ljiro/soilscan

How ljiro/soilscan works

SoilScan combines a mobile soil-analysis app with a Python NPK model-training and analysis pipeline. The user-facing flow is photo capture or upload, model inference, then a soil classification and fertilizer guidance; the README describes the inference backend as Hugging Face Spaces, while this tree contains only selected app screens, not that backend or the classifier screen implementation. The training flow covers configured data preparation, model fitting, evaluation, and embedding-similarity analysis. Edges involving unsampled app behavior rely on README statements; code-level links use the shown source excerpts or SOURCE INDEX.

Main entry points: config.py Experiment config · dataset.py Dataset preparation · model.py Model architectures · train.py Model training

Components 16

Mobile app

Soil scanner SoilScan/screens
Fertilizer guidance SoilScan/screens/FertilizerRecommendationScreen.js
Soil guide SoilScan/screens/GuideScreen.js

Training pipeline

UUID label maps datamaps
Dataset preparation train/dataset.py
Model architectures train/model.py
Model training train/train.py
Training metrics train/metrics.py

Model analysis

Batch inference train/inference.py
Checkpoint evaluation train/evaluate.py
Embedding similarity train/similarity_matrix.py
NPK combo similarity train/combo_similarity.py

Experiment setup

Experiment config train/config.py
Experiment presets experiments

Other components

SoilScan user
Inference backend

Connections 28

SoilScan user → Soil scanner: submits soil photo README.md
Soil scanner → Inference backend: requests inference README.md
Inference backend → Soil scanner: returns classification README.md
Soil scanner → Fertilizer guidance: opens guidance
SoilScan user → Fertilizer guidance: enters field values README.md
SoilScan user → Soil guide: views guidance
UUID label maps → Dataset preparation: provides labels
Experiment config → Dataset preparation: configures loading train/dataset.py
Experiment config → Model architectures: configures model train/model.py
Dataset preparation → Model training: builds loaders train/train.py
Model architectures → Model training: builds network train/train.py
Model training → Training metrics: computes metrics train/train.py
Experiment config → Model training: sets run parameters train/train.py
Experiment config → Batch inference: configures prediction train/inference.py
Dataset preparation → Batch inference: supplies transforms train/inference.py
Model architectures → Batch inference: builds predictor train/inference.py
Experiment config → Checkpoint evaluation: configures evaluation train/evaluate.py
Dataset preparation → Checkpoint evaluation: builds test loader train/evaluate.py
Model architectures → Checkpoint evaluation: builds evaluation model train/evaluate.py
Model training → Checkpoint evaluation: provides checkpoint train/evaluate.py
Training metrics → Checkpoint evaluation: scores predictions train/evaluate.py
Experiment config → Embedding similarity: configures analysis train/similarity_matrix.py
Dataset preparation → Embedding similarity: supplies samples train/similarity_matrix.py
Model architectures → Embedding similarity: extracts embeddings train/similarity_matrix.py
Experiment config → NPK combo similarity: configures analysis train/combo_similarity.py
Dataset preparation → NPK combo similarity: supplies samples train/combo_similarity.py
Model architectures → NPK combo similarity: extracts embeddings train/combo_similarity.py
Experiment presets → Experiment config: provides presets

More Python diagrams: zhaoyang97/paper-notes · microsoft/pylance-release · jmfernandes/robin_stocks · rupeshs/fastsdcpu · cfzd/ultra-fast-lane-detection

Drawn by GitDiagram on October 3, 2026 from the repository's file tree, README and source files. Read as Markdown ·