Binary Files

Graviti SDK use the File and RemoteFile to represent a specific file.

In addition, SDK also provides several commonly used file formats, including Image, Audio and PointCloud.

More details about the Copy Files and Storage Config are as follows:

• Copy Files
• Storage Config

File

SDK supports all various binary files including video files and text files by File.

Load the local text files to DataFrame:

import graviti.portex as pt
from graviti import DataFrame, File

std = pt.build_package("https://github.com/Project-OpenBytes/portex-standard", "main")
schema = pt.record(
    {
        "filename": pt.string(),
        "file": std.file.File(),
    }
)

data = [
    {
        "filename": "EXAMPLE1.txt",
        "file": File("PATH/TO/YOUR/EXAMPLE1.txt")
    },
    {
        "filename": "EXAMPLE2.txt",
        "file": File("PATH/TO/YOUR/EXAMPLE2.txt")
    },
]

df = DataFrame(data, schema)

Read the text in DataFrame:

text = df["text"][0]
with text.open() as fp:
    fp.read().decode("utf-8")

For all binary files, SDK supports viewing their basic information, including extension, size and checksum:

text.extension
text.size
text.get_checksum()

Image

Load the local image into DataFrame:

import graviti.portex as pt
from graviti import DataFrame, Image

std = pt.build_package("https://github.com/Project-OpenBytes/portex-standard", "main")
schema = pt.record(
    {
        "filename": pt.string(),
        "image": std.file.Image(),
    }
)

data = [
    {
        "filename": "EXAMPLE1.png",
        "image": Image("PATH/TO/YOUR/EXAMPLE1.png")
    },
    {
        "filename": "EXAMPLE2.png",
        "image": Image("PATH/TO/YOUR/EXAMPLE2.png")
    },
]

df = DataFrame(data, schema)

Read the image in DataFrame:

import PIL

image = df["image"][0]
with image.open() as fp:
    PIL.Image.open(fp)

For image files, SDK supports viewing their height and width:

image.height
image.width

Audio

Load the local audio into DataFrame:

import graviti.portex as pt
from graviti import DataFrame, Audio

std = pt.build_package("https://github.com/Project-OpenBytes/portex-standard", "main")
schema = pt.record(
    {
        "filename": pt.string(),
        "audio": std.file.Audio(),
    }
)

data = [
    {
        "filename": "EXAMPLE1.wav",
        "audio": Audio("PATH/TO/YOUR/EXAMPLE1.wav")
    },
    {
        "filename": "EXAMPLE2.wav",
        "audio": Audio("PATH/TO/YOUR/EXAMPLE2.wav")
    },
]

df = DataFrame(data, schema)

Read the audio in DataFrame:

audio = df["audio"][0]
with audio.open() as fp:
    fp.read()

Point Cloud

Load the local point_cloud into DataFrame:

import graviti.portex as pt
from graviti import DataFrame, PointCloud

std = pt.build_package("https://github.com/Project-OpenBytes/portex-standard", "main")
schema = pt.record(
    {
        "filename": pt.string(),
        "point_cloud": std.file.PointCloud(),
    }
)

data = [
    {
        "filename": "EXAMPLE1.pcd",
        "point_cloud": PointCloud("PATH/TO/YOUR/EXAMPLE1.pcd")
    },
    {
        "filename": "EXAMPLE2.pcd",
        "point_cloud": PointCloud("PATH/TO/YOUR/EXAMPLE2.pcd")
    },
]

df = DataFrame(data, schema)

Read the point_cloud in DataFrame:

point_cloud = df["point_cloud"][0]
with point_cloud.open() as fp:
    fp.read()