Spaces:
Running on Zero
Running on Zero
Delete CLAUDE.md
Browse files
CLAUDE.md
DELETED
|
@@ -1,31 +0,0 @@
|
|
| 1 |
-
### Project deliverables β instructions/project.md
|
| 2 |
-
|
| 3 |
-
### CURRENT STATUS
|
| 4 |
-
|
| 5 |
-
**Phase 1 - Data Fusion & Graph Construction: β
COMPLETE**
|
| 6 |
-
|
| 7 |
-
- β
Data downloads (Hub'Eau hydrometric, ERA5 meteorological, elevations)
|
| 8 |
-
- β
Data loaders (hydrometric.py, safran.py) - aligned with actual API structures
|
| 9 |
-
- β
Fusion pipeline (fusion_engine.py) - merges Hub'Eau + ERA5
|
| 10 |
-
- β
Graph construction (graph_from_fused.py) - PyTorch Geometric compatible
|
| 11 |
-
- β
Graph visualization (visualize.py) - 7 plot types
|
| 12 |
-
|
| 13 |
-
**Working Pipeline:**
|
| 14 |
-
```
|
| 15 |
-
Hub'Eau (53K obs) + ERA5 (10K obs) β Fusion β Graphs β Visualization β Ready for GNN
|
| 16 |
-
```
|
| 17 |
-
|
| 18 |
-
**Datasets:**
|
| 19 |
-
- Hub'Eau: 101,218 observations, 14 stations (1969-2026)
|
| 20 |
-
- ERA5: Sample data (Jan 2020), full download ready (1960-2026, 67 years)
|
| 21 |
-
- Station elevations: 27 stations from EU-DEM
|
| 22 |
-
- ADES groundwater: 113 wells, 272K measurements
|
| 23 |
-
- IDPR KARST indicators: 27 stations
|
| 24 |
-
|
| 25 |
-
**Next:** GNN model training or full ERA5 download
|
| 26 |
-
|
| 27 |
-
### INSTRUCTIONS
|
| 28 |
-
- Role: Elite ML engineer, efficient, direct, best practices
|
| 29 |
-
- Focus: Graph construction for PyTorch Geometric compatible with spatial-temporal data
|
| 30 |
-
- Datasets: Located in `datasets/`
|
| 31 |
-
- Code: All implementation in `src/`
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|