1. Create a client with `CLOUDANT` default service name =============
client = CloudantV1.new_instance()
# 2. Update the document ==============================================
example_db_name = "orders"
example_doc_id = "example"
# Try to get the document if it previously existed in the database
try:
document = client.get_document(
db=example_db_name,
doc_id=example_doc_id
).get_result()
# =================================================================
# Note: for response byte stream use:
"""
document_as_byte_stream = client.get_document_as_stream(
db=example_db_name,
doc_id=example_doc_id
).get_result()
"""
# =================================================================
# Add Bob Smith's address to the document
document["address"] = "19 Front Street, Darlington, DL5 1TY"
# Remove the joined property from document object
if "joined" in document:
document.pop("joined")
# Update the document in the database
update_document_response = client.post_document(
db=example_db_name,
document=document
).get_result()
# =================================================================
# Note 1: for request byte stream use:
"""
update_document_response = client.post_document(
db=example_db_name,
document=document_as_byte_stream
).get_result()
"""
# =================================================================
# =================================================================
# Note 2: updating the document can also be done with the
# "put_document" function. `doc_id` and `rev` are required for an
# UPDATE operation, but `rev` can be provided in the document
# object as `_rev` too:
"""
update_document_response = client.put_document(
db=example_db_name,
doc_id=example_doc_id, # doc_id is a required parameter
rev=document["_rev"],
document=document # _rev in the document object CAN replace above `rev` parameter
).get_result()
"""
# =================================================================
# Keeping track of the latest revision number of the document
# object is necessary for further UPDATE/DELETE operations:
document["_rev"] = update_document_response["rev"]
print(f'You have updated the document:\n' +
json.dumps(document, indent=2))
except ApiException as ae:
if ae.status_code == 404:
print('Cannot delete document because either ' +
f'"{example_db_name}" database or "{example_doc_id}" ' +
'document was not found.')
{
"_id": "example",
"_rev": "2-4e2178e85cffb32d38ba4e451f6ca376",
"name": "Bob Smith",
"address": "19 Front Street, Darlington, DL5 1TY"
}
Note: This example code assumes that you have created both the orders
database and the example document by
running the previous example code
successfully. Otherwise, the following error message occurs, "Cannot delete document because either 'orders'
database or 'example' document was not found."
from ibm_cloud_sdk_core import ApiException
from ibmcloudant.cloudant_v1 import CloudantV1
# 1. Create a client with `CLOUDANT` default service name =============
client = CloudantV1.new_instance()
# 2. Delete the document ==============================================
example_db_name = "orders"
example_doc_id = "example"
# Try to get the document if it previously existed in the database
try:
document = client.get_document(
db=example_db_name,
doc_id=example_doc_id
).get_result()
delete_document_response = client.delete_document(
db=example_db_name,
doc_id=example_doc_id, # `doc_id` is required for DELETE
rev=document["_rev"] # `rev` is required for DELETE
).get_result()
if delete_document_response["ok"]:
print('You have deleted the document.')
except ApiException as ae:
if ae.status_code == 404:
print('Cannot delete document because either ' +
f'"{example_db_name}" database or "{example_doc_id}"' +
'document was not found.')
You have deleted the document.
For a complete list of code examples, see the examples directory.
For sample code on handling errors, see Cloudant API docs.
For endpoints that read or write document content it is possible to bypass usage of the built-in object with byte streams.
Depending on the specific SDK operation it may be possible to:
Request byte stream can be supplied for arguments that accept the BinaryIO type.
For these cases you can pass this byte stream directly to the HTTP request body.
Response byte stream is supported in functions with the suffix of _as_stream.
The returned byte stream allows the response body to be consumed
without triggering JSON unmarshalling that is typically performed by the SDK.
The update document section contains examples for both request and response byte stream cases.
The API reference contains further examples of using byte streams. They are titled "Example request as stream" and are initially collapsed. Expand them to see examples of:
Byte requests:
Byte responses:
This SDK supports two possible formats to define an HTTP request. One approach uses only model classes and the other only dictionaries.
from ibmcloudant.cloudant_v1 import DesignDocument, CloudantV1, DesignDocumentOptions, SearchIndexDefinition
client = CloudantV1.new_instance()
price_index = SearchIndexDefinition(
index='function (doc) { index("price", doc.price); }'
)
design_document_options = DesignDocumentOptions(
partitioned=True
)
partitioned_design_doc = DesignDocument(
indexes={'findByPrice': price_index},
options=design_document_options
)
response = client.put_design_document(
db='products',
design_document=partitioned_design_doc,
ddoc='appliances'
).get_result()
print(response)
from ibmcloudant.cloudant_v1 import CloudantV1
client = CloudantV1.new_instance()
price_index = {
'index': 'function (doc) { index("price", doc.price); }'
}
partitioned_design_doc = {
'indexes': {'findByPrice': price_index},
'options': {'partitioned': True},
}
response = client.put_design_document(
db='products',
design_document=partitioned_design_doc,
ddoc='appliances'
).get_result()
print(response)
Since model classes and dicts are different data structures, they cannot be combined.
from ibmcloudant.cloudant_v1 import CloudantV1, DesignDocument
client = CloudantV1.new_instance()
price_index = {
'index': 'function (doc) { index("price", doc.price); }'
}
partitioned_design_doc = DesignDocument(
indexes={'findByPrice': price_index},
options={'partitioned': True}
)
response = client.put_design_document(
db='products',
design_document=partitioned_design_doc,
ddoc='appliances'
).get_result()
print(response)
The SDK provides a changes feed follower utility (currently beta).
This helper utility connects to the _changes endpoint and returns the individual change items.
It removes some of the complexity of using the _changes endpoint by setting some options automatically
and providing error suppression and retries.
Tip: the changes feed often does not meet user expectations or assumptions.
Consult the Cloudant changes feed FAQ to get a better understanding of the limitations and suitable use-cases before using the changes feed in your application.
There are two modes of operation:
since sequence (by default the feed will start from now).since sequence (by default the feed will start from the beginning).The SDK's model of changes feed options is also used to configure the follower. However, a subset of the options are invalid as they are configured internally by the implementation. Supplying these options when instantiating the follower causes an error. The invalid options are:
descendingfeedheartbeatlastEventId - use since insteadtimeout_selector is permitted for the filter option. This restriction is because selector
based filters perform better than JavaScript backed filters. Configuring a non-selector based filter will
cause the follower to error.Note that the limit parameter will terminate the follower at the given number of changes in either
operating mode.
The changes follower requires the client to have HTTP timeouts of at least 1 minute and will error during instantiation if it is insufficient. The default client configuration has sufficiently long timeouts.
For use-cases where these configuration limitations are deemed too restrictive then it is recommended to
write code to use the SDK's POST _changes API instead of the follower.
By default, the changes follower will suppress transient errors indefinitely and attempt to run to completion or listen forever as dictated by the operating mode. For applications where that is not desirable an optional error tolerance duration may be specified to control the time since the last successful response that transient errors will be suppressed. This can be used, for example, by applications as a grace period before reporting an error and requiring intervention.
There are some additional points to consider for error suppression:
For both modes:
400, 401, 403 404).limit.As is true for the _changes endpoint change items have at least once delivery and an individual item
may be received multiple times. When using the follower change items may be repeated even within a limited
number of changes (i.e. using the limit option) this is a minor difference from using limit on the HTTP native API.
The follower is not optimized for some use cases and it is not recommended to use it in cases where:
include_docs and larger document sizes (for example > 10 kiB).In these cases use-case specific control over the number of change requests made and the content size of the responses
may be achieved by using the SDK's POST _changes API.
The changes follower does not checkpoint since it has no information about whether a change item has been
processed by the consuming application after being received. It is the application developer's responsibility
to store the sequence IDs to have appropriate checkpoints and to re-initialize the follower with the required
since value after, for example, the application restarts.
The frequency and conditions for checkpointing are application specific and some applications may be tolerant of dropped changes. This section is intended only to provide general guidance on how to avoid missing changes.
To guarantee processing of all changes the sequence ID from a change item must not be persisted until after
the processing of the change item by the application has completed. As indicated previously change items are
delivered at least once so application code must be able to handle repeated changes already and it is
preferable to restart from an older since value and receive changes again than risk missing them.
The sequence IDs are available on each change item by default, but may be omitted from some change items when
using the seq_interval configuration option. Infrequent sequence IDs may improve performance by reducing
the amount of data that needs to be transferred, but the trade-off is that more changes will be repeated if
it is necessary to resume the changes follower.
Extreme care should be taken with persisting sequences if choosing to process change items in parallel as there is a considerable risk of missing changes on a restart if the sequence is recorded out of order.
import ChangesFollower
from ibmcloudant.cloudant_v1 import CloudantV1
client = CloudantV1.new_instance()
cf_params = {
'db': 'example', # Required: the database name.
'limit': 100, # Optional: return only 100 changes (including duplicates).
'since': '3-g1AG3...' # Optional: start from this sequence ID (e.g. with a value read from persistent storage).
}
changes_follower = ChangesFollower(
service=client, # Required: the Cloudant service client instance.
error_tolerance=10000, # Optional: suppress transient errors for at least 10 seconds before terminating.
**cf_params # Required: changes feed configuration options dict.
)
import Iterable
from ibmcloudant import ChangesFollower
from ibmcloudant.cloudant_v1 import CloudantV1, ChangesResultItem
client = CloudantV1.new_instance()
changes_follower = ChangesFollower(
service=client,
**{'db': 'example'})
changes_items: Iterable[ChangesResultItem] = changes_follower.start()
# Note: iterable will not do anything until it is iterated
# Create a for loop to iterate over the flow of changes
# for changes_item in changes_items: ...
import Iterable
from ibmcloudant import ChangesFollower
from ibmcloudant.cloudant_v1 import CloudantV1, ChangesResultItem
client = CloudantV1.new_instance()
changes_follower = ChangesFollower(
service=client,
**{'db': 'example'})
changes_items: Iterable[ChangesResultItem] = changes_follower.start_one_off()
# Note: iterable will not do anything until it is iterated
# Create a for loop to iterate over the flow of changes
# for changes_item in changes_items: ...
import ChangesFollower
from ibmcloudant.cloudant_v1 import CloudantV1
client = CloudantV1.new_instance()
# Start from a previously persisted seq
# Normally this would be read by the app from persistent storage
# e.g. previously_persisted_seq = your_app_persistence_read_func()
previously_persisted_seq = '3-g1AG3...'
changes_follower = ChangesFollower(
service=client,
**{'db': 'example', 'since': previously_persisted_seq})
changes_items = changes_follower.start()
for changes_item in changes_items:
# do something with changes
print(changes_item.id)
for change in changes_item.changes:
print(change.rev)
# when change item processing is complete app can store seq
seq = changes_item.seq
# write seq to persistent storage for use as since if required to resume later
# e.g. your_app_persistence_write_func(seq)
# keep processing changes until the application is terminated or some other stop condition is reached
# Note: iterator above is blocking, code here will be unreachable
# until the iteration is stopped or another stop condition is reached.
# For long running followers careful consideration should be made of where to call stop on the iterator.
import ChangesFollower
from ibmcloudant.cloudant_v1 import CloudantV1
client = CloudantV1.new_instance()
# Start from a previously persisted seq
# Normally this would be read by the app from persistent storage
# e.g. previously_persisted_seq = your_app_persistence_read_func()
previously_persisted_seq = '3-g1AG3...'
changes_follower = ChangesFollower(
service=client,
**{'db': 'example', 'since': previously_persisted_seq})
changes_items = changes_follower.start_one_off()
for changes_item in changes_items:
# do something with changes
print(changes_item.id)
for change in changes_item.changes:
print(change.rev)
# when change item processing is complete app can store seq
seq = changes_item.seq
# write seq to persistent storage for use as since if required to resume later
# e.g. your_app_persistence_write_func(seq)
# Note: iterator above is blocking, code here will be unreachable
# until all changes are processed (or another stop condition is reached).
import ChangesFollower
from ibmcloudant.cloudant_v1 import CloudantV1
client = CloudantV1.new_instance()
changes_follower = ChangesFollower(
service=client,
**{'db': 'example'})
changes_items = changes_follower.start()
for changes_item in changes_items:
# Option 1: call stop after some condition
# Note that since the iterator is blocking at least one item
# must be returned from it to reach to this point.
# Additional changes may be processed before the iterator stops.
changes_follower.stop()
# Option 2: call stop method when you want to end the continuous loop from
# outside the iterator. For example, you've put the changes follower in a
# separate thread and need to call stop on the main thread.
# Note: in this context the call must be made from a different thread because
# code immediately following the iterator is unreachable until the iterator
# has stopped.
changes_follower.stop()
If you are having difficulties using this SDK or have a question about the IBM Cloud services, ask a question on Stack Overflow.
If you encounter an issue with the project, you are welcome to submit a bug report.
Before you submit a bug report, search for similar issues and review the KNOWN_ISSUES file to verify that your issue hasn't been reported yet.
Please consult the security policy before opening security related issues.
This SDK follows semantic versioning with respect to the definition of user facing APIs. This means under some circumstances breaking changes may occur within a major or minor version of the SDK related to changes in supported language platforms.
The SDK is supported on the available LTS releases of the language platform. The LTS language versions are listed in the prerequisites:
Incompatible changes from new language versions are not added to the SDK until they are available in the minimum supported language version.
When language LTS versions move out of support the following will happen:
Find more open source projects on the IBM GitHub page.
For more information, see CONTRIBUTING.
This SDK is released under the Apache 2.0 license. To read the full text of the license, see LICENSE.
FAQs
Python client library for IBM Cloudant
We found that ibmcloudant demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 2 open source maintainers collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
Socket CEO Feross Aboukhadijeh discusses open source security challenges, including zero-day attacks and supply chain risks, on the Cyber Security Council podcast.
Security News
Research
Socket researchers uncover how threat actors weaponize Out-of-Band Application Security Testing (OAST) techniques across the npm, PyPI, and RubyGems ecosystems to exfiltrate sensitive data.
Security News
Research
A malicious npm campaign is targeting Ethereum developers by impersonating Hardhat plugins and the Nomic Foundation, stealing sensitive data like private keys.