A thin Python client for the Draftable API which wraps all available endpoints and handles authentication and signing.
See the full API documentation for an introduction to the API, usage notes, and other reference material.
- Operating system: Any maintained Linux, macOS, or Windows release
- Python runtime: Any maintained version (currently 3.8, 3.9, 3.10, 3.11, 3.12, and 3.13)
- Create a free API account
- Retrieve your credentials
- Install the library
pip install draftable-compare-api- Instantiate a client
import draftable
client = draftable.Client('<yourAccountId>', '<yourAuthToken>')
comparisons = client.comparisons- Start creating comparisons
comparison = comparisons.create(
'https://api.draftable.com/static/test-documents/code-of-conduct/left.rtf',
'https://api.draftable.com/static/test-documents/code-of-conduct/right.pdf'
)
print("Comparison created: {}".format(comparison))
# Generate a signed viewer URL to access the private comparison. The expiry
# time defaults to 30 minutes if the valid_until parameter is not provided.
viewer_url = comparisons.signed_viewer_url(comparison.identifier)
print("Viewer URL (expires in 30 mins): {}".format(viewer_url))A helper utility, dr-compare, is included for interacting with the API from the command-line. After installing the library run dr-compare to view the built-in help.
The CLI tool will only be available in the environment under which the library is installed (e.g. a given virtualenv). To ensure it's accessible outside of any given environment install the library user or system-wide (e.g. sudo pip install draftable-compare-api).
- The handling of
datetimeobjects is as follows:- Any naive
datetimeobjects provided in a method call are assumed to be in UTC time. - Returned
datetimeobjects are always "aware" (include timezone information) and use UTC.
- Any naive
The package provides a module, draftable, which exports a class to create a Client for your API account.
Client provides a comparisons property which yields a ComparisonsEndpoint to manage the comparisons for your API account.
Creating a Client differs slightly based on the API endpoint being used:
import draftable
# Draftable API (default endpoint)
account_id = '<yourAccountId>' # Replace with your API credentials from:
auth_token = '<yourAuthToken>' # https://api.draftable.com/account/credentials
client = draftable.Client(account_id, auth_token)
comparisons = client.comparisons
# Draftable API regional endpoint or Self-hosted
base_url = 'https://draftable.example.com/api/v1' # Replace with the endpoint URL
account_id = '<yourAccountId>' # Replace with your API credentials from the regional
auth_token = '<yourAuthToken>' # Draftable API endpoint or your Self-hosted container
client = draftable.Client(account_id, auth_token, base_url)
comparisons = client.comparisonsFor API Self-hosted you may need to suppress TLS certificate validation if the server is using a self-signed certificate (the default).
Instances of the ComparisonsEndpoint class provide the following methods for retrieving comparisons:
all()
Returns alistof all your comparisons, ordered from newest to oldest. This is potentially an expensive operation.get(identifier: str)
Returns the specifiedComparisonor raises aNotFoundexception if the specified comparison identifier does not exist.
Comparison objects have the following properties:
identifier: str
The unique identifier of the comparison.left: object/right: object
Information about each side of the comparison.file_type: str
The file extension.source_url: str(optional)
The URL for the file if the original request was specified by URL, otherwiseNone.display_name: str(optional)
The display name for the file if given in the original request, otherwiseNone.
public: bool
Indicates if the comparison is public.creation_time: datetime
Time in UTC when the comparison was created.expiry_time: datetime(optional)
The expiry time if the comparison is set to expire, otherwiseNone.ready: bool
Indicates if the comparison is ready to display.
If a Comparison is ready (i.e. it has been processed) it has the following additional properties:
ready_time: datetime
Time in UTC the comparison became ready.failed: bool
Indicates if comparison processing failed.error_message: str(only present iffailed)
Reason processing of the comparison failed.
from draftable.endpoints import exceptions
identifier = '<identifier>'
try:
comparison = comparisons.get(identifier)
print("Comparison '{identifier}' ({visibility}) is {status}.".format(
identifier = identifier,
visibility = 'public' if comparison.public else 'private',
status = 'ready' if comparison.ready else 'not ready'
))
if comparison.ready:
elapsed = comparison.ready_time - comparison.creation_time
print("The comparison took {} seconds.".format(elapsed.total_seconds()))
if comparison.failed:
print("The comparison failed with error: {}".format(comparison.error_message))
except exceptions.NotFound:
print("Comparison '{}' does not exist.".format(identifier))Instances of the ComparisonsEndpoint class provide the following methods for deleting comparisons:
delete(identifier: str)
Returns nothing on successfully deleting the specified comparison or raises aNotFoundexception if no such comparison exists.
oldest_comparisons = comparisons.all()[-10:]
print("Deleting oldest {} comparisons ...".format(len(oldest_comparisons)))
for comparison in oldest_comparisons:
comparisons.delete(comparison.identifier)
print("Comparison '{}' deleted.".format(comparison.identifier))Instances of the ComparisonsEndpoint class provide the following methods for retrieving comparisons:
create(left: ComparisonSide, right: ComparisonSide, identifier: str = None, public: bool = False, expires: datetime | timedelta = None)
Returns aComparisonrepresenting the newly created comparison.
create accepts the following arguments:
left/right
Describes the left and right files (see following section).identifier(optional)
Identifier to use for the comparison:- If specified, the identifier must be unique (i.e. not already be in use).
- If unspecified or
None, the API will automatically generate a unique identifier.
public(optional)
Specifies the comparison visibility:- If
Falseor unspecified authentication is required to view the comparison. - If
Truethe comparison can be accessed by anyone with knowledge of the URL.
- If
expires(optional)
Time at which the comparison will be deleted:- Must be specified as a
datetimeor atimedelta(UTC if naive). - If specified, the provided expiry time must be UTC and in the future.
- If unspecified or
None, the comparison will never expire (but may be explicitly deleted).
- Must be specified as a
Note: The comparison must be retrieved via the comparisons.get(<identifier>) call to check the 'ready' status for new comparisons. This is an important
step before exporting or accessing consecutive comparisons in any code loops.
The following exceptions may be raised:
BadRequest
The request could not be processed (e.g.identifieralready in use).InvalidArgument
Failure in parameter validation (e.g.expiresis in the past).
The draftable module provides the following static methods for creating comparison sides:
draftable.make_side(url_or_file_path: str, file_type: str = None, display_name: str = None)
Returns aComparisonSidefor a file or URL by attempting to guess from theurl_or_file_pathparameter.
Alternatively, for explicitly creating a file or URL comparison side, the following static methods can be used:
draftable.endpoints.comparisons.sides.side_from_file(file: object, file_type: str, display_name: str = None)
Returns aComparisonSidefor a locally accessible file.draftable.endpoints.comparisons.sides.side_from_url(url: str, file_type: str, display_name: str = None)
Returns aComparisonSidefor a remotely accessible file referenced by URL.
These methods accept the following arguments:
url_or_file_path(make_sideonly)
The file or URL path for a comparison side.file(side_from_fileonly)
A file object to be read and uploaded.- The file must be opened for reading in binary mode.
url(side_from_urlonly)
The URL from which the server will download the file.file_type
The type of file being submitted:- PDF:
pdf - Word:
docx,docm,doc,rtf - PowerPoint:
pptx,pptm,ppt - Other:
txt
- PDF:
display_name(optional)
The name of the file shown in the comparison viewer.
The following exceptions may be raised:
InvalidArgument
Failure in parameter validation (e.g.file_typeis invalid,urlis malformed, orfileis not opened in binary mode).
The make_side method may additionally raise the following exceptions:
InvalidPath
The provided path uses a file URI scheme (i.e.file:) but references a remote host or the file doesn't exist.ValueError
The provided path was not a URL or the file doesn't exist.
from datetime import timedelta
identifier = draftable.generate_identifier()
comparison = comparisons.create(
identifier = identifier,
left = draftable.make_side(
'https://domain.com/left.pdf',
file_type='pdf',
display_name='Document.pdf'
),
right = draftable.make_side(
'path/to/right/file.docx',
file_type='docx',
display_name='Document (revised).docx'
),
# Expire this comparison in 2 hours (default is no expiry)
expires = timedelta(hours=2)
)
print("Created comparison: {}".format(comparison))Instances of the ComparisonsEndpoint class provide the following methods for displaying comparisons:
public_viewer_url(identifier: str, wait: bool = False)
Generates a public viewer URL for the specified comparison.signed_viewer_url(identifier: str, valid_until: datetime | timedelta = None, wait: bool = False)
Generates a signed viewer URL for the specified comparison.
Both methods use the following common parameters:
identifier
Identifier of the comparison for which to generate a viewer URL.wait(optional)
Specifies the behaviour of the viewer if the provided comparison does not exist.- If
Falseor unspecified, the viewer will show an error if theidentifierdoes not exist. - If
True, the viewer will wait for a comparison with the providedidentifierto exist.
Note this will result in a perpetual loading animation if theidentifieris never created.
- If
The signed_viewer_url method also supports the following parameters:
valid_until(optional)
Time at which the URL will expire (no longer load).- Must be specified as a
datetimeor atimedelta. - If specified, the provided expiry time must be UTC and in the future.
- If unspecified or
None, the URL will be generated with the default 30 minute expiry.
- Must be specified as a
See the displaying comparisons section in the API documentation for additional details.
identifier = '<identifier>'
# Retrieve a signed viewer URL which is valid for 1 hour. The viewer will wait
# for the comparison to exist in the event processing has not yet completed.
viewer_url = comparisons.signed_viewer_url(identifier, timedelta(hours=1), wait=True)
print("Viewer URL (expires in 1 hour): {}".format(viewer_url))To perform comparison exports retrieve an ExportsEndpoint instance via the exports property of your Client instance:
exports = client.exportsInstances of the ExportsEndpoint class provide the following methods for exporting comparisons:
create(comparison: Union[Comparison, str], kind: str = 'single_page', include_cover_page: bool = False)
Returns aExportrepresenting the newly created export.
create accepts the following arguments:
comparison
The comparison to export provided as aComparisoninstance or a comparison identifier.kind
The type of export to perfom.include_cover_page
Whether a cover page should be included (combinedexport kind only).
The following export kinds are supported for the kind parameter:
single_page
Exports only the right-side of the comparison with highlights and deletion markers.combined
Exports both the left and right sides.left
Export the left side only.right
Exports the right side only.
exports = client.exports
export = exports.create(comparison, kind='single_page')Instances of the ExportsEndpoint class provide the following methods for retrieving exports:
get(identifier: str)
Returns the specifiedExportor raises aNotFoundexception if the specified export identifier does not exist.
Export objects have the same properties as Comparison objects.
As with comparison requests, export requests need to be processed. Typically processing only takes a few seconds, but this can vary based on numerous factors (e.g. size of the documents being exported, number of changes in the comparison, etc ...). To determine if an export has been processed the ready property of the Export instance should be inspected and polling can be utilised to wait until the export has been completed.
import time
export = client.exports.create(comparison.identifier)
while not export.ready:
time.sleep(1)
export = client.exports.get(export.identifier)
if export.ready:
print(export.url)A dictionary that describes the changes between the two documents is available, once the comparison is ready. This method returns a draftable.endpoints.comparisons.changes.ChangeDetails object.
identifier = '<identifier>'
# Retrieve a ChangeDetails object that captures all changes between the two documents.
change_details = comparisons.change_details(identifier)
print("Changes summary: {}".format(change_details.summary))The draftable module provides the following static methods for generating comparison identifiers:
draftable.generate_identifier()
Generates a random unique comparison identifier.
If connecting to an API Self-hosted endpoint which is using a self-signed certificate (the default) you will need to suppress certificate validation. This can be done by setting the CURL_CA_BUNDLE environment variable to an empty string. On Windows, this must be done from within the Python interpreter due to platform limitations.
See the below examples for different operating systems and shell environments. Note that all examples only set the variable for the running shell or Python interpreter and it will not persist. To persist the setting consult the documentation for your shell environment. This should be done with caution as this setting suppresses certificate validation for all connections made by the Python runtime!
(ba)sh (Linux, macOS, WSL)
export CURL_CA_BUNDLE=0PowerShell:
$env:CURL_CA_BUNDLE=0Setting an environment variable to the empty string is not valid in Windows and is treated as equivalent to removing any existing environment variable of the same name. As such, suppressing certificate validation requires an alternate approach. The most straightforward is to set the environment variable from within Python, instead of before launch.
import os
os.environ['CURL_CA_BUNDLE'] = ''If your code spawns Python subprocesses they must separately modify their environment as the change will not be inherited as you'd normally expect.
Disabling certificate validation in production environments is strongly discouraged as it significantly lowers security. We only recommend setting this environment variable in development environments if configuring a CA signed certificate for API Self-hosted is not possible.
All content is licensed under the terms of The MIT License.