GEOPY-3049: Review capabilities of ferreus_rbf - #60
Conversation
…lities and ferreus vs scipy rbf comparison
github-actions
Bot
changed the title
There was a problem hiding this comment.
🔵 Needs a closer look
The changed artifacts are Git LFS notebook pointers, so the notebook contents can’t be reviewed here and require manual inspection/execution to validate.
Pull request overview
Adds/updates Git LFS–tracked Jupyter notebooks under surface_apps-assets/ferreus/ to investigate ferreus_rbf capabilities, including comparisons against SciPy RBF and some real-data/geoh5 POC exploration.
Changes:
- Added a new
ferreus_rbfvs SciPy RBF comparison notebook. - Updated existing notebooks used for geochemistry comparison and geoh5 POC investigation (LFS objects updated).
File summaries
| File | Description |
|---|---|
| surface_apps-assets/ferreus/ferreus_rmt_real_geochem_comparison.ipynb | Updates the LFS-tracked notebook used for real geochemistry comparison. |
| surface_apps-assets/ferreus/ferreus_rmt_geoh5_poc.ipynb | Updates the LFS-tracked notebook used for a geoh5 proof-of-concept workflow. |
| surface_apps-assets/ferreus/ferreus_rbf_scipy_comparison.ipynb | Adds a new LFS-tracked notebook comparing ferreus_rbf and SciPy RBF. |
Review details
- Files reviewed: 3/3 changed files
- Comments generated: 3
- Review effort level: Lite
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
| version https://git-lfs.github.com/spec/v1 | ||
| oid sha256:82ae2d99c4ae8061c0f1b32d2054385221acc8be05f20b25a9fe4cdd77e5f885 | ||
| size 32367 |
| version https://git-lfs.github.com/spec/v1 | ||
| oid sha256:4643cbf88711cda598ec36e191a2284466a3ee6fe5e07668253bd2ba383709fd | ||
| size 947791 | ||
| oid sha256:759ed1449688380bd3544cf03b1d9f2888f8f4496ea09e9422debe8dfdcf77e3 | ||
| size 56727 |
| version https://git-lfs.github.com/spec/v1 | ||
| oid sha256:cbf1eeb37496c29e7f9c2c494f6905036da3ff77e4a0e4e0e6a7d7fd406a1224 | ||
| size 38170 | ||
| oid sha256:78760f3bda845a93fa168a2c20eeaa20901cb04a9523fce44d81204833285c92 | ||
| size 39514 |
There was a problem hiding this comment.
Correct me if I am wrong, but the main outcomes of the analysis are
- Both scipy and ferrous compute very similar solutions
- Ferrous becomes much cheaper than scipy as the number of inputs increases
Can you expand the tests to a larger set beyond 4000, just so we confirm the deviation between the two algos
Also, can you add the test for peak memory? Since the solutions are very similar, hooking up to ferreus mainly becomes a compute/memory play, so would be good to know before we go forward.
GEOPY-3049 - Review capabilities of ferreus_rbf
add notebook specifically for the investigation of ferreus_rbf capabilities and ferreus vs scipy rbf comparison