Convenient filesystem interface over Oracle Cloud's Object Storage

Oracle, Cloud, Infrastructure, OCI, Object, Storage, filesystem, fsspec, object-storage, python3
pip install ocifs==1.3.1


Oracle Cloud Infrastructure Object Storage fsspec Implementation

PyPI Python

​ The Oracle Cloud Infrastructure Object Storage service is an internet-scale, high-performance storage platform that offers reliable and cost-efficient data durability. With Object Storage, you can safely and securely store or retrieve data directly from the internet or from within the cloud platform. ​ ocifs is part of the fsspec intake/filesystem_spec ecosystem

a template or specification for a file-system interface, that specific implementations should follow, so that applications making use of them can rely on a common interface and not have to worry about the specific internal implementation decisions with any given backend. ​ ocifs joins the list of file systems supported with this package. ​ The intake/filesystem_spec project is used by Pandas, Dask and other data libraries in python, this package adds Oracle OCI Object Storage capabilties to these libraries. ​

OCIFS file system style operations Example:

from ocifs import OCIFileSystem

fs = OCIFilesystem("~/.oci/config")
# 1.Create empty file or truncate in OCI objectstorage bucket
 fs.touch("oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt", truncate=True, data=b"Writing to Object Storage!")
 # 2.Fetch(potentially multiple paths' contents"oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt")
 # 3.Get metadata about a file from a head or list call"oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt")
 # 4.Get directory listing page"oci://<my_bucket>@<my_namespace>/<my_prefix>/", detail=True)
 # 5.Is this entry directory-like?
 # 6.Is this entry file-like?
 # 7.If there is a file at the given path (including broken links)
 # 8.List of files for the given path
 fs.listdir("oci://<my_bucket>@<my_namespace>/<my_prefix>", detail=True)
 # 9.Get the first ``size`` bytes from file
 fs.head("oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt", size=1024)
 # 10.Get the last ``size`` bytes from file
 fs.tail("oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt", size=1024)
 # 11.Hash of file properties, to tell if it has changed
 # 12.Size in bytes of file
 # 13.Size in bytes of each file in a list of paths
 paths = ["oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt"]
 # 14.Normalise OCI path string into bucket and key.
 # 15.Delete  a file from the  bucket
 # 16.Get the contents of the file as a byte
 fs.read_bytes("oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt", start=0, end=13)
 # 17.Get the contents of the file as a string
 fs.read_text("oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt", encoding=None, errors=None, newline=None)
 # 18.Get the contents of the file as a byte
 fs.read_block("oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt", 0, 13)
 # 19.Open a file for writing/flushing into file in OCI objectstorage bucket
 # Ocifs sets the best-guessed content-type for hello.txt i.e "text/plain"
 with"oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt", 'w', autocommit=True) as f:
        f.write("Writing data to buffer, before manually flushing and closing.") # data is flushed and file closed
# Ocifs uses the specified content-type passed in the open while writing to OCI objectstorage bucket
 with"oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt", 'w',content_type='text/plain') as f:
        f.write("Writing data to buffer, before manually flushing and closing.") # data is flushed and file closed
 # 20.Open a file for reading a file from OCI objectstorage bucket
 with"oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt") as f:
 # 21.Space used by files and optionally directories within a path
 # 22.Find files by glob-matching.
 # 23.Renames an object in a particular bucket in tenancy namespace on OCI
 fs.rename("oci://<my_bucket>@<my_namespace>/<my_prefix>/hello.txt", "oci://<my_bucket>@<my_namespace>/<my_prefix>/hello2.txt")
 # 24.Delete multiple files from the same bucket
 pathlist = ["oci://<my_bucket>@<my_namespace>/<my_prefix>/hello2.txt"]

Or Use With Pandas

import pandas as pd
import ocifsdf = pd.read_csv(
    storage_options={"config": "~/.oci/config"},

Or Use With PyArrow

import pandas as pd
import ocifsdf = pd.read_csv(
    "oci://my_bucket@my_namespace/my_object.csv",storage_options={"config": "~/.oci/config"})

Or Use With ADSDataset

import ads
import pandas as pd
from ads.common.auth import default_signer
from ads.dataset.dataset import ADSDatasetads.set_auth(auth="api_key", oci_config_location="~/.oci/config", profile="<profile_name>")
    ds = ADSDataset(
        df=pd.read_csv(f"oci://my_bucket@my_namespace/my_object.csv", storage_options=default_signer()),

Getting Started

python3 -m pip install ocifs

Software Prerequisites

Python >= 3.6

Environment Variables for Authentication:

export OCIFS_IAM_TYPE=api_key
export OCIFS_CONFIG_LOCATION=~/.oci/config

Note, if you are operating on OCI with an alternative valid signer, such as resource principal, instead set the following:

export OCIFS_IAM_TYPE=resource_principal

Environment Variables for enabling Logging:

To quickly see all messages, you can set the environment variable OCIFS_LOGGING_LEVEL=DEBUG.




The built-in filesystems in fsspec are maintained by the intake project team, where as ocifs is an external implementation (similar to s3fs, gcsfs, adl/abfs, and so on), which is maintained by Oracle.


This project welcomes contributions from the community. Before submitting a pull request, please review our contribution guide


Please consult the security guide for our responsible security vulnerability disclosure process


Copyright (c) 2021, 2023 Oracle and/or its affiliates.

Released under the Universal Permissive License v1.0 as shown at