image/transform.py
Resize, crop, convert, and compress images (PNG, JPEG, WebP, GIF, TIFF, BMP), reporting the byte savings on stdout. Set max_dimension to fit the longest edge inside a box while preserving aspect ratio, or give width and/or height directly; crop takes "x,y,w,h" in pixels and is applied before any resize. Shrinks oversized screenshots before they consume an agent's context or get embedded in a PDF.
Use when an image must be resized, cropped, converted to another format, or compressed to a smaller file size.
- the file
- image/transform.py
- its skill
- image/SKILL.md (the whole image pack)
- as knowledge
- okf/image/transform.md
- on GitHub
- image/transform.py ↗
Run it
1. Once per machine: install fragletc (Docker must already be running).
curl -fsSL https://raw.githubusercontent.com/ofthemachine/fraglet/main/install.sh | sh
2. Fetch the tool and print its contract.
curl -fsSLO https://tools.ofthemachine.com/image/transform.py && chmod +x transform.py && ./transform.py --fraglet-help
3. Run it.
./transform.py -p image="<image>" [-p max_dimension="<max_dimension>"] [-p width="<width>"] [-p height="<height>"] [-p crop="<crop>"] [-p format="<format>"] [-p quality=85] [-p grayscale=false] --output image
Parameters
| name | type | requirement | description |
|---|---|---|---|
| image | file | required | Source image mounted at /input/image |
| max_dimension | string | optional | Fit longest edge inside this many pixels |
| width | string | optional | Target width in pixels |
| height | string | optional | Target height in pixels |
| crop | string | optional | Crop box as x,y,w,h in pixels (before resize) |
| format | string | optional | Output format (PNG, JPEG, WEBP, …); default keeps source |
| quality | string | default 85 | JPEG/WebP quality 1-100 |
| grayscale | string | default false | Convert to grayscale when true |
Outputs
image
Provenance
- image
- ofthemachine/python3@sha256:a015b4f9f0e7c1648f8fcd4164f6a296dd3c9189404f88f333a9764d08d13522
- reach
network=none— a run is a pure function of its inputs; results are safe to memoize.- procedure_hash
- sha256:66606990e1aec01379992222a56644b0e8c46f95effd1b4139783a07b3738e85
- receipt
- add
--receipt run.jsonto any run;receipt/attest.pyverifies it against this hash without re-running.
Source
transform.py · 101 lines · sha256 66606990e1ae…
#!/usr/bin/env -S fragletc --image ofthemachine/python3@sha256:a015b4f9f0e7c1648f8fcd4164f6a296dd3c9189404f88f333a9764d08d13522
#: d=Resize, crop, convert, and compress images (PNG, JPEG, WebP, GIF, TIFF, BMP), reporting the byte savings on stdout. Set max_dimension to fit the longest edge inside a box while preserving aspect ratio, or give width and/or height directly; crop takes "x,y,w,h" in pixels and is applied before any resize. Shrinks oversized screenshots before they consume an agent's context or get embedded in a PDF.
#: when=Use when an image must be resized, cropped, converted to another format, or compressed to a smaller file size.
#: network=none
#: stdin=none
#: param=image:required:file:d=Source image mounted at /input/image
#: param=max_dimension:d=Fit longest edge inside this many pixels
#: param=width:d=Target width in pixels
#: param=height:d=Target height in pixels
#: param=crop:d=Crop box as x,y,w,h in pixels (before resize)
#: param=format:d=Output format (PNG, JPEG, WEBP, …); default keeps source
#: param=quality:default=85:d=JPEG/WebP quality 1-100
#: param=grayscale:default=false:d=Convert to grayscale when true
#: output=image
import os
import sys
from PIL import Image
src = os.environ.get("IMAGE", "")
crop_spec = os.environ.get("CROP", "").strip()
out_format = os.environ.get("FORMAT", "").strip().upper()
quality = int(os.environ["QUALITY"])
grayscale = os.environ["GRAYSCALE"].strip().lower() in ("1", "true", "yes")
if not src or not os.path.exists(src):
print(f"Error: image file '{src}' does not exist", file=sys.stderr)
sys.exit(1)
def positive_int(name: str):
raw = os.environ.get(name, "").strip()
if not raw:
return None
try:
value = int(raw)
except ValueError:
print(f"Error: {name.lower()} must be an integer, got '{raw}'", file=sys.stderr)
sys.exit(1)
if value <= 0:
print(f"Error: {name.lower()} must be positive, got {value}", file=sys.stderr)
sys.exit(1)
return value
max_dimension = positive_int("MAX_DIMENSION")
width = positive_int("WIDTH")
height = positive_int("HEIGHT")
img = Image.open(src)
src_format = img.format or "PNG"
src_size = img.size
src_bytes = os.path.getsize(src)
if crop_spec:
try:
x, y, w, h = (int(v.strip()) for v in crop_spec.split(","))
except ValueError:
print(f"Error: crop must be 'x,y,w,h' in pixels, got '{crop_spec}'", file=sys.stderr)
sys.exit(1)
img = img.crop((x, y, x + w, y + h))
if max_dimension:
img.thumbnail((max_dimension, max_dimension), Image.LANCZOS)
elif width and height:
img = img.resize((width, height), Image.LANCZOS)
elif width:
img = img.resize((width, round(img.height * width / img.width)), Image.LANCZOS)
elif height:
img = img.resize((round(img.width * height / img.height), height), Image.LANCZOS)
if grayscale:
img = img.convert("L")
target_format = out_format or src_format
if target_format == "JPG":
target_format = "JPEG"
# JPEG has no alpha channel; flatten transparency onto white rather than failing.
if target_format == "JPEG" and img.mode in ("RGBA", "LA", "P"):
img = img.convert("RGBA")
flattened = Image.new("RGB", img.size, (255, 255, 255))
flattened.paste(img, mask=img.split()[-1])
img = flattened
save_args = {}
if target_format in ("JPEG", "WEBP"):
save_args["quality"] = quality
dest = "/output/image"
try:
img.save(dest, format=target_format, **save_args)
except (KeyError, OSError) as e:
print(f"Error: cannot write {target_format}: {e}", file=sys.stderr)
sys.exit(1)
out_bytes = os.path.getsize(dest)
delta = 100 * (out_bytes / src_bytes - 1) if src_bytes else 0.0
print(f"source: {src_size[0]}x{src_size[1]} {src_format} {src_bytes:,} bytes")
print(f"output: {img.size[0]}x{img.size[1]} {target_format} {out_bytes:,} bytes")
print(f"change: {delta:+.1f}% bytes")