Report data

Measurements arrive on the order itself, so the webhook body, GET /orders/:id and GET /orders/:id/status all carry the same numbers — you never need a second call to get them. totals covers the whole property; each entry in structures[] is one building (house, detached garage, shed) with the identical shape.

Field	What it is
structureCount	Separate buildings measured. totals is the sum across all of them. Inferred from geometry, not read off the document.
structures[].index	1 is the largest building. Ordered by roof area, descending — so structures[0] is the house and outbuildings follow. Stable: the same document always indexes the same way, so an index is safe to store or put in a URL.
totalAreaSqFt / totalSquares	Sloped roof area. Squares is area ÷ 100.
facetCount	Number of roof planes.
predominantPitch	The pitch covering the most area. Flat or N/12.
pitchAreas	Roof area at each pitch — what you need to price steep-slope labour.
edgeTotals	Length and count per category: EAVE, RAKE, RIDGE, HIP, VALLEY, STEP_FLASHING, WALL_FLASHING, PARAPET, BEND, OTHER. Always all ten, zero-filled. BEND is a pitch transition — where two different slopes meet; PARAPET is the run of parapet wall along the roof edge. These are the same names the PDF's measurement tables print, so a number you read on the report can be found in the JSON without a lookup.
totalEdgeLengthFt	Every edge added up.
hipsAndRidgesLengthFt	Hips and ridges together — ridge cap is ordered against the pair.
complexity	simple / normal / complex, from facet count and total hip+valley run.
estimatedAtticSqFt	Sloped area projected back to horizontal. For ventilation and insulation sizing, not a measured floor area.
diagrams	The four drawings for this target, already addressed. Sits beside the areas on purpose: the moment you have a square footage to show, you have the picture to show with it.
faces[]	Every roof plane, with its own area, pitch and edge list.

Edges shared between adjacent faces are counted once per structure, so adding up faces[].edges[] yourself will overcount — use edgeTotals. The category names are ours, not the document's. Raw <LINE> types vary by exporter — the same eave arrives as FASCIA in one dialect and type="EDGE" subtype="EAVE" in another, and a pitch transition as BATTEN — so these nine categories are the normalized form, and they're what the PDF prints too. Parse edgeTotals rather than joining on raw XML strings. Pitches are snapped to a reporting ladder (0, ¼, ½, then whole numbers): measured slopes arrive as values like 10.31, and the PDF and these fields round through the same function so they can't disagree.