Building pages fails when encountering empty field

[caufieldjh]

The generate-pages action fails when provided a source YAML with an empty field.
This is specifically a bug in browser_export.py.

See https://github.com/monarch-initiative/dismech/actions/runs/20855636245/job/59921394677

Run just gen-browser-data
uv run python -c "from pathlib import Path; from dismech.export import BrowserExporter; BrowserExporter().export_to_js(sorted(Path('kb/disorders').glob('*.yaml')), Path('app/data.js'))"
Traceback (most recent call last):
  File "<string>", line 1, in <module>
    from pathlib import Path; from dismech.export import BrowserExporter; BrowserExporter().export_to_js(sorted(Path('kb/disorders').glob('*.yaml')), Path('app/data.js'))
                                                                          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/runner/work/dismech/dismech/src/dismech/export/browser_export.py", line 151, in export_to_js
    record = self.extract_disorder(disorder, file_path.name)
  File "/home/runner/work/dismech/dismech/src/dismech/export/browser_export.py", line 85, in extract_disorder
    genes = [g.get("name", "") for g in disorder.get("genetic", []) if g.get("name")]
                                        ~~~~~~~~~~~~^^^^^^^^^^^^^^^
TypeError: 'NoneType' object is not iterable
error: Recipe `gen-browser-data` failed on line 262 with exit code 1

[github-actions]

Issue Analysis

This is a defensive programming issue in browser_export.py where empty YAML fields are not handled correctly.

Root Cause

The error occurs at line 85 in src/dismech/export/browser_export.py:

genes = [g.get("name", "") for g in disorder.get("genetic", []) if g.get("name")]

When a disorder YAML file contains an empty field like:

genetic:
environmental:

YAML parses this as None (not an empty list). When the code calls disorder.get("genetic", []), it returns None instead of the default [], causing TypeError: 'NoneType' object is not iterable.

Affected Files

128 disorder files have empty genetic: fields, including the recently added Semicircular_Canal_Dehiscence_Syndrome.yaml (PR #84) which likely triggered this failure.

Other Vulnerable Fields

The same pattern is used for multiple fields that could fail with empty values:

• Line 42: has_subtypes
• Line 50: pathophysiology
• Line 71: phenotypes
• Line 85: genetic ⚠️ (current failure)
• Line 88: treatments ⚠️
• Line 91: environmental ⚠️
• Line 94: biochemical ⚠️

Solution

Replace .get(field, []) with .get(field) or [] to handle None values:

# Current (fails on None)
genes = [g.get("name", "") for g in disorder.get("genetic", []) if g.get("name")]

# Fixed (handles None correctly)
genes = [g.get("name", "") for g in (disorder.get("genetic") or []) if g.get("name")]

Recommended Fix

Update all list comprehensions in extract_disorder() to use the or [] pattern:

subtypes = [s.get("name", "") for s in (disorder.get("has_subtypes") or []) if s.get("name")]
genes = [g.get("name", "") for g in (disorder.get("genetic") or []) if g.get("name")]
treatments = [t.get("name", "") for t in (disorder.get("treatments") or []) if t.get("name")]
environmental = [e.get("name", "") for e in (disorder.get("environmental") or []) if e.get("name")]
biochemical = [b.get("name", "") for b in (disorder.get("biochemical") or []) if b.get("name")]

And for loop-based iterations:

for patho in (disorder.get("pathophysiology") or []):
for pheno in (disorder.get("phenotypes") or []):

Testing

After the fix, verify with:

just gen-browser-data

This should handle all 128 files with empty genetic: fields without error.