Process Credentials for GCP Client Library - python
Google Cloud Credential provider which allows sourcing credentials from an external process.
Essentially, its a credential source which allows the delegation of acquiring GCP access_tokens to arbitrary binaries you have access to at runtime.
The arbitrary binary would use whatever means it has available (kerberos, ldap, saml-cli, etc) to get a GCP access_token.
From there, the token is given surfaced as a refreshable credential source you can directly use with a GCP library.
This is similar to several systems that provide such delegation.
- Kubernetes kubectl credential plugin
- AWS Process Credentials
Needless to say, use this after very careful consideration: this library will attempt to execute a binary on the system where its run (ofcource the process running using the library would need access to run the binary anyway)
NOTE these samples are NOT supported by google; its just something done on a weekend...
Implementations
As its a weekend project, caveat emptor. The code is alpha quality and I didn't have time to push it to maven central, npm, etc.
If you want it there, please review the code, provide suggestions and improvements
golang: https://github.com/salrashid123/gcp_process_credentials_gopython: https://github.com/salrashid123/gcp_process_credentials_pyjava: https://github.com/salrashid123/gcp_process_credentials_javanode: https://github.com/salrashid123/gcp_process_credentials_node
See the "examples" folder in each
Binary Response Contract
Each library above will invoke a binary, pass it some args and env var.
The response back from the binary must
be valid JSON in the form
{
"access_token": "ya29....",
"expires_in": 3600,
"token_type": "Bearer"
}
access_token: your access tokenexpires_in: how many seconds this token is valid fortoken_type: usually just a bearer token
Quickstart
For a quick example in python, the following will read a token file and use that for credentials:
pip install google-auth gcp-process-credentials
from gcp_process_credentials.credentials import ProcessCredentials
e = os.environ.copy()
e.update({'foo': 'bar'})
pc = ProcessCredentials(command=["/usr/bin/cat"], args=['/tmp/token.txt'], env=e,)
storage_client = storage.Client(project=project_id, credentials=pc)
buckets = storage_client.list_buckets()
for bkt in buckets:
print(bkt.name)
ofcourse the file here /tmp/token.txt must be the json file format described above
Parser Interface
If your binary does not provide the exact json format, your can define a parser interface to 'translate' the credential for you.
For example, gcloud auth print-access-token returns just the access token with an annoying newline character from stdout.
You an provide an interface to do the translation like this:
def gcloud_parser(req):
data = {}
data["access_token"] = ''.join(req.decode().split('\n'))
data["expires_in"] = 3600
data["token_type"] = "Bearer"
return str(json.dumps(data))
pc = ProcessCredentials(command=["gcloud"], args=['auth', 'print-access-token'], env=e, parser=gcloud_parser)
Injecting tokens vs Wrapped Credentials
You might be asking...why cant' i just run the binary on my own in code, get the token an inject it as a credential like this??
import google.oauth2.credentials
credentials = google.oauth2.credentials.Credentials("thetoken")
storage_client = storage.Client(credentials=credentials)
for b in storage_client.list_buckets():
print(b.name)
Well, the, token is not refreshable and your client library will need to manage that. On the other hand, if you use this library, it will automatically refresh the token by calling the binary when its nearing expiration
Other References AWS->GCP Process Credential Plugin
to generate the library from scratch and run local, run
python3 setup.py sdist bdist_wheel
cd example
virtualenv env
pip3 install ../
pip3 install -r requirements.txt
Metadata
Release files for gcp-process-credentials 0.0.2
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| gcp_process_credentials-0.0.2.tar.gz | 8.1 kB | Without provenance | Details for gcp_process_credentials-0.0.2.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| gcp_process_credentials-0.0.2-py3-none-any.whl | Python 3 | none | any | Details for gcp_process_credentials-0.0.2-py3-none-any.whl |
Total release size: 16.7 kB
Release files / gcp_process_credentials-0.0.2.tar.gz
| Download URL | gcp_process_credentials-0.0.2.tar.gz |
|---|---|
| Size | 8.1 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
4156e8af57330683eb483b8d98e59c2d5861b7dabeb83e112c241997f95a64c8
|
|
BLAKE2b-256 checksum How to use checksums |
a8193230056223b28acf6c36ddabaf0e737b5c709698bfd9739962f397497b89
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/4.0.2 CPython/3.11.4
|
Release files / gcp_process_credentials-0.0.2-py3-none-any.whl
| Download URL | gcp_process_credentials-0.0.2-py3-none-any.whl |
|---|---|
| Size | 8.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
6e08a26273bfedc9ac213b33b4c367d7fc203dbc9a60ea5d04d58ece62cf2e82
|
|
BLAKE2b-256 checksum How to use checksums |
6960ffbae6d61d3d8e3d5342fc104f6503dc22ca1f92fad1e7f2d0ea74f05f0a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/4.0.2 CPython/3.11.4
|
Security for gcp-process-credentials 0.0.2
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.