The disagg dataset (2019–2026, 468 markets) was in the DB but invisible.
This wires it into every layer of the app:
Backend:
- models.py: add has_disagg to CommodityMeta; add DisaggPositionPoint,
DisaggHistoryResponse, DisaggScreenerRow models
- commodities.py: join disagg_reports to populate has_disagg flag and
correct first/last dates; HAVING filter removes markets with no data
- disagg.py (new): /api/disagg/{code}/history, /api/disagg/screener,
/api/disagg/{code}/net-position-percentile, /api/disagg/compare
- main.py: register disagg router
Frontend:
- Metric selector shows Disaggregated optgroup (Managed Money, Prod/Merchant,
Swap Dealer, Other Rept) when market has has_disagg=true, hides Legacy group
- Detail view auto-switches to disagg endpoint and defaults to m_money_net
for disagg markets; shows green "Disaggregated" badge
- Screener always uses disagg endpoint (Managed Money percentile rank)
- Compare uses /api/disagg/compare for disagg metrics
- style.css: add .badge-disagg green variant
Result: wheat markets (SRW, HRW, HRSpring, Black Sea) now show 7 years of
disaggregated positioning data with Managed Money as the default metric.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Integrates the CFTC Disaggregated Commitments of Traders reports
(com_disagg_txt_YYYY.zip) which break positions down by Producer/Merchant,
Swap Dealers, Managed Money, and Other Reportables — a different report
type from the existing legacy COT data.
- schema.sql: add disagg_reports, disagg_positions, disagg_concentration tables
- parser.py: add DisaggPositionRow/DisaggBlock dataclasses and
parse_disagg_csv_text()/parse_disagg_zip_file() for c_year.txt format
- importer.py: add import_disagg_block(), import_disagg_zip_file(),
run_disagg_historical_import() for 2019–2026 yearly ZIPs
- cli.py: add import-disagg-history subcommand
- docker-compose.yaml: run import-disagg-history on startup (idempotent
via import_log, so re-deploys skip already-imported years)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
FastAPI application that ingests CFTC Commitments of Traders data into SQLite
and exposes it via a REST API with analytics endpoints (screener, percentile rank,
concentration). Includes CLI for historical and weekly data ingestion, Docker setup,
and a frontend.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
