drc-eval-workflow

name	drc-eval-workflow
description	Use when running KLayout DRC evaluation pipelines — executing DRC rules against GDS test files, parsing results, and computing metrics.

DRC Evaluation Workflow

Running KLayout DRC in Batch Mode

klayout -b -r rule.drc -rd input=test.gds -rd output=result.lyrdb

• -b: Batch mode (no GUI)
• -r rule.drc: DRC rule file
• -rd input=...: Pass GDS input path as variable $input
• -rd output=...: Pass output path as variable $output

Parsing .lyrdb Results

The .lyrdb file is XML (KLayout Report Database). Parse it to extract violations:

import xml.etree.ElementTree as ET

def parse_lyrdb(lyrdb_path):
    """Parse KLayout RDB file and return violations by category."""
    tree = ET.parse(lyrdb_path)
    root = tree.getroot()
    violations = {}
    for cat in root.iter("category"):
        name_el = cat.find("name")
        if name_el is not None:
            name = name_el.text
            count = sum(1 for _ in cat.iter("value"))
            violations[name] = count
    return violations

Evaluation Workflow

1. Prepare: Collect DRC rule (.drc) and test GDS files (pass/ and fail/ directories)
2. Run: Execute KLayout on each GDS file
3. Parse: Extract violation counts from each .lyrdb output
4. Score: Compare against ground truth (pass files should have 0 violations, fail files should have >0)
5. Metrics: Compute per-category F1, compile rate, and overall success

Computing F1 Score

def compute_f1(predictions, labels):
    """Compute F1 from binary predictions and labels.
    predictions: dict of {filename: {category: violation_count}}
    labels: dict of {filename: {category: 0_or_1}}
    """
    tp = fp = fn = 0
    for fname, cats in labels.items():
        for cat, label in cats.items():
            pred = 1 if predictions.get(fname, {}).get(cat, 0) > 0 else 0
            if pred == 1 and label == 1:
                tp += 1
            elif pred == 1 and label == 0:
                fp += 1
            elif pred == 0 and label == 1:
                fn += 1
    precision = tp / (tp + fp) if (tp + fp) > 0 else 0
    recall = tp / (tp + fn) if (tp + fn) > 0 else 0
    f1 = 2 * precision * recall / (precision + recall) if (precision + recall) > 0 else 0
    return {"f1": f1, "precision": precision, "recall": recall, "tp": tp, "fp": fp, "fn": fn}

Quick Single-File Check

For a quick validation of one DRC rule against one GDS file:

# Create temp output
TMPFILE=$(mktemp /tmp/drc_XXXXXX.lyrdb)
klayout -b -r my_rule.drc -rd input=test.gds -rd output=$TMPFILE
# Check if any violations found
grep -c "<value>" $TMPFILE  # 0 = no violations