compare_basket
ChatGPTPrice up to 30 leaf categories with exact decimal quantities and explicit units. Fetch each category's quantity basis and supported area IDs first. Use the user's physical/count unit: five pounds is quantity 5, unit pounds; 24 eggs is quantity 24, unit units; 12 ounces is quantity 12, unit ounces. display_units counts the whole displayed equivalent, not pounds or ounces. Incomplete baskets show gaps and are excluded from complete rankings.
compare_price_quotes
ChatGPTCompare user-supplied USD package prices using compatible physical/count units and a stated common quantity. Optional dated category benchmark. Supplied quotes are distinct from CostInflation observations.
decompose_basket_change
ChatGPTAttribute price changes for fixed item quantities across two dates. Complete item contributions reconcile exactly with total change; partial evidence has no complete total. Quantity changes are not mixed into this attribution.
fetch
ChatGPTData-first retrieval of one resolved category's quantity basis, supported areas, dated public evidence and canonical citation URL. Prepare a final history or basket result without creating an intermediate view for each category.
get_basket_history
ChatGPTCalculate a basket on each calendar date with explicit quantities and units. Same-day observations by default; only an explicit bounded_as_of policy permits earlier observations. Incomplete totals are null.
get_catalog
ChatGPTEnumerate canonical category IDs, quantities, units, links and supported operations. Stable filter-aware cursor pagination; no data reads.
get_coverage
ChatGPTInspect selected categories and areas for available dates, observed and missing days, date gaps and freshness. Does not invent coverage for missing areas.
get_history_statistics
ChatGPTCompute equally weighted daily min, max, nearest-rank median and current historical rank with explicit ties and gaps. These are statistics of category observations, not product-price distributions or forecasts.
get_package_size_changes
ChatGPTRead consecutive-day matched-listing package quantity and unit-price change summaries. Counts expose matched, unmatched and ambiguous cohorts. Only stable identity/title/brand pairs are matched; do not label all changes shrinkflation or infer causality.
get_price_changes
ChatGPTCompare two actual dated boundaries for selected categories and areas. Exact USD and percentage changes, zero-baseline handling and optional numeric ordering; earlier boundary dates are explicit.
get_price_distribution
ChatGPTRead daily p05/p25/p75/p95 bands and datapoint counts at the canonical measured quantity. Actual dates, missing areas and published weighting method are explicit. Averages of panel quantiles are not pooled percentiles; summed marginal bands are not basket percentiles.
get_price_history
ChatGPTExplore one category, supported area IDs, actual dates and calculated change. Omit areas to discover the panel. Dates choose latest observations on or before boundaries; at most 512 rows. The interactive view includes every observation; keep chat to requested findings, scope and citations instead of repeating its complete table.
get_price_matrix
ChatGPTRead exact category prices for the displayed measured quantities. Returns actual observation dates, age and missing/earlier cells for an optional as-of date.
render_price_view
ChatGPTRender the view_data returned by a calculation tool as a line, bars, contribution, matrix or accessible table. Copy the typed prepared data unchanged; this presentation tool does not recalculate evidence. Line views require dated series.
search
ChatGPTSearch the bounded public category catalog. Returns IDs and canonical citation URLs.
