Skip to main content

functions

Pure functions for PDF report generation.

Extracted from bitfount.federated.algorithms.ophthalmology.ga_trial_pdf_algorithm_amethyst and ga_trial_pdf_algorithm_base so that the same logic can be called from composable Prefect steps without depending on class inheritance.

PDF rendering is handled by bitfount.steps.pdf_report.pdf_render.

Module​

Functions​

eligible_filenames_from_evaluations​

def eligible_filenames_from_evaluations(    evaluations: list[CriteriaEvaluation],) ‑> set[str]:

Return the set of filenames whose every criterion passed.

Arguments

  • evaluations: The structured evaluations from criteria_matching.

Returns The eligible scan filenames.

extract_record_info​

def extract_record_info(    datasource_row: pd.Series, report_metadata: ReportMetadata,) ‑> RecordInfo:

Extract text field values from a datasource row for the PDF header.

Parses column values according to their type (datetime, float, int, string).

generate_pdf_for_row​

def generate_pdf_for_row(    datasource_row: pd.Series,    results_df: pd.DataFrame,    row_index: Any,    ga_metric: GAMetrics | GAMetricsWithFovea,    report_metadata: ReportMetadata,    base_path: Path,    task_id: str,    original_filename: str,    total_ga_area_lower_bound: float,    total_ga_area_upper_bound: float,    pdf_filename_columns: list[str] | None = None,    filename_prefix: str | None = None,    trial_name: str | None = None,    eligibility: str | None = None,) ‑> pathlib.Path | None:

Generate a complete PDF report for a single patient row.

Orchestrates: scan extraction → record info → path generation → rendering.

Returns Path to the generated PDF, or None if generation failed.

generate_pdf_output_path​

def generate_pdf_output_path(    base_path: Path,    original_filename: str,    row: pd.Series,    task_id: str,    pdf_filename_columns: list[str] | None = None,    filename_prefix: str | None = None,    trial_name: str | None = None,    eligibility: str | None = None,) ‑> pathlib.Path:

Generate a unique output path for a PDF report file.

Arguments

  • base_path: Base output directory.
  • original_filename: The original scan filename.
  • row: The datasource row (for extracting column values for filename).
  • task_id: Task run ID (used as subdirectory).
  • pdf_filename_columns: Columns whose values form the PDF filename.
  • filename_prefix: Optional prefix for the filename.
  • trial_name: Trial name to include in filename.
  • eligibility: Eligibility label to include in filename.

Returns A unique Path that does not conflict with existing files.

get_scan_data​

def get_scan_data(    datasource_row: pd.Series,    results_df: pd.DataFrame,    row_index: Any,    bscan_idx: int | None,    segmentation_labels: dict[str, int] | None = None,) ‑> ScanData:

Extract B-scan image, parse mask, overlay segmentation, return scan data.

Arguments

  • datasource_row: The row from the source DataFrame containing image data.
  • results_df: The predictions DataFrame.
  • row_index: The index of this row in results_df.
  • bscan_idx: Which B-scan slice to use (from ga_metrics.max_ga_bscan_index).
  • segmentation_labels: Label-to-index mapping for mask parsing.

Returns An AltrisScan with images and metadata.

map_subfoveal_indicator​

def map_subfoveal_indicator(value: str | None) ‑> str | None:

Map raw subfoveal indicator 'Y'/'N' to 'Yes'/'No'.

render_pdf​

def render_pdf(    output_path: Path,    record_info: AltrisRecordInfo,    scan: AltrisScan,    ga_metrics: GAMetrics | GAMetricsWithFovea,    task_id: str,    total_ga_area_lower_bound: float,    total_ga_area_upper_bound: float,    eligibility: str | None = None,) ‑> bool:

Render a single patient PDF report.

Arguments

  • output_path: Where to save the PDF.
  • record_info: Patient/record text fields for the header.
  • scan: B-scan image data with segmentation overlay.
  • ga_metrics: GA metrics for this scan.
  • task_id: Task run ID.
  • total_ga_area_lower_bound: Lower bound for slider visualization.
  • total_ga_area_upper_bound: Upper bound for slider visualization.
  • eligibility: Eligibility label (added to text fields if provided).

Returns True if PDF was generated successfully, False otherwise.

validate_pdf_inputs​

def validate_pdf_inputs(    filenames: list[str], ga_dict: Mapping[str, Any], results_df: pd.DataFrame,) ‑> None:

Validate that PDF generation inputs are consistent.

Raises

  • ValueError: If lengths don't match.