SPECIES_MAPPING.md

Species to Parquet File Mapping

This document maps each parquet shard file to its corresponding species and image count.

Dataset Information

• Total Shards: 60
• Mammals: 30 species
• Aves (Birds): 30 species

---

Mammals Species Mapping

Parquet File	Species Name	Image Count
mammals-00001-of-00030.parquet	Alpaca	1,864
mammals-00002-of-00030.parquet	American Black Bear	2,912
mammals-00003-of-00030.parquet	Black-backed Jackal	2,067
mammals-00004-of-00030.parquet	Black Rhinoceros	1,428
mammals-00005-of-00030.parquet	Bushpig	1,241
mammals-00006-of-00030.parquet	Common Brushtail Possum	1,365
mammals-00007-of-00030.parquet	Crab-eating Mongoose	2,364
mammals-00008-of-00030.parquet	Crested Porcupine	1,449
mammals-00009-of-00030.parquet	Dromedary Camel	1,809
mammals-00010-of-00030.parquet	Eastern Gray Squrril	2,434
mammals-00011-of-00030.parquet	Ferret Badger	1,743
mammals-00012-of-00030.parquet	Gemsbok	3,963
mammals-00013-of-00030.parquet	Gigant Armadillo	463
mammals-00014-of-00030.parquet	Giraff	2,985
mammals-00015-of-00030.parquet	Greater Kudu	4,410
mammals-00016-of-00030.parquet	Hippotamus	2,319
mammals-00017-of-00030.parquet	Jaguar	6,855
mammals-00018-of-00030.parquet	Least Weasel	1,869
mammals-00019-of-00030.parquet	L´Hoest´s Monkey	2,941
mammals-00020-of-00030.parquet	Northern Treeshrew	699
mammals-00021-of-00030.parquet	NZ Sea lion	2,296
mammals-00022-of-00030.parquet	Raccoon	2,700
mammals-00023-of-00030.parquet	Serval	625
mammals-00024-of-00030.parquet	Ship Rat	822
mammals-00025-of-00030.parquet	Spotted Paca	1,715
mammals-00026-of-00030.parquet	Stump-tailed Macaque	3,558
mammals-00027-of-00030.parquet	Sun Bear	713
mammals-00028-of-00030.parquet	Wartog	1,671
mammals-00029-of-00030.parquet	White Nosed Coatis	1,996
mammals-00030-of-00030.parquet	Wolf	2,307

---

Aves (Birds) Species Mapping

Parquet File	Species Name	Image Count
aves-00001-of-00030.parquet	American Crow	1,240
aves-00002-of-00030.parquet	Australasian Swamphen	2,772
aves-00003-of-00030.parquet	Australian Magpie	2,739
aves-00004-of-00030.parquet	Black Curassow	1,949
aves-00005-of-00030.parquet	Blue Whistling Thrush	2,157
aves-00006-of-00030.parquet	Brown Quail	2,662
aves-00007-of-00030.parquet	Chicken	1,647
aves-00008-of-00030.parquet	Common chafflinch	2,384
aves-00009-of-00030.parquet	Common Myna	2,707
aves-00010-of-00030.parquet	Dunnock	2,066
aves-00011-of-00030.parquet	European Starling	1,876
aves-00012-of-00030.parquet	Fantails	2,127
aves-00013-of-00030.parquet	Greenfinch	2,503
aves-00014-of-00030.parquet	Kea	2,833
aves-00015-of-00030.parquet	Kiwi	1,634
aves-00016-of-00030.parquet	Kori Bustard	1,485
aves-00017-of-00030.parquet	Mountain Quail	2,381
aves-00018-of-00030.parquet	New Zealand Robin	2,753
aves-00019-of-00030.parquet	Ostrich	3,727
aves-00020-of-00030.parquet	Petrel	2,370
aves-00021-of-00030.parquet	Pipit	2,316
aves-00022-of-00030.parquet	Red Junglefowls	2,684
aves-00023-of-00030.parquet	Spix's Guan	3,032
aves-00024-of-00030.parquet	Swamp Harrier	2,917
aves-00025-of-00030.parquet	Takahe	2,085
aves-00026-of-00030.parquet	Tui	2,767
aves-00027-of-00030.parquet	Vulturine Guinea Fowl	4,420
aves-00028-of-00030.parquet	Weka	2,421
aves-00029-of-00030.parquet	Wild Turkey	2,057
aves-00030-of-00030.parquet	Yellow-eyed Penguin	2,817

---

Summary Statistics

• Total Mammals Images: 65,583
• Total Aves Images: 73,528
• Grand Total: 139,111

---

Usage

Each parquet file contains:

• species_name: Folder name for the species
• filename: Original image filename
• image_bytes: Raw image data (JPG format)
• original_url: Source URL from original dataset
• dataset_name: Name of source dataset
• bbox: Bounding box coordinates [x, y, width, height]
• category: Animal category/classification
• source_image: Original source image filename

Loading a Species

from huggingface_hub import hf_hub_download
import pyarrow.parquet as pq

# Download specific shard
parquet_file = hf_hub_download(
    repo_id="AI-EcoNet/ViT_animal_data",
    filename="data/mammals-00001-of-00030.parquet",
    repo_type="dataset"
)

# Read parquet file
table = pq.read_table(parquet_file)
df = table.to_pandas()

# Extract images
for idx, row in df.iterrows():
    with open(f"output/{row['filename']}", 'wb') as f:
        f.write(row['image_bytes'])

---

Generated on: 2025-11-14