Skip to content

Latest commit

 

History

History
276 lines (187 loc) · 11.7 KB

README.md

File metadata and controls

276 lines (187 loc) · 11.7 KB
Raw

pysnyk

Build status

A Python client for the Snyk API.

Client

Using the client requires you to provide your Snyk API token.

import snyk
client = snyk.SnykClient("<your-api-token>")

By default the client will connect to the public Snyk service. If you are using a local installation then you can provide the API url as the second argument.

import snyk
client = snyk.SnykClient("<your-api-token>", "<your-instance-of-snyk>")

By default the User-Agent string for the API requests will be pysnyk/<version>. If you would like to send a custom user agent you can do so as follows:

import snyk
client = snyk.SnykClient("<your-api-token>", user_agent="<your-instance-of-snyk>")

By default the requests are not retried. If you would like to retry failed requests with backoff and delay you can do so as follows:

import snyk
client = snyk.SnykClient("<your-api-token>", tries=4, delay=1, backoff=2)
  • tries - the maximum number of attempts. Default: 1 (no retries)
  • delay - initial delay between attempts. Default: 1
  • backoff - multiplier applied to delay between attempts. Default: 2
  • debug - run client in debug mode, useful for debugging API requests. Default: False

Organizations

With the client we can get a list of Snyk organizations you are a member of:

client.organizations.all()

This returns a list of snyk.models.Organization objects.

If you already have the ID of the organization you're after you can grab it directly:

client.organizations.get("<organization-id>")

This will return a single snyk.models.Organization object.

Most of the API is scoped to organizations, so most other methods are found on the snyk.models.Organization objects returned by these two methods.

The snyk.models.Organization object has the following properties related to the API:

  • entitlements - returns the set of Snyk features available to this account
  • dependencies- returns a Manager for packages in use in this organization
  • licenses - returns a Manager for licenses currently in use by projects in this organisation
  • members - returns a Manager for members
  • projects - returns a Manager for associated projects
  • integrations - returns a Manager for active integrations

A note on Managers

Managers provide a consistent API for accessing objects from the Snyk API. Each manager implements the following methods:

  • all() - return a list of all of the relevant objects
  • get("<id>") - return a single instance of the object if it exists
  • first() - grab the first instance of the object if one exists
  • filter(<key>="<value>") - return a list filtered by one or more key/value pairs

Projects

Once you have an organization you're likely to want to grab the related projects:

client.organizations.first().projects.all()

This will return a list of snyk.models.Project objects.

In the case where you want to get all of the projects across all of your organizations then you can use the handy method on the client.

client.projects.all()

The snyk.models.Project object has the following useful properties and methods:

  • delete() - deletes the project in question. Be careful as this will delete all associated data too
  • move(new_org_id) - moves the project to a new organization
  • dependencies - returns a Manager for packages in use in this project
  • dependency_graph - returns a snyk.models.DependencyGraph object which represents the full dependency graph of package dependencies
  • ignores - returns a Manager for ignore rules set on the project
  • vulnerabilities - returns a list of snyk.models.Vulnerability objects with information about vulnerabilities in this project
  • jira_issues - returns a Manager with access to any associated Jira issues
  • licenses - returns a Manager for licenses currently in use by this project
  • settings - returns a Manager for interacting with the current project settings
  • tags - returns a Manager for interacting with the current project tags

You can add and delete tags using the manager:

  • tags.add(key, value) - adds a tag with the provided key/value pair to the project
  • tags.delete(key, value) - deletes a tag with the provided key/value pair from the project

In the case of Projects, as well as filtering by properties (as mentioned above) you can also filter by tag:

client.organizations.first().projects.filter(tags = [{"key": "some-key", "value": "some-value"}])

Note that the settings Manager can also be used to update settings like so, assuming you have a snyk.models.Project object in the variable project.

project.settings.update(pull_request_test_enabled=True)

Importing new projects

The client supports a high-level import_project method on organizations for adding new projects to be monitored by Snyk.

org = client.organizations.first()
org.import_project("github.com/user/project@branch")
org.import_project("docker.io/repository:tag")

If you are targetting a specific manifest file or files you can pass those as an optional argument, for instance:

org.import_project("github.com/user/project@branch", files=["Gemfile.lock"])

This method currently only supports importing projects from GitHub and Docker Hub. For other integrations you will need to grab the lower-level snyk.models.Integration object from the snyk.models.Organization.integrations manager noted above. Other services will be added to this API soon.

Testing for vulnerabilities

The API also exposes methods to discover vulnerability information about individual packages. These methods are found on the Organization object.

  • test_maven(<package_group_id>, <package_artifact_id>, <version>) - returns an IssueSet containing vulnerability information for a Maven artifact
  • test_rubygem(<name>, <version>) - returns an IssueSet containing vulnerability information for a Ruby Gem
  • test_python(<name>, <version>) - returns an IssueSet containing vulnerability information for Python package from PyPi
  • test_npm(<name>, <version>) - returns an IssueSet containing vulnerability information for an NPM package

Here's an example of checking a particular Python package.

>>> org = client.organizations.first()
>>> result = org.test_python("flask", "0.12.2")
>>> assert result.ok
False
# You can access details of the vulnerabilities too, for example
>>> result.issues.vulnerabilities[0].title
'Improper Input Validation'
>>> result.issues.vulnerabilities[0].identifiers
{'CVE': ['CVE-2018-1000656'], 'CWE': ['CWE-20']

As well as testing individual packages you can also test all packages found in various dependency management manifests. The client currently supports the following methods:

  • test_pipfile(<file-handle-or-string>) - returns an IssueSet for all Python dependencies in a Pipfile
  • test_gemfilelock(<file-handle-or-string>) - returns an IssueSet for all Ruby dependencies in a Gemfile
  • test_packagejson(<file-handle-or-string>, (<lock-file-handle-or-string>)) - returns an IssueSet for all Javascript dependencies in a package.json file. Optionally takes a package.lock file
  • test_gradlefile(<file-handle-or-string>) - returns an IssueSet for all dependencies in a Gradlefile
  • test_sbt(<file-handle-or-string>) - returns an IssueSet for all dependencies defined in a .sbt file
  • test_pom(<file-handle-or-string>) - returns an IssueSet for all dependencies in a Maven pom.xml file
  • test_yarn(<file-handle-or-string>, <lock-file-handle-or-string>) - returns an IssueSet for all dependencies in Yarn package.json and yarn.lock files
  • test_composer(<file-handle-or-string>, <lock-file-handle-or-string>) - returns an IssueSet for all dependencies in Composer composer.json and composer.lock files

For example, here we are testing a Python Pipfile.

>>> org = client.organizations.first()
>>> file = open("Pipfile")
>>> org.test_pipfile(file)

Inviting new users

You can invite new users to the organization via the API.

>>> org = client.organizations.first()
>>> org.invite("[email protected]")

You can also invite new users as administrators:

>>> org = client.organizations.first()
>>> org.invite("[email protected]", admin=True)

Low-level client

As well as the high-level API of the Snyk client you can use the HTTP methods directly. For these you simply need to pass the path, and optionally a data payload. The full domain, and the authentication details, are already provided by the client.

client.get("<path>")
client.delete("<path>")
client.put("<path>", <data>)
client.post("<path>", <data>)

Most of the time you shouldn't need to use these. They are mainly useful if new methods are added to the API which are not yet supported in the client. This can also be useful if you want to pass very specific parameters, or to parse the raw JSON output from the API.

Experimental rest low-level client

pysnyk >= 0.9.0 now includes support for basic rest (formerly referred to as v3) compatibility. To switch to use a rest client, pass the rest API url and version when initializing a client. Right now it supports the GET method. Refer to the rest API docs for more information and examples.

Getting the rest information of an organization:

# To get this value, get it from a Snyk organizations settings page
snyk_org = "df734bed-d75c-4f11-bb47-1d119913bcc7"

# to use the rest endpoint you MUST include a version value and the url of the v3 api endpoint as shown below
rest_client = SnykClient(snyk_token, version="2022-02-16~experimental", url="https://api.snyk.io/rest")

print(rest_client.get(f"/orgs/{snyk_org}").json())

# this supports overriding rest versions for a specific GET requests:
user = rest_client.get(f"orgs/{snyk_org}/users/{snyk_user}", version="2022-02-01~experimental").json()

# pass parameters such as how many results per page
params = {"limit": 10}

targets = rest_client.get(f"orgs/{snyk_org}/targets", params=params)

V1 and rest can work at the same time by instantiating two clients:

snyk_org = "df734bed-d75c-4f11-bb47-1d119913bcc7"

v1client = SnykClient(snyk_token)

rest_client = SnykClient(snyk_token, version="2022-02-16~experimental", url="https://api.snyk.io/rest")

v1_org = v1client.organizations.get(snyk_org)

rest_org = rest_client.get(f"/orgs/{snyk_org}").json()

The rest API introduces consistent pagination across all endpoints. The v3 client includes a helper method .get_rest_pages which collects the paginated responses and returns a single list combining the contents of the "data" key from all pages. It takes the same values as the get method.

rest_client = SnykClient(snyk_token, version="2022-02-16~experimental", url="https://api.snyk.io/rest")

params = {"limit": 10}

targets = rest_client.get(f"orgs/{snyk_org}/targets", params=params).json()

print(len(targets["data"]))
# returns 10 targets

all_targets = rest_client.get_rest_pages(f"orgs/{snyk_org}/targets", params=params)

print(len(all_targets))
# returns 33 targets, note we don't have to add .json() to the call or access the "data" key, get_rest_pages does that for us

# Override Content-Type header (or others) as needed for 
response = rest_client.post(f"orgs/{snyk_org}/invites", body={"email": "[email protected]", "role": "f8223479-01a9-4774-8e29-06ce9a0513d6", headers={"Content-Type": "application/vnd.api+json"}})

For backwards compatibility the get_rest_pages method has an alternative name of get_v3_pages to not break code already rewritten replatformed to the 0.9.0 pysnyk module.