[](https://pypi.org/project/geococo/)
[](https://github.com/jaspersiebring/GeoCOCO/actions/workflows/main.yml)
[](https://geococo.readthedocs.io/en/latest/?badge=latest)
Easily transform your GIS annotations into [Microsoft's Common Objects In Context (COCO)](https://cocodataset.org/#format-data) datasets with GeoCOCO. This tool allows users to leverage the advanced digitizing solutions of modern GIS software for the annotations of image objects in geographic imagery.
Built with `Pydantic` and `pycocotools`, it features a complete implementation of the COCO standard for object detection with out-of-the-box support for JSON-encoding and RLE compression. The resulting datasets are versioned, easily extendable with new annotations and fully compatible with other data applications that accept the COCO format.
## Key features
- User-friendly: GeoCOCO is designed for ease of use, requiring minimal configuration and domain knowledge
- Version Control: Datasets created with GeoCOCO are versioned and designed for expansion with future annotations
- Command-line Tool: Use GeoCOCO from your terminal to create, append and copy COCO datasets
- Python Module: Integrate GeoCOCO in your own data applications with the `geococo` package
- Representation: GeoCOCO maximizes label representation through an adaptive moving window approach
- COCO Standard: Output datasets are fully compatible with other COCO-accepting applications
- Compact File Size: JSON-encoding and RLE compression are employed to ensure compact file sizes
## Installation
Installing from the Python Package Index (PyPI):
````
# Install from PYPI with Python's package installer (pip)
pip install geococo
````
## Example of usage
After installing `geococo`, there are a number of ways you can interact with its API.
#### Command line interface
The easiest way to use `geococo` is to simply call it from your preferred terminal with one of three commands: `new`, `add` and `copy`.
````
$ geococo --help
Usage: geococo [OPTIONS] COMMAND [ARGS]...
Transform your GIS annotations into COCO datasets.
Options:
--help Show this message and exit.
Commands:
add Transform and add GIS annotations to an existing CocoDataset
copy Copy and (optionally) update the metadata of an existing CocoDataset
new Initialize a new CocoDataset with user-prompted metadata
````
Starting a new CocoDataset (will prompt user for metadata like Description and Contributor)
````
$ geococo new dataset.json
````
Adding new annotations to existing CocoDataset (will increment version based on input data and update any new categories/images)
````
$ geococo add image.tif labels.shp dataset.json images/ 512 512 --id-attribute ids
````
For more information on the different commands, call `geococo COMMAND` with `--help`.
````
$ geococo add --help
Usage: geococo add [OPTIONS] IMAGE_PATH LABELS_PATH JSON_PATH OUTPUT_DIR WIDTH
HEIGHT
Transform and add GIS annotations to an existing COCO dataset.
This method generates a COCO dataset by moving across the given image
(image_path) with a moving window (width, height), constantly checking for
intersecting annotations (labels_path) that represent image objects in said
image (e.g. buildings in satellite imagery; denoted by (super)category name
and/or id). Each valid intersection will add n Annotations entries to the
dataset (json_path) and save a subset of the input image that contained
these entries (output_dir).
The output data size depends on your input labels, as the moving window
adjusts its step size to accommodate the average annotation size, optimizing
dataset representation and minimizing tool configuration. Each addition will
also increment the dataset version: patch if using the same image_path,
minor if using a new image_path, and major if using a new output_dir.
Arguments:
IMAGE_PATH Path to geospatial image containing image objects [required]
LABELS_PATH Path to vector file containing annotated image objects
[required]
JSON_PATH Path to json file containing the COCO dataset [required]
OUTPUT_DIR Path to output directory for image subsets [required]
WIDTH Width of image subsets [required]
HEIGHT Height of image subsets [required]
Options:
--id-attribute TEXT Name of column containing category_id values
(optional if --name_attribute is given)
--name-attribute TEXT Name of column containing category_name values
(optional if --id_attribute is given)
--super-attribute TEXT Name of column containing supercategory values
--help Show this message and exit.
````
#### Python module
This is recommended for most developers as it gives you more granular control over the various steps. It does assume a basic familiarity with the `geopandas` and `rasterio` packages (i.e. GIS modules that help you manage vector and raster data respectively).
````
import pathlib
import geopandas as gpd
import rasterio
from geococo import create_dataset, load_dataset, save_dataset, append_dataset
# Replace this with your preferred output paths
data_path = pathlib.Path("path/to/your/coco/output/images")
json_path = pathlib.Path("path/to/your/coco/json/file")
# Dimensions of the moving window and output images
width, height = 512, 512
# Starting a new CocoDataset
description = "My First Dataset"
contributor = "User"
# version and date_created are automatically set
dataset = create_dataset(description=description, contributor=contributor)
# You can also load existing COCO datasets
# dataset = load_dataset(json_path=json_path)
# Loading GIS data with rasterio and geopandas
labels = gpd.read_file(some_labels_path)
raster_source = rasterio.open(some_image_path)
# (Optional) Apply any spatial or attribute queries here
# labels = labels.loc[labels["ids"].isin([1, 2, 3])]
# labels = labels.loc[labels.within(some_polygon)]
# Find and save all Annotation instances
dataset = append_dataset(
dataset=dataset,
images_dir=data_path,
src=raster_source,
labels=labels,
window_bounds=[(width, height)],
id_attribute=None, # column with category_id values
name_attribute="ids", # column with category_name values
super_attribute=None, # optional column with super_category values
)
# Encode CocoDataset instance as JSON and save to json_path
save_dataset(dataset=dataset, json_path=json_path)
````
# Visualization with FiftyOne
Like the official COCO project, the open source tool [FiftyOne](https://docs.voxel51.com/) can be used to visualize and evaluate your datasets. To do this, you'll need the `fiftyone` and `pycocotools` packages. Note that `geococo` does not install `fiftyone` by default, so you'll need to install it separately (instructions for installation can be found [here](https://docs.voxel51.com/getting_started/install.html)). Once you have `fiftyone` installed, you can use the following command to inspect your COCO dataset in your web browser.
````
# requires pycocotools and fiftyone
import fiftyone as fo
import fiftyone.zoo as foz
import pathlib
data_path = pathlib.Path("path/to/your/coco/output/images")
json_path = pathlib.Path("path/to/your/coco/json/file")
# Load COCO formatted dataset
coco_dataset = fo.Dataset.from_dir(
dataset_type=fo.types.COCODetectionDataset,
data_path=data_path,
labels_path=json_path,
include_id=True,
)
# Launch the app
session = fo.launch_app(coco_dataset, port=5151)
````
<p float="left">
<img src="https://github.com/jaspersiebring/GeoCOCO/assets/25051531/f8ab55da-b3cd-4beb-b082-7946e712ea5c" width="45%" height = 250/>
<img src="https://github.com/jaspersiebring/GeoCOCO/assets/25051531/9a796a54-ffc2-49c3-95bc-59e5c0dd1d7c" width="45%" height = 250 />
</p>
Raw data
{
"_id": null,
"home_page": "https://geococo.readthedocs.io/",
"name": "geococo",
"maintainer": null,
"docs_url": null,
"requires_python": "<4.0,>=3.9",
"maintainer_email": null,
"keywords": "coco, annotation, deep-learning, gis, qgis, arcgis, labeling, ai, yolo",
"author": "Jasper Siebring",
"author_email": "j.siebring92@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/f0/af/3bf0cd399893147fbd328563a76e65155e5c67a982cdec3bc8a86abf2163/geococo-0.5.5.tar.gz",
"platform": null,
"description": "\n\n[](https://pypi.org/project/geococo/)\n\n[](https://github.com/jaspersiebring/GeoCOCO/actions/workflows/main.yml)\n[](https://geococo.readthedocs.io/en/latest/?badge=latest)\n\nEasily transform your GIS annotations into [Microsoft's Common Objects In Context (COCO)](https://cocodataset.org/#format-data) datasets with GeoCOCO. This tool allows users to leverage the advanced digitizing solutions of modern GIS software for the annotations of image objects in geographic imagery.\n \nBuilt with `Pydantic` and `pycocotools`, it features a complete implementation of the COCO standard for object detection with out-of-the-box support for JSON-encoding and RLE compression. The resulting datasets are versioned, easily extendable with new annotations and fully compatible with other data applications that accept the COCO format.\n\n## Key features\n- User-friendly: GeoCOCO is designed for ease of use, requiring minimal configuration and domain knowledge\n- Version Control: Datasets created with GeoCOCO are versioned and designed for expansion with future annotations\n- Command-line Tool: Use GeoCOCO from your terminal to create, append and copy COCO datasets\n- Python Module: Integrate GeoCOCO in your own data applications with the `geococo` package\n- Representation: GeoCOCO maximizes label representation through an adaptive moving window approach\n- COCO Standard: Output datasets are fully compatible with other COCO-accepting applications\n- Compact File Size: JSON-encoding and RLE compression are employed to ensure compact file sizes\n\n## Installation\nInstalling from the Python Package Index (PyPI):\n````\n# Install from PYPI with Python's package installer (pip)\npip install geococo\n````\n\n## Example of usage\n\nAfter installing `geococo`, there are a number of ways you can interact with its API.\n\n#### Command line interface\n The easiest way to use `geococo` is to simply call it from your preferred terminal with one of three commands: `new`, `add` and `copy`.\n\n ````\n$ geococo --help\nUsage: geococo [OPTIONS] COMMAND [ARGS]...\n\n Transform your GIS annotations into COCO datasets.\n\nOptions:\n --help Show this message and exit.\n\nCommands:\n add Transform and add GIS annotations to an existing CocoDataset\n copy Copy and (optionally) update the metadata of an existing CocoDataset\n new Initialize a new CocoDataset with user-prompted metadata\n````\n\nStarting a new CocoDataset (will prompt user for metadata like Description and Contributor)\n\n````\n$ geococo new dataset.json\n````\n\nAdding new annotations to existing CocoDataset (will increment version based on input data and update any new categories/images)\n````\n$ geococo add image.tif labels.shp dataset.json images/ 512 512 --id-attribute ids\n````\n\nFor more information on the different commands, call `geococo COMMAND` with `--help`.\n````\n$ geococo add --help\nUsage: geococo add [OPTIONS] IMAGE_PATH LABELS_PATH JSON_PATH OUTPUT_DIR WIDTH\n HEIGHT\n\n Transform and add GIS annotations to an existing COCO dataset.\n\n This method generates a COCO dataset by moving across the given image\n (image_path) with a moving window (width, height), constantly checking for\n intersecting annotations (labels_path) that represent image objects in said\n image (e.g. buildings in satellite imagery; denoted by (super)category name\n and/or id). Each valid intersection will add n Annotations entries to the\n dataset (json_path) and save a subset of the input image that contained\n these entries (output_dir).\n\n The output data size depends on your input labels, as the moving window\n adjusts its step size to accommodate the average annotation size, optimizing\n dataset representation and minimizing tool configuration. Each addition will\n also increment the dataset version: patch if using the same image_path,\n minor if using a new image_path, and major if using a new output_dir.\n\nArguments:\n IMAGE_PATH Path to geospatial image containing image objects [required]\n LABELS_PATH Path to vector file containing annotated image objects\n [required]\n JSON_PATH Path to json file containing the COCO dataset [required]\n OUTPUT_DIR Path to output directory for image subsets [required]\n WIDTH Width of image subsets [required]\n HEIGHT Height of image subsets [required]\n\nOptions:\n --id-attribute TEXT Name of column containing category_id values\n (optional if --name_attribute is given)\n --name-attribute TEXT Name of column containing category_name values\n (optional if --id_attribute is given)\n --super-attribute TEXT Name of column containing supercategory values\n --help Show this message and exit.\n````\n\n\n#### Python module\nThis is recommended for most developers as it gives you more granular control over the various steps. It does assume a basic familiarity with the `geopandas` and `rasterio` packages (i.e. GIS modules that help you manage vector and raster data respectively).\n\n````\nimport pathlib\nimport geopandas as gpd\nimport rasterio\nfrom geococo import create_dataset, load_dataset, save_dataset, append_dataset\n\n# Replace this with your preferred output paths\ndata_path = pathlib.Path(\"path/to/your/coco/output/images\")\njson_path = pathlib.Path(\"path/to/your/coco/json/file\")\n\n# Dimensions of the moving window and output images\nwidth, height = 512, 512\n\n# Starting a new CocoDataset\ndescription = \"My First Dataset\"\ncontributor = \"User\"\n\n# version and date_created are automatically set\ndataset = create_dataset(description=description, contributor=contributor)\n# You can also load existing COCO datasets\n# dataset = load_dataset(json_path=json_path)\n\n# Loading GIS data with rasterio and geopandas\nlabels = gpd.read_file(some_labels_path)\nraster_source = rasterio.open(some_image_path)\n\n# (Optional) Apply any spatial or attribute queries here\n# labels = labels.loc[labels[\"ids\"].isin([1, 2, 3])]\n# labels = labels.loc[labels.within(some_polygon)]\n\n# Find and save all Annotation instances\ndataset = append_dataset(\n dataset=dataset,\n images_dir=data_path,\n src=raster_source,\n labels=labels,\n window_bounds=[(width, height)],\n id_attribute=None, # column with category_id values\n name_attribute=\"ids\", # column with category_name values\n super_attribute=None, # optional column with super_category values\n)\n\n# Encode CocoDataset instance as JSON and save to json_path\nsave_dataset(dataset=dataset, json_path=json_path)\n````\n\n# Visualization with FiftyOne\nLike the official COCO project, the open source tool [FiftyOne](https://docs.voxel51.com/) can be used to visualize and evaluate your datasets. To do this, you'll need the `fiftyone` and `pycocotools` packages. Note that `geococo` does not install `fiftyone` by default, so you'll need to install it separately (instructions for installation can be found [here](https://docs.voxel51.com/getting_started/install.html)). Once you have `fiftyone` installed, you can use the following command to inspect your COCO dataset in your web browser.\n\n````\n# requires pycocotools and fiftyone\nimport fiftyone as fo\nimport fiftyone.zoo as foz\nimport pathlib\n\ndata_path = pathlib.Path(\"path/to/your/coco/output/images\")\njson_path = pathlib.Path(\"path/to/your/coco/json/file\")\n\n# Load COCO formatted dataset\ncoco_dataset = fo.Dataset.from_dir(\n dataset_type=fo.types.COCODetectionDataset,\n data_path=data_path,\n labels_path=json_path,\n include_id=True,\n)\n\n# Launch the app\nsession = fo.launch_app(coco_dataset, port=5151)\n````\n\n<p float=\"left\">\n <img src=\"https://github.com/jaspersiebring/GeoCOCO/assets/25051531/f8ab55da-b3cd-4beb-b082-7946e712ea5c\" width=\"45%\" height = 250/>\n <img src=\"https://github.com/jaspersiebring/GeoCOCO/assets/25051531/9a796a54-ffc2-49c3-95bc-59e5c0dd1d7c\" width=\"45%\" height = 250 />\n</p>\n",
"bugtrack_url": null,
"license": null,
"summary": "Converts GIS annotations to Microsoft's Common Objects In Context (COCO) dataset format",
"version": "0.5.5",
"project_urls": {
"Homepage": "https://geococo.readthedocs.io/",
"Repository": "https://github.com/jaspersiebring/GeoCOCO"
},
"split_keywords": [
"coco",
" annotation",
" deep-learning",
" gis",
" qgis",
" arcgis",
" labeling",
" ai",
" yolo"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "f71b87e2a672933abc910559fa8a46a74135f89adde5a1a7db995ebf4d4f2f92",
"md5": "aa2ad90627c6108992ef6bdfd8baef4e",
"sha256": "5fd8dc35571ea602188dc63a8f16d51c9208947f736e6db059dc291de30f434a"
},
"downloads": -1,
"filename": "geococo-0.5.5-py3-none-any.whl",
"has_sig": false,
"md5_digest": "aa2ad90627c6108992ef6bdfd8baef4e",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.9",
"size": 30110,
"upload_time": "2024-07-20T17:05:07",
"upload_time_iso_8601": "2024-07-20T17:05:07.183650Z",
"url": "https://files.pythonhosted.org/packages/f7/1b/87e2a672933abc910559fa8a46a74135f89adde5a1a7db995ebf4d4f2f92/geococo-0.5.5-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "f0af3bf0cd399893147fbd328563a76e65155e5c67a982cdec3bc8a86abf2163",
"md5": "d091cc6d676046a8571ffff4fdb5b0e5",
"sha256": "cf3c1b0de66d986cc66c8e9e008ecbebfb5c2cf712f9a574ae9b77110c71bbb7"
},
"downloads": -1,
"filename": "geococo-0.5.5.tar.gz",
"has_sig": false,
"md5_digest": "d091cc6d676046a8571ffff4fdb5b0e5",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.9",
"size": 29656,
"upload_time": "2024-07-20T17:05:08",
"upload_time_iso_8601": "2024-07-20T17:05:08.214505Z",
"url": "https://files.pythonhosted.org/packages/f0/af/3bf0cd399893147fbd328563a76e65155e5c67a982cdec3bc8a86abf2163/geococo-0.5.5.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-07-20 17:05:08",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "jaspersiebring",
"github_project": "GeoCOCO",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "geococo"
}
JSON
