Plugins and Extensibility

The phenotyping pipeline supports optional plugin registries for metadata extractors, report sections, and property functions.

---

Table of Contents

1. Property Functions
2. Metadata Extractors
3. Report Sections

---

Property Functions

Register custom property functions for use with compute_properties. Property functions have signature (image2d, mask2d) -> Dict[str, float].

Registering a Property

from phenome.plugins import register_property

def my_blob_counter(image2d, mask2d):
    """Count blobs in image."""
    if image2d is None:
        return {"my_blob_count": float("nan")}
    # ... your logic ...
    return {"my_blob_count": 42.0}

register_property("my_blobs", my_blob_counter)

Using a Registered Property

from phenome.plugins import get_property, get_blob_properties

df = pheno.compute_properties(
    property_preset="none",
    additional_property_functions={"image": [get_blob_properties]}
)

blob_fn = get_property("blob")
if blob_fn:
    df = pheno.compute_properties(
        property_preset="none",
        additional_property_functions={"image": [blob_fn]}
    )

Built-in Properties

Name	Function	Description
blob	get_blob_properties	DoG blob detection (count, area, intensity stats)

---

Metadata Extractors

Register metadata extractors for use with find_files(metadata_fn=...) or for discovery.

Registering an Extractor

from phenome.plugins import register_metadata_extractor
import re

def extract_from_path_template(path: str):
    """Extract metadata from path like /data/plate1/well_A01.tif"""
    match = re.search(r"plate(\d+)/well_([A-Z]\d+)", path)
    if match:
        return {"file_path": path, "plate": match.group(1), "well": match.group(2)}
    return {"file_path": path}

register_metadata_extractor("plate_well", extract_from_path_template)

Using a Registered Extractor

from phenome.plugins import get_metadata_extractor

extractor = get_metadata_extractor("plate_well")
if extractor:
    df = pheno.find_files("/data/images", metadata_fn=extractor)

---

Report Sections

Register custom HTML section generators for inclusion in generate_report().

Registering a Section

from phenome.plugins import register_report_section

def my_custom_section(pipeline, title: str = "Custom Analysis") -> str:
    """Generate HTML for a custom report section."""
    n_images = pipeline.results.n_images
    return f"""
    <section id="custom">
        <h2>{title}</h2>
        <p>Processed {n_images} images.</p>
    </section>
    """

register_report_section("custom_analysis", my_custom_section)

Report integration: Custom report sections are registered in the plugin registry but are not automatically included in generate_report() by default. To add them to your reports, you need to extend generate_report in your project to iterate over get_report_sections() and append the generated HTML to the report body. The registry provides the discovery mechanism; the main report generator does not call it by default.

---

See Also

Topic	Document
Extension points overview	Extending the Pipeline
Metadata extractors	Experiment Details
Property functions	Select properties
Report API	Report Generation