Merge pull request #42 from DigitalSlideArchive/initial-rule-algorithm

imagedephi/__init__.py

@@ -1,74 +0,0 @@
-from __future__ import annotations
-
-from enum import Enum
-from pathlib import Path
-from typing import TYPE_CHECKING, Any
-
-import click
-import tifftools
-import tifftools.constants
-
-if TYPE_CHECKING:
-    from tifftools.tifftools import IFD, TiffInfo
-
-
-class RedactMethod(Enum):
-    REPLACE = 1
-    DELETE = 2
-
-
-def get_tags_to_redact() -> dict[int, dict[str, Any]]:
-    return {
-        270: {
-            "id": 270,
-            "name": "ImageDescription",
-            "method": RedactMethod.REPLACE,
-            "replace_value": "Redacted by ImageDePHI",
-        }
-    }
-
-
-def redact_one_tag(ifd: IFD, tag: tifftools.TiffTag, redact_instructions: dict[str, Any]) -> None:
-    if redact_instructions["method"] == RedactMethod.REPLACE:
-        ifd["tags"][tag.value]["data"] = redact_instructions["replace_value"]
-    elif redact_instructions["method"] == RedactMethod.DELETE:
-        del ifd["tags"][tag.value]
-
-
-def redact_tiff_tags(ifds: list[IFD], tags_to_redact: dict[int, dict[str, Any]]) -> None:
-    for ifd in ifds:
-        for tag_id, tag_info in sorted(ifd["tags"].items()):
-            tag: tifftools.TiffTag = tifftools.constants.get_or_create_tag(
-                tag_id,
-                datatype=tifftools.Datatype[tag_info["datatype"]],
-            )
-            if not tag.isIFD():
-                if tag.value in tags_to_redact:
-                    redact_one_tag(ifd, tag, tags_to_redact[tag.value])
-            else:
-                # tag_info['ifds'] contains a list of lists
-                # see tifftools.read_tiff
-                for sub_ifds in tag_info.get("ifds", []):
-                    redact_tiff_tags(sub_ifds, tags_to_redact)
-
-
-def redact_one_image(tiff_info: TiffInfo, output_path: Path) -> None:
-    ifds = tiff_info["ifds"]
-    tags_to_redact = get_tags_to_redact()
-    redact_tiff_tags(ifds, tags_to_redact)
-    tifftools.write_tiff(tiff_info, output_path)
-
-
-def get_output_path(file_path: Path, output_dir: Path) -> Path:
-    return output_dir / f"REDACTED_{file_path.name}"
-
-
-def redact_images(image_dir: Path, output_dir: Path) -> None:
-    for child in image_dir.iterdir():
-        try:
-            tiff_info: TiffInfo = tifftools.read_tiff(child)
-        except tifftools.TifftoolsError:
-            click.echo(f"Could not open {child.name} as a tiff. Skipping...")
-            continue
-        click.echo(f"Redacting {child.name}...")
-        redact_one_image(tiff_info, get_output_path(child, output_dir))

imagedephi/base_rules.yaml

@@ -0,0 +1,84 @@
+---
+name: Base rules
+description: These are the default rules for each potential piece of redactable material (metadata, image) to be used as a fallback if there are no user-defined rules for a piece of redactable material.
+rules:
+  # For images that strictly adhere to the tiff standards e.g. svs files
+  tiff:
+    - tag_name: NewSubfileType
+      method: keep
+      type: metadata
+    - tag_name: ImageWidth
+      method: keep
+      type: metadata
+    - tag_name: ImageLength
+      method: keep
+      type: metadata
+    - tag_name: BitsPerSample
+      method: keep
+      type: metadata
+    - tag_name: Compression
+      method: keep
+      type: metadata
+    - tag_name: Photometric
+      method: keep
+      type: metadata
+    - tag_name: ImageDescription
+      method: delete
+      type: metadata
+    - tag_name: Orientation
+      method: keep
+      type: metadata
+    - tag_name: SamplesPerPixel
+      method: keep
+      type: metadata
+    - tag_name: XResolution
+      method: keep
+      type: metadata
+    - tag_name: YResolution
+      method: keep
+      type: metadata
+    - tag_name: PlanarConfig
+      method: keep
+      type: metadata
+    - tag_name: ResolutionUnit
+      method: keep
+      type: metadata
+    - tag_name: TileWidth
+      method: keep
+      type: metadata
+    - tag_name: TileHeight
+      method: keep
+      type: metadata
+    - tag_name: TileOffsets
+      method: keep
+      type: metadata
+    - tag_name: TileByteCounts
+      method: keep
+      type: metadata
+    - tag_name: SampleFormat
+      method: keep
+      type: metadata
+    - tag_name: JPEGTables
+      method: keep
+      type: metadata
+    - tag_name: YCbCrSubsampling
+      method: keep
+      type: metadata
+    - tag_name: ImageDepth
+      method: keep
+      type: metadata
+    - tag_name: ICCProfile
+      method: keep
+      type: metadata
+    - tag_name: StripOffsets
+      method: keep
+      type: metadata
+    - tag_name: RowsPerStrip
+      method: keep
+      type: metadata
+    - tag_name: StripByteCounts
+      method: keep
+      type: metadata
+    - tag_name: Predictor
+      method: keep
+      type: metadata

imagedephi/main.py

@@ -1,35 +1,65 @@
 from __future__ import annotations
 
 import asyncio
+from dataclasses import dataclass
 from pathlib import Path
+from typing import TextIO
 import webbrowser
 
 import click
 from hypercorn import Config
 from hypercorn.asyncio import serve
+import yaml
 
 from imagedephi.async_utils import run_coroutine, wait_for_port
 from imagedephi.gui import app, shutdown_event
-from imagedephi.redact import redact_images
+from imagedephi.redact import RuleSource, build_ruleset, redact_images, show_redaction_plan
+from imagedephi.rules import RuleSet
+
+
+@dataclass
+class ImagedephiContext:
+    override_rule_set: RuleSet | None = None
 
 
 @click.group
-def imagedephi() -> None:
+@click.option(
+    "-r",
+    "--override-rules",
+    type=click.File("r"),
+    help="Specify user-defined rules to override defaults",
+)
+@click.pass_context
+def imagedephi(ctx: click.Context, override_rules: TextIO | None) -> None:
     """Redact microscopy whole slide images."""
-    pass
+    obj = ImagedephiContext()
+    # Store separately, to preserve the type of "obj"
+    ctx.obj = obj
+
+    if override_rules:
+        obj.override_rule_set = build_ruleset(yaml.safe_load(override_rules), RuleSource.OVERRIDE)
 
 
 @imagedephi.command
 @click.argument(
     "input-dir", type=click.Path(exists=True, file_okay=False, readable=True, path_type=Path)
 )
 @click.argument(
-    "output-dir", type=click.Path(exists=True, file_okay=False, writable=True, path_type=Path)
+    "output-dir",
+    type=click.Path(exists=True, file_okay=False, readable=True, writable=True, path_type=Path),
 )
-def run(input_dir: Path, output_dir: Path) -> None:
-    """Run in CLI-only mode."""
-    redact_images(input_dir, output_dir)
-    click.echo("Done!")
+@click.pass_obj
+def run(obj: ImagedephiContext, input_dir: Path, output_dir: Path):
+    """Redact images in a folder according to given rule sets."""
+    redact_images(input_dir, output_dir, obj.override_rule_set)
+
+
+@imagedephi.command
+@click.argument("image", type=click.Path())
+@click.pass_obj
+def plan(obj: ImagedephiContext, image: click.Path) -> None:
+    """Print the redaction plan for a given image and rules."""
+    show_redaction_plan(image, obj.override_rule_set)
 
 
 @imagedephi.command