PolicyEngine / PolicyEngine/chronicle
String year keys in manifest files: crash the {year} templating in source_package.py
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 0
- Forks
- 1
- Avg merge
- 11h 48m
- Merged PRs (30d)
- 45
Description
Problem
ledger/source_package.py:1735 runs value.format(year=year, filing_year=year + 1) while iterating manifest files: keys. When a manifest uses a non-numeric key (a vintage/release label rather than a year), year arrives as a string and year + 1 raises TypeError.
Two manifests currently trip it:
cms_nhe/historical_service_sourcessa .../extracted_targets
Neither loses facts today (NHE builds fully via its column_by_year data years — CY1966–2024, 59 facts; the SSA extracted set is redundant with its numeric 2024 key), but the crash blocks the affected (package, key) builds and will bite the first manifest where a labeled key is the only source of facts.
Related edge, same root: _year_mapping(column_by_year, <label-year>) raises "No source artifact for year ..." for release-label keys with no data column (seen with cbo/individual_income_tax_receipts_2026_02 [2026] and federal_reserve/z1_household_net_worth [2026]).
Suggested fix
Treat non-integer files: keys as labels: skip the {year}-templating path for them (or coerce with an explicit int() guard and a clear "label key, no year templating" branch), and make the per-key build report distinguish "label key, nothing to build" from a real failure.
Found 2026-07-03 while rebuilding the canonical consumer_facts.jsonl for the populace US build.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Read ledger/source_package.py around line 1735, then inspect the files: entries in cms_nhe/historical_service_source and ssa .../extracted_targets. Reproduce the affected package/key builds and verify that non-integer keys avoid year templating, while the per-key report distinguishes a label key with nothing to build from a real failure.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- build-system, data-engineering
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 68/100