Files
ComfyUI-Lora-Manager/tests/utils/test_exif_utils.py
T
Will Miao 69691b17a1 feat(recipes): preserve embedded ComfyUI workflow on remote imports
CivitAI serves a re-encoded, metadata-free optimized rendition as the recipe
preview, so the ComfyUI workflow embedded in the original image was dropped:
imported recipes reported has_workflow=false and never offered "Send Workflow
to ComfyUI" even when the source image carried one.

Recover the workflow from the original rendition and carry it to the save step
as data, so the stored preview stays the small optimized image:

- ExifUtils: embed a caller-supplied workflow during optimize_image's single
  encode pass, and add embed_workflow() to patch WebP EXIF in place (used by
  the verbatim skip_optimize branch and as a safety net).
- RecipePersistenceService.save_recipe: embed metadata["workflow"] before
  detecting has_workflow.
- analyze_remote_image: return the workflow recovered from the original
  rendition it already downloads for EXIF parsing.
- RecipeManagementHandler: add _fetch_original_media() and workflow helpers;
  _do_import_from_url reuses them, and _do_import_remote_recipe fetches the
  original only when CivitAI reports a ComfyUI payload (meta.comfy) so
  workflow-less images pay no extra bandwidth.
- Batch URL imports and the import modal forward the recovered workflow.

Verified against the reported image: has_workflow flips from false to true and
the recovered workflow matches the original (25 nodes, same graph id).
2026-09-29 07:20:12 +08:00

514 lines
19 KiB
Python

import json
from typing import Any, Dict
import piexif # pyright: ignore[reportMissingTypeStubs]
from PIL import Image, PngImagePlugin
from py.utils.exif_utils import ExifUtils
def test_append_recipe_metadata_includes_checkpoint(monkeypatch, tmp_path):
captured = {}
monkeypatch.setattr(
ExifUtils, "extract_image_metadata", staticmethod(lambda _path: None)
)
def fake_update_image_metadata(image_path, metadata):
captured["path"] = image_path
captured["metadata"] = metadata
return image_path
monkeypatch.setattr(
ExifUtils, "update_image_metadata", staticmethod(fake_update_image_metadata)
)
checkpoint = {
"type": "checkpoint",
"modelId": 827184,
"modelVersionId": 2167369,
"modelName": "WAI-illustrious-SDXL",
"modelVersionName": "v15.0",
"hash": "ABC123",
"file_name": "WAI-illustrious-SDXL",
"baseModel": "Illustrious",
}
recipe_data = {
"title": "Semi-realism",
"base_model": "Illustrious",
"loras": [],
"tags": [],
"checkpoint": checkpoint,
}
image_path = tmp_path / "image.webp"
image_path.write_bytes(b"")
ExifUtils.append_recipe_metadata(str(image_path), recipe_data)
assert captured["path"] == str(image_path)
assert captured["metadata"].startswith("Recipe metadata: ")
payload = json.loads(captured["metadata"].split("Recipe metadata: ", 1)[1])
assert payload["checkpoint"] == {
"type": "checkpoint",
"modelId": 827184,
"modelVersionId": 2167369,
"modelName": "WAI-illustrious-SDXL",
"modelVersionName": "v15.0",
"hash": "abc123",
"file_name": "WAI-illustrious-SDXL",
"baseModel": "Illustrious",
}
assert payload["base_model"] == "Illustrious"
def test_append_recipe_metadata_pixel_preserving_webp(tmp_path):
"""pixel_preserving=True must rewrite only the EXIF chunk of a WebP,
leaving the pixel chunks byte-identical and the old block replaced."""
img = Image.new("RGB", (64, 48), (120, 30, 200))
original_params = (
"masterpiece, best quality\n"
"Negative prompt: lowres\n"
"Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 1, "
"Size: 512x768, Model hash: abc123, Model: foo_v1, Clip skip: 2\n"
' Recipe metadata: {"title": "Old", "loras": []}'
)
exif = piexif.dump(
{
"0th": {},
"Exif": {
piexif.ExifIFD.UserComment: b"UNICODE\x00"
+ original_params.encode("utf-16be")
},
}
)
image_path = tmp_path / "recipe.webp"
img.save(str(image_path), format="WEBP", exif=exif, quality=85)
with open(image_path, "rb") as fh:
before = fh.read()
new_recipe = {
"title": "New",
"base_model": "SDXL",
"loras": [],
"gen_params": {"steps": 25},
}
ExifUtils.append_recipe_metadata(
str(image_path), new_recipe, pixel_preserving=True
)
with open(image_path, "rb") as fh:
after = fh.read()
def chunks(data: bytes) -> Dict[bytes, bytes]:
pos, result = 12, {}
while pos + 8 <= len(data):
fourcc = data[pos : pos + 4]
size = int.from_bytes(data[pos + 4 : pos + 8], "little")
result[fourcc] = data[pos + 8 : pos + 8 + size]
pos += 8 + size + (size % 2)
return result
before_chunks = chunks(before)
after_chunks = chunks(after)
for fourcc, payload in before_chunks.items():
if fourcc == b"EXIF":
assert after_chunks[fourcc] != payload, "EXIF must be replaced"
else:
assert after_chunks[fourcc] == payload, f"{fourcc} was re-encoded"
# The appended block is updated; the original parameters stay in front.
metadata = ExifUtils.extract_image_metadata(str(image_path))
assert "Steps: 20" in metadata
assert 'Recipe metadata: {"title": "New"' in metadata
assert '"title": "Old"' not in metadata
def test_optimize_image_preserves_workflow_when_converting_png_to_webp(tmp_path):
image_path = tmp_path / "source.png"
png_info = PngImagePlugin.PngInfo()
png_info.add_text("parameters", "prompt text\nSteps: 20")
png_info.add_text("workflow", json.dumps({"nodes": [{"id": 1}]}))
Image.new("RGB", (64, 32), color="red").save(image_path, pnginfo=png_info)
optimized_data, extension = ExifUtils.optimize_image(
str(image_path),
target_width=32,
format="webp",
quality=85,
preserve_metadata=True,
)
optimized_path = tmp_path / f"optimized{extension}"
optimized_path.write_bytes(optimized_data)
exif_dict = piexif.load(str(optimized_path))
assert exif_dict["0th"] is not None
assert (
exif_dict["0th"][piexif.ImageIFD.ImageDescription].decode("utf-8")
== 'Workflow:{"nodes": [{"id": 1}]}'
)
assert exif_dict["Exif"] is not None
user_comment = exif_dict["Exif"][piexif.ExifIFD.UserComment]
assert user_comment.startswith(b"UNICODE\0")
assert user_comment[8:].decode("utf-16be") == "prompt text\nSteps: 20"
def test_update_image_metadata_preserves_webp_workflow(tmp_path):
image_path = tmp_path / "recipe.webp"
exif_dict = {
"0th": {
piexif.ImageIFD.ImageDescription: 'Workflow:{"nodes":[{"id":1}]}',
},
"Exif": {
piexif.ExifIFD.UserComment: b"UNICODE\0"
+ "prompt text".encode("utf-16be")
},
}
Image.new("RGB", (32, 32), color="blue").save(
image_path, format="WEBP", exif=piexif.dump(exif_dict), quality=85
)
ExifUtils.update_image_metadata(
str(image_path), 'prompt text\nRecipe metadata: {"title":"recipe"}'
)
updated_exif = piexif.load(str(image_path))
assert updated_exif["0th"] is not None
assert (
updated_exif["0th"][piexif.ImageIFD.ImageDescription].decode("utf-8")
== 'Workflow:{"nodes":[{"id":1}]}'
)
assert updated_exif["Exif"] is not None
updated_comment = updated_exif["Exif"][piexif.ExifIFD.UserComment]
assert (
updated_comment[8:].decode("utf-16be")
== 'prompt text\nRecipe metadata: {"title":"recipe"}'
)
def test_update_image_metadata_preserves_png_workflow(tmp_path):
image_path = tmp_path / "recipe.png"
png_info = PngImagePlugin.PngInfo()
png_info.add_text("parameters", "prompt text")
png_info.add_text("workflow", '{"nodes":[{"id":1}]}')
Image.new("RGB", (32, 32), color="green").save(image_path, pnginfo=png_info)
ExifUtils.update_image_metadata(
str(image_path), 'prompt text\nRecipe metadata: {"title":"recipe"}'
)
with Image.open(image_path) as img:
assert img.info["workflow"] == '{"nodes":[{"id":1}]}'
assert (
img.info["parameters"]
== 'prompt text\nRecipe metadata: {"title":"recipe"}'
)
def test_optimize_image_embeds_supplied_workflow_when_source_has_none(tmp_path):
"""Import paths hand the workflow over as data when the preview source is
metadata-free (CivitAI's optimized rendition); optimize_image must embed
it while re-encoding, otherwise the recipe loses has_workflow."""
image_path = tmp_path / "optimized.webp"
Image.new("RGB", (64, 32), color="red").save(image_path, format="WEBP", quality=85)
workflow = {"nodes": [{"id": 1}], "last_node_id": 1}
optimized_data, extension = ExifUtils.optimize_image(
str(image_path),
target_width=32,
format="webp",
quality=85,
preserve_metadata=True,
workflow=workflow,
)
optimized_path = tmp_path / f"embedded{extension}"
optimized_path.write_bytes(optimized_data)
metadata = ExifUtils._load_structured_metadata(str(optimized_path))
assert metadata["workflow"] == json.dumps(workflow)
def test_optimize_image_keeps_source_workflow_over_supplied(tmp_path):
image_path = tmp_path / "source.png"
png_info = PngImagePlugin.PngInfo()
png_info.add_text("workflow", '{"nodes":[{"id":7}]}')
Image.new("RGB", (64, 32), color="red").save(image_path, pnginfo=png_info)
optimized_data, extension = ExifUtils.optimize_image(
str(image_path),
target_width=32,
format="webp",
quality=85,
preserve_metadata=True,
workflow={"nodes": [{"id": 1}]},
)
optimized_path = tmp_path / f"sourcewins{extension}"
optimized_path.write_bytes(optimized_data)
metadata = ExifUtils._load_structured_metadata(str(optimized_path))
assert metadata["workflow"] == '{"nodes":[{"id":7}]}'
def test_embed_workflow_adds_workflow_to_metadata_free_webp(tmp_path):
image_path = tmp_path / "preview.webp"
Image.new("RGB", (32, 32), color="blue").save(image_path, format="WEBP", quality=85)
workflow = json.dumps({"nodes": [{"id": 1}]})
returned = ExifUtils.embed_workflow(str(image_path), workflow)
assert returned == str(image_path)
metadata = ExifUtils._load_structured_metadata(str(image_path))
assert metadata["workflow"] == workflow
with Image.open(image_path) as img:
assert img.size == (32, 32)
def test_embed_workflow_adds_workflow_to_metadata_free_png(tmp_path):
image_path = tmp_path / "preview.png"
Image.new("RGB", (32, 32), color="blue").save(image_path)
workflow = {"nodes": [{"id": 3}]}
ExifUtils.embed_workflow(str(image_path), workflow)
metadata = ExifUtils._load_structured_metadata(str(image_path))
assert metadata["workflow"] == json.dumps(workflow)
def test_embed_workflow_leaves_existing_workflow_untouched(tmp_path):
image_path = tmp_path / "preview.png"
png_info = PngImagePlugin.PngInfo()
png_info.add_text("workflow", '{"nodes":[{"id":9}]}')
Image.new("RGB", (32, 32), color="green").save(image_path, pnginfo=png_info)
ExifUtils.embed_workflow(str(image_path), {"nodes": [{"id": 1}]})
with Image.open(image_path) as img:
assert img.info["workflow"] == '{"nodes":[{"id":9}]}'
def test_embed_workflow_ignores_unsupported_payloads_and_containers(tmp_path):
image_path = tmp_path / "preview.webp"
Image.new("RGB", (16, 16), color="black").save(image_path, format="WEBP")
# Nothing to embed / unsupported payload types are no-ops.
assert ExifUtils.embed_workflow(str(image_path), None) == str(image_path)
assert ExifUtils.embed_workflow(str(image_path), "") == str(image_path)
assert ExifUtils.embed_workflow(str(image_path), 123) == str(image_path)
assert ExifUtils._load_structured_metadata(str(image_path))["workflow"] is None
video_path = tmp_path / "clip.mp4"
video_path.write_bytes(b"video")
assert ExifUtils.embed_workflow(str(video_path), {"nodes": []}) == str(video_path)
# --- ISOBMFF / brotli extraction tests ---
import struct
import brotli # pyright: ignore[reportMissingTypeStubs]
def _build_jxl_with_brob(payload_json: Dict[str, Any]) -> bytes:
"""Build a minimal JXL container with a brob box containing brotli-compressed JSON."""
# ISOBMFF box 1: JXL signature box (size=12, type='JXL ', signature)
box1 = struct.pack(">I", 12) + b"JXL " + bytes([0x0d, 0x0a, 0x87, 0x0a])
# ISOBMFF box 2: ftyp (size=16, type='ftyp', major='jxl ', minor=0)
box2 = struct.pack(">I", 16) + b"ftyp" + b"jxl " + struct.pack(">I", 0)
# ISOBMFF box 3: brob — payload is b'comf' + brotli(json)
compressed = brotli.compress(json.dumps(payload_json).encode("utf-8"))
brob_payload = b"comf" + compressed
box3 = struct.pack(">I", 8 + len(brob_payload)) + b"brob" + brob_payload
return box1 + box2 + box3
def _build_avif_with_brob(payload_json: Dict[str, Any]) -> bytes:
"""Build a minimal AVIF container with a brob box containing brotli-compressed JSON."""
compressed = brotli.compress(json.dumps(payload_json).encode("utf-8"))
brob_payload = b"comf" + compressed
ftyp_box = struct.pack(">I", 20) + b"ftyp" + b"avif" + struct.pack(">I", 0) + b"avif"
brob_box = struct.pack(">I", 8 + len(brob_payload)) + b"brob" + brob_payload
return ftyp_box + brob_box
class TestIsobmffBrotliExtraction:
"""Tests for ISOBMFF brotli metadata extraction in ExifUtils."""
def test_extract_jxl_brotli_happy_path(self, tmp_path):
"""JXL container with valid brob box extracts prompt and workflow."""
payload = {"prompt": "a cute cat", "workflow": {"nodes": [{"id": 1}]}}
data = _build_jxl_with_brob(payload)
path = tmp_path / "test.jxl"
path.write_bytes(data)
result = ExifUtils._load_structured_metadata(str(path))
assert result["prompt"] == "a cute cat"
assert result["workflow"] == '{"nodes": [{"id": 1}]}'
assert result["parameters"] is None
assert result["comment"] is None
def test_extract_avif_brotli_happy_path(self, tmp_path):
"""AVIF container with valid brob box extracts prompt and workflow."""
payload = {"prompt": "landscape", "workflow": {"nodes": []}}
data = _build_avif_with_brob(payload)
path = tmp_path / "test.avif"
path.write_bytes(data)
result = ExifUtils._load_structured_metadata(str(path))
assert result["prompt"] == "landscape"
assert result["workflow"] == '{"nodes": []}'
def test_extract_no_brob_box_returns_none(self, tmp_path):
"""JXL container without a brob box returns None from _extract_isobmff_brotli."""
# Only JXL signature + ftyp, no brob
box1 = struct.pack(">I", 12) + b"JXL " + bytes([0x0d, 0x0a, 0x87, 0x0a])
box2 = struct.pack(">I", 16) + b"ftyp" + b"jxl " + struct.pack(">I", 0)
path = tmp_path / "test.jxl"
path.write_bytes(box1 + box2)
# The low-level extraction should return None (no brob box)
result = ExifUtils._extract_isobmff_brotli(str(path))
assert result is None
def test_extract_corrupt_brob_returns_none(self, tmp_path):
"""Broken brob box payload gracefully returns None."""
box1 = struct.pack(">I", 12) + b"JXL " + bytes([0x0d, 0x0a, 0x87, 0x0a])
box2 = struct.pack(">I", 16) + b"ftyp" + b"jxl " + struct.pack(">I", 0)
# brob with garbage payload that doesn't start with b'comf'
garbage = b"\xff\xff\xff\xff" * 32
box3 = struct.pack(">I", 8 + len(garbage)) + b"brob" + garbage
path = tmp_path / "test.jxl"
path.write_bytes(box1 + box2 + box3)
result = ExifUtils._extract_isobmff_brotli(str(path))
assert result is None
def test_extract_non_isobmff_file_falls_through(self, tmp_path):
"""A regular PNG file is not processed as ISOBMFF and returns PIL metadata."""
png_info = PngImagePlugin.PngInfo()
png_info.add_text("prompt", "from png")
path = tmp_path / "test.png"
Image.new("RGB", (4, 4), color="red").save(path, pnginfo=png_info)
result = ExifUtils._load_structured_metadata(str(path))
assert result["prompt"] == "from png"
def test_extract_skip_on_update_and_optimize(self, tmp_path):
"""AVIF/JXL files are skipped for write operations (update/append/optimize)."""
path = tmp_path / "test.avif"
path.write_bytes(b"fake avif data")
# update_image_metadata should return the path unchanged
result = ExifUtils.update_image_metadata(str(path), "some metadata")
assert result == str(path)
# append_recipe_metadata should also skip
result = ExifUtils.append_recipe_metadata(str(path), {"title": "test"})
assert result == str(path)
# optimize_image should passthrough for AVIF/JXL paths
result_data, ext = ExifUtils.optimize_image(str(path))
assert ext == ".avif"
assert result_data == b"fake avif data"
def test_extract_prompt_as_dict(self, tmp_path):
"""prompt field as dict is JSON-serialized."""
payload = {"prompt": {"text": "hello", "negative": "bad"}}
data = _build_jxl_with_brob(payload)
path = tmp_path / "test.jxl"
path.write_bytes(data)
result = ExifUtils._load_structured_metadata(str(path))
assert result["prompt"] is not None
assert json.loads(result["prompt"]) == {"text": "hello", "negative": "bad"}
def test_extract_workflow_as_list(self, tmp_path):
"""workflow field as list is JSON-serialized."""
payload = {"workflow": [{"id": 1}, {"id": 2}]}
data = _build_avif_with_brob(payload)
path = tmp_path / "test.avif"
path.write_bytes(data)
result = ExifUtils._load_structured_metadata(str(path))
assert result["workflow"] is not None
assert json.loads(result["workflow"]) == [{"id": 1}, {"id": 2}]
def test_over_decompressed_size_limit(self, tmp_path, monkeypatch):
"""Decompressed data exceeding _BROTLI_MAX_DECOMPRESSED is rejected."""
# Monkey-patch the limit to a small value to avoid large test data
monkeypatch.setattr(ExifUtils, "_BROTLI_MAX_DECOMPRESSED", 100)
large_content = "x" * 200
payload = {"prompt": large_content}
data = _build_jxl_with_brob(payload)
path = tmp_path / "test.jxl"
path.write_bytes(data)
# Direct extraction should return None because decompressed size exceeds limit
result = ExifUtils._extract_isobmff_brotli(str(path))
assert result is None
# --- get_image_dimensions tests ---
def test_get_image_dimensions_returns_actual_size(tmp_path):
"""(a) A valid image returns its real (width, height)."""
image_path = tmp_path / "preview.png"
Image.new("RGB", (64, 32), color="red").save(image_path)
assert ExifUtils.get_image_dimensions(str(image_path)) == (64, 32)
def test_get_image_dimensions_missing_path_returns_none(tmp_path):
"""(b) A nonexistent path returns None without raising."""
assert ExifUtils.get_image_dimensions(str(tmp_path / "missing.png")) is None
def test_get_image_dimensions_skips_video_extension_without_pil(tmp_path, monkeypatch):
"""(c) A .mp4 path returns None and never invokes PIL."""
video_path = tmp_path / "preview.mp4"
video_path.write_bytes(b"not really a video")
def fail_if_called(*args, **kwargs):
raise AssertionError("PIL Image.open must not be called for video paths")
monkeypatch.setattr("py.utils.exif_utils.Image.open", fail_if_called)
assert ExifUtils.get_image_dimensions(str(video_path)) is None
def test_get_image_dimensions_corrupt_file_returns_none(tmp_path):
"""(d) A corrupt file returns None without raising."""
image_path = tmp_path / "corrupt.png"
image_path.write_bytes(b"\x00\x01\x02\x03 not a real image")
assert ExifUtils.get_image_dimensions(str(image_path)) is None
def test_get_image_dimensions_cache_key_includes_mtime(tmp_path):
"""(e) Replacing a path with a different-size image returns the new size."""
image_path = tmp_path / "replaced.png"
Image.new("RGB", (64, 32), color="red").save(image_path)
assert ExifUtils.get_image_dimensions(str(image_path)) == (64, 32)
Image.new("RGB", (100, 50), color="blue").save(image_path)
assert ExifUtils.get_image_dimensions(str(image_path)) == (100, 50)
def test_get_image_dimensions_skips_unreadable_formats(tmp_path):
"""(f) .avif/.jxl paths return None without raising."""
for ext in (".avif", ".jxl"):
image_path = tmp_path / f"preview{ext}"
image_path.write_bytes(b"fake container data")
assert ExifUtils.get_image_dimensions(str(image_path)) is None