H-6617: Create custom graph rendering pipeline inside frontend - #8952
H-6617: Create custom graph rendering pipeline inside frontend#8952indietyp wants to merge 517 commits into
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
1 Skipped Deployment
|
This stack of pull requests is managed by Graphite. Learn more about stacking. |
2fbcafb to
e313dac
Compare
Merging this PR will regress 2 benchmarks
|
| Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|
| ❌ | bit_matrix/dense/union_rows[1000] |
102.4 µs | 119.3 µs | -14.16% |
| ❌ | bit_matrix/sparse/union_rows[1000] |
126.3 µs | 144.8 µs | -12.82% |
| ⚡ | n50000_d256[32] |
1,467.9 ms | 696.5 ms | ×2.1 |
| ⚡ | n10000_d256[128] |
4 s | 1.9 s | ×2.1 |
| ⚡ | n10000_d256[32] |
1,033.5 ms | 517.7 ms | +99.63% |
| ⚡ | n10000_d256[8] |
94.3 ms | 47.9 ms | +96.69% |
| ⚡ | d256[64] |
68.3 µs | 40.7 µs | +67.74% |
| ⚡ | d1536[16] |
111.6 µs | 70.6 µs | +57.92% |
| ⚡ | d3072[16] |
221.2 µs | 140.2 µs | +57.76% |
| ⚡ | d256[16] |
19.5 µs | 12.5 µs | +56.75% |
| ⚡ | d256[15] |
18.6 µs | 12.4 µs | +50.22% |
| ⚡ | kernel/micro_4x2[256] |
4.9 µs | 3.8 µs | +27.89% |
| ⚡ | kernel/micro_4x2[1536] |
23.9 µs | 18.8 µs | +27.01% |
| ⚡ | kernel/micro_4x2[3072] |
46.8 µs | 36.8 µs | +26.99% |
| ⚡ | bit_matrix/dense/iter_row[1000] |
496.4 ns | 440.8 ns | +12.6% |
| ⚡ | kernel/dot[256] |
1.3 µs | 1.2 µs | +10.56% |
Tip
Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.
Comparing bm/h-6617-experiment-with-graph-rendering-inside-of-hash-graph (6c41b64) with main (276e17d)1
Footnotes
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #8952 +/- ##
==========================================
+ Coverage 59.19% 61.51% +2.32%
==========================================
Files 1389 1248 -141
Lines 135001 123443 -11558
Branches 6183 5785 -398
==========================================
- Hits 79910 75936 -3974
+ Misses 54151 46623 -7528
+ Partials 940 884 -56 Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
7c7fadb to
26a8078
Compare
f4af41d to
a4a2e3d
Compare
a4a2e3d to
f91ab64
Compare
f91ab64 to
efb273c
Compare
- clarify that the rank-inverse check is a bounded sample, not a full proof - replace comma splices in schedule and cut docs
- replace each format's own endianness flag with a shared Machine word - bump every file format's layout version by one
- make EdgeRecord/GroupRecord generic over validated row and scalar types - reject out-of-domain values at the file parse instead of index validation
- make ScopeRow, ScopeSchedule::over, and the cut re-export module-private - move their white-box tests into schedule::tests alongside them
- add heap_bytes to the bitset, visibility proof, and scope schedule
- split cache.rs into cache::{scope, tests} and weigh entries by retained memory
- store typed row ids and domain-typed confidences in InstanceRecord - reject out-of-domain confidences at the spool's mapped parse - delete the unchecked constructors from the read-back path
- catch-all now merges per-bucket segments and re-sorts on every read instead of caching a materialized tail - heap_bytes no longer charges for a tail that isn't retained
- stop forcing the saturated memo to build when weighing a small scope - add a per-container allowance and the cache key's inline size to the estimate - saturate the weight fold instead of risking overflow
- compile disjoint column-equality tuples in an Any group into one UNNEST/row-membership predicate instead of one branch per tuple - add owned SQL parameters (SqlParameter, add_owned_parameter) so array literals can be bound in one shot - read_detail moves from query_opt to query_raw and now explicitly rejects a second row
- add capture/verify harness (route_fixture.rs) and the checked-in fixture pair - document the declaration-readback contract in wire.md - add the TypeScript conformance case that decodes it under the served schedule

🌟 What is the purpose of this PR?
🔗 Related links
🚫 Blocked by
🔍 What does this change?
Pre-Merge Checklist 🚀
🚢 Has this modified a publishable library?
This PR:
📜 Does this require a change to the docs?
The changes in this PR:
🕸️ Does this require a change to the Turbo Graph?
The changes in this PR:
turbo.json's have been updated to reflect this🐾 Next steps
🛡 What tests cover this?
❓ How to test this?
📹 Demo