mirror of
https://github.com/IfcOpenShell/IfcOpenShell.git
synced 2026-09-10 22:16:41 +00:00
0b8c787ac0
The .ifcview data is hugely redundant (repeated double instance matrices, patterned indices) — measured 12x zstd whole-file. Server Content-Encoding can't be used (it breaks HTTP Range), so compress PER-CHUNK into the format. Format (v16): geometry becomes per-chunk zstd(vertices)+zstd(indices) frames — each independently Range-fetchable, so streaming is intact — and the critical + deferred metadata blocks are single zstd frames. SidecarChunk carries the compressed blob offsets/sizes; applyStreamedChunk (render/upload) is UNCHANGED — decompression slots into the fetch. Full readSidecar (test/tooling) reconstructs by decompress+scatter. zstd: desktop links libzstd (also compresses at bake); the web build (Emscripten has no zstd port) FetchContent's the pinned zstd source and compiles its decompress-only subset for wasm — no vendored blob, same version as desktop. New SidecarCompress wraps it (compress guarded off under Emscripten). Both stream paths — desktop StreamingThread worker + sync fallback (readChunkGeometryCompressed) and web beginWebChunkLoad — decompress; readSidecarMetadataOnly / the web bootstrap / loadDeferredMetadataWeb decompress the metadata blocks. streamingByteProgress reports COMPRESSED bytes. MEASURED: a 752 MB v15 federation → 75 MB v16 (10x; per-file 6.7-15.3x); PP-PLP 118→15 MB, loads 13/13 chunks on web, 0 errors. Three fixes found while testing big federations on a real server: - Web-streamed race: streaming_from_web was set in the deferred-header callback (a round-trip after the model+chunks exist), so driveStreamingLoads could take the sync fopen path meanwhile → "failed to read/decompress chunk 0". Now set immediately after applyCachedModel. - OOM abort on 18 models: the pool grew unbounded until an alloc failed, but on web that's an uncatchable bad_alloc abort. Cap total pool capacity (setMaxTotalCapacity, 3 GB) so it stops before the heap ceiling, and raise MAXIMUM_MEMORY 2→4 GB (wasm32 max) for headroom. - Web never evicted (grow-or-block only). At the hard budget, fall through to the LRU/priority evictor so a big federation stays navigable (highest-contribution chunks win) instead of freezing with holes. 113/113 desktop + 6/6 web smoke pass. No back-compat: regenerate sidecars (desktop bakes v16; scratch conv tool migrates v15→v16). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
172 lines
9.0 KiB
C++
172 lines
9.0 KiB
C++
/********************************************************************************
|
|
* *
|
|
* This file is part of IfcOpenShell. *
|
|
* *
|
|
* IfcOpenShell is free software: you can redistribute it and/or modify *
|
|
* it under the terms of the Lesser GNU General Public License as published by *
|
|
* the Free Software Foundation, either version 3.0 of the License, or *
|
|
* (at your option) any later version. *
|
|
* *
|
|
* IfcOpenShell is distributed in the hope that it will be useful, *
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of *
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the *
|
|
* Lesser GNU General Public License for more details. *
|
|
* *
|
|
* You should have received a copy of the Lesser GNU General Public License *
|
|
* along with this program. If not, see <http://www.gnu.org/licenses/>. *
|
|
* *
|
|
********************************************************************************/
|
|
|
|
// NOTE: Sidecar format v3 is being rewritten to v4 (instanced geometry layout).
|
|
// During the instancing rewrite (Commit A) the cache is a no-op: reads always
|
|
// miss and writes always succeed without producing a file. Commit B will
|
|
// re-introduce the on-disk format with MeshInfo + InstanceGpu sections.
|
|
|
|
#ifndef SIDECARCACHE_H
|
|
#define SIDECARCACHE_H
|
|
|
|
#include "InstancedGeometry.h"
|
|
|
|
#include <cstdint>
|
|
#include <optional>
|
|
#include <string>
|
|
#include <vector>
|
|
#include <memory>
|
|
|
|
static constexpr uint32_t SIDECAR_MAGIC = 0x49465657; // "IFVW"
|
|
// v5 = MeshInfo extended with lod1_ebo_byte_offset + lod1_index_count (56 B).
|
|
// sd.indices may contain an appended LOD1 index slice for each mesh
|
|
// where meshoptimizer decimation produced useful output.
|
|
// v6 = VBO vertices quantized to 16 B/vertex (pos u16x3 + normal oct i16x2 +
|
|
// color u8x4). Dequant basis is per-mesh MeshInfo.local_aabb_min/max.
|
|
// v7 = VBO vertices shrunk to 12 B/vertex (normal oct i8x2 replaces i16x2,
|
|
// eliminating 2-byte pad + saving 2 bytes on normal).
|
|
// v8 = source_file_size field dropped from header. Sidecar is keyed purely
|
|
// on path stem (foo.ifc and foo.ifcdb/ both map to foo.ifcview) so the
|
|
// same cache serves either source format. Staleness is user-managed
|
|
// (delete the sidecar to force a rebuild).
|
|
// v9 = unused `reserved` field dropped from header (16 B -> 12 B).
|
|
// v10 = InstanceCpu gains placement_transformation[16] alongside transform[16]
|
|
// — record grew from 104 B to 168 B. placement_transformation is the
|
|
// raw streamer output; transform is the composed FederatedFalseOrigin ·
|
|
// ModelTransformation · CoordinateOperation · placement_transformation
|
|
// result. Sidecar serialises both; on load the transform is recomputed
|
|
// from placement_transformation + the ViewportWindow's current stage
|
|
// matrices, so v10 sidecars are reusable across .ifcfeds.
|
|
// v11 = SidecarData gains a per-model CoordinateOperation cache:
|
|
// coordinate_operation_meters[16] (column-major), unit scales, and a
|
|
// has_coordinate_operation flag. Lets sidecar-loaded models apply
|
|
// georef without re-parsing the IFC source. Edits to the IFC's
|
|
// IfcMapConversion do NOT invalidate the sidecar — delete the
|
|
// .ifcview manually if you change the source's georef parameters.
|
|
// v12 = InstanceCpu::placement_transformation is double[16], and
|
|
// InstanceChunk carries the streamer placement as double[16]. This keeps
|
|
// large IFC placements exact until CoordinateOperation / FederatedFalseOrigin
|
|
// composition has reduced them to viewport-local float-sized values.
|
|
// v13 = Map unit scale in cached ModelGeoref is derived from
|
|
// IfcMapConversion.Scale, not IfcProjectedCRS.MapUnit.
|
|
// v14 = Geometry is laid out in streaming-chunk order (SidecarLayout) and a
|
|
// chunk table-of-contents (`chunks`) is appended. The loader builds its
|
|
// chunks from the TOC instead of re-deriving the Morton/greedy plan, so
|
|
// each chunk is one CONTIGUOUS byte range — fixing network read
|
|
// amplification. The plan can't be re-derived at load because the float
|
|
// Morton quantisation isn't bit-identical across toolchains (x86 baker vs
|
|
// wasm loader), so it must be baked in. No back-compat: v13 sidecars are
|
|
// rejected (regenerate them).
|
|
// v15 = The post-index metadata is split into a render-CRITICAL block (meshes,
|
|
// instances, georef, chunk TOC) followed by a DEFERRED block (elements +
|
|
// string_table — the IFC element tree, used for UI/picking, never for
|
|
// rendering), with the critical block's byte length written just after
|
|
// the index section. The web loader reads only the critical block before
|
|
// painting, so first geometry no longer waits on the property data; the
|
|
// deferred block is fetched lazily (or skipped where unused). Desktop
|
|
// reads both. No back-compat: regenerate sidecars.
|
|
// v16 = Geometry + metadata are zstd-COMPRESSED. Each chunk's vertex bytes and
|
|
// index bytes are stored as two independent zstd frames (so per-chunk
|
|
// Range streaming still works — you fetch + decompress just one chunk),
|
|
// and the critical + deferred metadata blocks are single zstd frames.
|
|
// The chunk TOC records each chunk's compressed blob offsets/sizes plus
|
|
// the raw (decompressed) sizes. ~3-5x fewer bytes over the wire while
|
|
// keeping HTTP Range intact (unlike server Content-Encoding). No
|
|
// back-compat: regenerate sidecars.
|
|
static constexpr uint32_t SIDECAR_VERSION = 16;
|
|
static constexpr uint32_t SIDECAR_ENDIAN = 0x01020304;
|
|
|
|
// Chunk table-of-contents entry (v16). A chunk is a CONTIGUOUS range of meshes
|
|
// [first_mesh, first_mesh + mesh_count). Its vertex + index bytes are stored as
|
|
// two zstd frames in the geometry section; the loader fetches [v_comp_off,
|
|
// +v_comp_size) / [i_comp_off, +i_comp_size) (offsets relative to the geometry
|
|
// section start) and decompresses them to v_raw_size / i_raw_size bytes — the
|
|
// chunk-local (vbytes, idx) applyStreamedChunk consumes.
|
|
struct SidecarChunk {
|
|
uint32_t first_mesh;
|
|
uint32_t mesh_count;
|
|
uint64_t v_comp_off;
|
|
uint64_t v_comp_size;
|
|
uint64_t v_raw_size;
|
|
uint64_t i_comp_off;
|
|
uint64_t i_comp_size;
|
|
uint64_t i_raw_size;
|
|
};
|
|
|
|
// Fixed-size element record. Strings are stored as (offset, length) pairs
|
|
// into a separate string table.
|
|
struct PackedElementInfo {
|
|
uint32_t object_id;
|
|
uint32_t model_id;
|
|
int32_t ifc_id;
|
|
int32_t parent_id;
|
|
uint32_t guid_offset;
|
|
uint32_t guid_length;
|
|
uint32_t name_offset;
|
|
uint32_t name_length;
|
|
uint32_t type_offset;
|
|
uint32_t type_length;
|
|
};
|
|
|
|
// Everything needed to display an already-tessellated model without
|
|
// re-running the iterator. v6 schema: instanced + quantized geometry.
|
|
struct SidecarData {
|
|
// Per-model GPU geometry (local coords). Raw VBO bytes at the
|
|
// INSTANCED_VERTEX_STRIDE_BYTES layout (12 B/vertex as of v7).
|
|
std::vector<uint8_t> vertices;
|
|
std::vector<uint32_t> indices;
|
|
|
|
// Mesh dictionary and per-instance data.
|
|
std::vector<MeshInfo> meshes; // indexed by local_mesh_id
|
|
std::vector<InstanceCpu> instances; // sorted by mesh_id
|
|
|
|
// CoordinateOperation cache (v11+). Mirrors ModelGeoref so a sidecar
|
|
// load can apply georef without re-parsing the IFC source.
|
|
// has_coordinate_operation == 0 means the model has no
|
|
// IfcMapConversion; the matrix is then the identity placeholder.
|
|
double coordinate_operation_meters[16] = {
|
|
1, 0, 0, 0,
|
|
0, 1, 0, 0,
|
|
0, 0, 1, 0,
|
|
0, 0, 0, 1 };
|
|
double project_length_to_meters = 1.0;
|
|
double map_unit_to_meters = 1.0;
|
|
uint32_t has_coordinate_operation = 0;
|
|
|
|
// Element tree metadata.
|
|
std::vector<PackedElementInfo> elements;
|
|
std::string string_table;
|
|
|
|
// Streaming chunk TOC. Always written on disk (v14); geometry is laid out
|
|
// in this chunk order (see SidecarLayout) so each chunk is one contiguous
|
|
// range and the loader builds chunks directly from it. Stays empty only for
|
|
// in-memory direct loads (finalizeModel), which don't stream and fall back
|
|
// to deriving the plan.
|
|
std::vector<SidecarChunk> chunks;
|
|
};
|
|
|
|
// Sidecar is keyed on the path stem: foo.ifc and foo.ifcdb/ both resolve to
|
|
// foo.ifcview alongside the source. No staleness check — callers delete the
|
|
// file to invalidate.
|
|
bool writeSidecar(const std::string& ifc_path, const SidecarData& data);
|
|
|
|
std::optional<SidecarData> readSidecar(const std::string& ifc_path);
|
|
|
|
#endif // SIDECARCACHE_H
|