RYDE-WORK/lnp_ml
1
0
Fork 0
mirror of https://github.com/RYDE-WORK/lnp_ml.git synced 2026-03-21 09:36:32 +08:00
Code Issues Packages Projects Releases Wiki Activity

整理Makefile结构

Browse Source
This commit is contained in:
RYDE-WORK 2026-02-26 18:06:07 +08:00
parent 0e917ef0d4
commit 72c292a91f
1 changed files with 68 additions and 83 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

151
Makefile
View File

@ -6,18 +6,33 @@ PROJECT_NAME = lnp-ml
PYTHON_VERSION = 3.8 PYTHON_VERSION = 3.8
PYTHON_INTERPRETER = python PYTHON_INTERPRETER = python
################################################################################# # --- CLI flag 变量 ---
# COMMANDS # MPNN_FLAG = $(if $(USE_MPNN),--use-mpnn,)
################################################################################# FREEZE_FLAG = $(if $(FREEZE_BACKBONE),--freeze-backbone,)
DEVICE_FLAG = $(if $(DEVICE),--device $(DEVICE),)
SCAFFOLD_SPLIT_FLAG = $(if $(filter 1,$(SCAFFOLD_SPLIT)),--scaffold-split,)
SEED_FLAG = $(if $(SEED),--seed $(SEED),)
N_TRIALS_FLAG = $(if $(N_TRIALS),--n-trials $(N_TRIALS),)
EPOCHS_PER_TRIAL_FLAG = $(if $(EPOCHS_PER_TRIAL),--epochs-per-trial $(EPOCHS_PER_TRIAL),)
MIN_STRATUM_FLAG = $(if $(MIN_STRATUM_COUNT),--min-stratum-count $(MIN_STRATUM_COUNT),)
OUTPUT_DIR_FLAG = $(if $(OUTPUT_DIR),--output-dir $(OUTPUT_DIR),)
USE_SWA_FLAG = $(if $(USE_SWA),--use-swa,)
INIT_PRETRAIN_FLAG = $(if $(NO_PRETRAIN),,--init-from-pretrain $(or $(INIT_PRETRAIN),models/pretrain_delivery.pt))
#################################################################################
# ENVIRONMENT & CODE QUALITY #
#################################################################################
## Install Python dependencies ## Install Python dependencies
.PHONY: requirements .PHONY: requirements
requirements: requirements:
pixi install pixi install
## Set up Python interpreter environment
.PHONY: create_environment
create_environment:
@echo ">>> Pixi environment will be created when running 'make requirements'"
@echo ">>> Activate with:\npixi shell"
## Delete all compiled Python files ## Delete all compiled Python files
.PHONY: clean .PHONY: clean
@ -25,7 +40,6 @@ clean:
find . -type f -name "*.py[co]" -delete find . -type f -name "*.py[co]" -delete
find . -type d -name "__pycache__" -delete find . -type d -name "__pycache__" -delete
## Lint using ruff (use `make format` to do formatting) ## Lint using ruff (use `make format` to do formatting)
.PHONY: lint .PHONY: lint
lint: lint:
@ -38,26 +52,10 @@ format:
ruff check --fix ruff check --fix
ruff format ruff format
## Set up Python interpreter environment
.PHONY: create_environment
create_environment:
@echo ">>> Pixi environment will be created when running 'make requirements'"
@echo ">>> Activate with:\npixi shell"
################################################################################# #################################################################################
# PROJECT RULES # # DATA PROCESSING #
################################################################################# #################################################################################
## Preprocess internal data (raw -> interim) ## Preprocess internal data (raw -> interim)
.PHONY: preprocess .PHONY: preprocess
preprocess: requirements preprocess: requirements
@ -85,44 +83,24 @@ data_pretrain_cv: requirements
## Process internal data with CV splitting (interim -> processed/cv) ## Process internal data with CV splitting (interim -> processed/cv)
## Use SCAFFOLD_SPLIT=1 to enable amine-based scaffold splitting (default: random shuffle) ## Use SCAFFOLD_SPLIT=1 to enable amine-based scaffold splitting (default: random shuffle)
SCAFFOLD_SPLIT_FLAG = $(if $(filter 1,$(SCAFFOLD_SPLIT)),--scaffold-split,)
.PHONY: data_cv .PHONY: data_cv
data_cv: requirements data_cv: requirements
$(PYTHON_INTERPRETER) scripts/process_data_cv.py $(SCAFFOLD_SPLIT_FLAG) $(PYTHON_INTERPRETER) scripts/process_data_cv.py $(SCAFFOLD_SPLIT_FLAG)
# MPNN 支持：使用 USE_MPNN=1 启用 MPNN encoder #################################################################################
# 例如：make pretrain USE_MPNN=1 # TRAINING #
MPNN_FLAG = $(if $(USE_MPNN),--use-mpnn,) #################################################################################
# Backbone 冻结：使用 FREEZE_BACKBONE=1 冻结 backbone，只训练 heads
# 例如：make finetune FREEZE_BACKBONE=1
FREEZE_FLAG = $(if $(FREEZE_BACKBONE),--freeze-backbone,)
# 设备选择：使用 DEVICE=xxx 指定设备
# 例如：make train DEVICE=cuda:0 或 make test_cv DEVICE=mps
DEVICE_FLAG = $(if $(DEVICE),--device $(DEVICE),)
## Pretrain on external data (delivery only) ## Pretrain on external data (delivery only)
.PHONY: pretrain .PHONY: pretrain
pretrain: requirements pretrain: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.pretrain main $(MPNN_FLAG) $(DEVICE_FLAG) $(PYTHON_INTERPRETER) -m lnp_ml.modeling.pretrain main $(MPNN_FLAG) $(DEVICE_FLAG)
## Evaluate pretrain model (delivery metrics)
.PHONY: test_pretrain
test_pretrain: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.pretrain test $(MPNN_FLAG) $(DEVICE_FLAG)
## Pretrain with cross-validation (5-fold) ## Pretrain with cross-validation (5-fold)
.PHONY: pretrain_cv .PHONY: pretrain_cv
pretrain_cv: requirements pretrain_cv: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.pretrain_cv main $(MPNN_FLAG) $(DEVICE_FLAG) $(PYTHON_INTERPRETER) -m lnp_ml.modeling.pretrain_cv main $(MPNN_FLAG) $(DEVICE_FLAG)
## Evaluate CV pretrain models on test sets (auto-detects MPNN from checkpoint)
.PHONY: test_pretrain_cv
test_pretrain_cv: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.pretrain_cv test $(DEVICE_FLAG)
## Train model (multi-task, from scratch) ## Train model (multi-task, from scratch)
.PHONY: train .PHONY: train
train: requirements train: requirements
@ -143,48 +121,64 @@ train_final: requirements
--init-from-pretrain models/pretrain_delivery.pt \ --init-from-pretrain models/pretrain_delivery.pt \
$(FREEZE_FLAG) $(MPNN_FLAG) $(DEVICE_FLAG) $(FREEZE_FLAG) $(MPNN_FLAG) $(DEVICE_FLAG)
## Finetune with cross-validation on internal data (5-fold, amine-based split) with pretrained weights
.PHONY: finetune_cv
finetune_cv: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.train_cv main --init-from-pretrain models/pretrain_delivery.pt $(FREEZE_FLAG) $(MPNN_FLAG) $(DEVICE_FLAG)
## Train with cross-validation on internal data only (5-fold, amine-based split) ## Train with cross-validation on internal data only (5-fold, amine-based split)
.PHONY: train_cv .PHONY: train_cv
train_cv: requirements train_cv: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.train_cv main $(FREEZE_FLAG) $(MPNN_FLAG) $(DEVICE_FLAG) $(PYTHON_INTERPRETER) -m lnp_ml.modeling.train_cv main $(FREEZE_FLAG) $(MPNN_FLAG) $(DEVICE_FLAG)
## Finetune with cross-validation on internal data (5-fold) with pretrained weights
.PHONY: finetune_cv
finetune_cv: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.train_cv main --init-from-pretrain models/pretrain_delivery.pt $(FREEZE_FLAG) $(MPNN_FLAG) $(DEVICE_FLAG)
#################################################################################
# EVALUATION #
#################################################################################
## Evaluate pretrain model (delivery metrics)
.PHONY: test_pretrain
test_pretrain: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.pretrain test $(MPNN_FLAG) $(DEVICE_FLAG)
## Evaluate CV pretrain models on test sets (auto-detects MPNN from checkpoint)
.PHONY: test_pretrain_cv
test_pretrain_cv: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.pretrain_cv test $(DEVICE_FLAG)
## Evaluate CV finetuned models on test sets (auto-detects MPNN from checkpoint) ## Evaluate CV finetuned models on test sets (auto-detects MPNN from checkpoint)
.PHONY: test_cv .PHONY: test_cv
test_cv: requirements test_cv: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.train_cv test $(DEVICE_FLAG) $(PYTHON_INTERPRETER) -m lnp_ml.modeling.train_cv test $(DEVICE_FLAG)
## Test model on test set (with detailed metrics, auto-detects MPNN from checkpoint)
.PHONY: test
test: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.predict test $(DEVICE_FLAG)
## Run predictions
.PHONY: predict
predict: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.predict $(DEVICE_FLAG)
#################################################################################
# HYPERPARAMETER TUNING #
#################################################################################
# 通用参数：
# SEED 随机种子 (默认: 42)
# N_TRIALS Optuna 试验数 (默认: 20)
# EPOCHS_PER_TRIAL 每个试验的最大 epoch (默认: 30)
# MIN_STRATUM_COUNT 复合分层标签的最小样本数 (默认: 5)
# OUTPUT_DIR 输出目录 (根据命令有不同默认值)
# INIT_PRETRAIN 预训练权重路径 (默认: models/pretrain_delivery.pt)
# NO_PRETRAIN=1 禁用预训练权重
## Train with hyperparameter tuning ## Train with hyperparameter tuning
.PHONY: tune .PHONY: tune
tune: requirements tune: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.train --tune $(MPNN_FLAG) $(DEVICE_FLAG) $(PYTHON_INTERPRETER) -m lnp_ml.modeling.train --tune $(MPNN_FLAG) $(DEVICE_FLAG)
# ============ 嵌套 CV + Optuna 调参（StratifiedKFold + 类权重） ============
# 通用参数：
# SEED: 随机种子 (默认: 42)
# N_TRIALS: Optuna 试验数 (默认: 20)
# EPOCHS_PER_TRIAL: 每个试验的最大 epoch (默认: 30)
# MIN_STRATUM_COUNT: 复合分层标签的最小样本数 (默认: 5)
# OUTPUT_DIR: 输出目录 (根据命令有不同默认值)
# INIT_PRETRAIN: 预训练权重路径 (默认: models/pretrain_delivery.pt)
SEED_FLAG = $(if $(SEED),--seed $(SEED),)
N_TRIALS_FLAG = $(if $(N_TRIALS),--n-trials $(N_TRIALS),)
EPOCHS_PER_TRIAL_FLAG = $(if $(EPOCHS_PER_TRIAL),--epochs-per-trial $(EPOCHS_PER_TRIAL),)
MIN_STRATUM_FLAG = $(if $(MIN_STRATUM_COUNT),--min-stratum-count $(MIN_STRATUM_COUNT),)
OUTPUT_DIR_FLAG = $(if $(OUTPUT_DIR),--output-dir $(OUTPUT_DIR),)
USE_SWA_FLAG = $(if $(USE_SWA),--use-swa,)
# 默认使用预训练权重，设置 NO_PRETRAIN=1 可禁用
INIT_PRETRAIN_FLAG = $(if $(NO_PRETRAIN),,--init-from-pretrain $(or $(INIT_PRETRAIN),models/pretrain_delivery.pt))
## Nested CV with Optuna: outer 5-fold (test) + inner 3-fold (tune) ## Nested CV with Optuna: outer 5-fold (test) + inner 3-fold (tune)
## 用于模型评估：外层 5-fold 产生无偏性能估计，内层 3-fold 做超参搜索 ## 用于模型评估：外层 5-fold 产生无偏性能估计，内层 3-fold 做超参搜索
## 默认加载 models/pretrain_delivery.pt 预训练权重，使用 NO_PRETRAIN=1 禁用
## 使用示例: make nested_cv_tune DEVICE=cuda N_TRIALS=30 ## 使用示例: make nested_cv_tune DEVICE=cuda N_TRIALS=30
.PHONY: nested_cv_tune .PHONY: nested_cv_tune
nested_cv_tune: requirements nested_cv_tune: requirements
@ -194,7 +188,6 @@ nested_cv_tune: requirements
## Final training with Optuna: 3-fold CV tune + full data train ## Final training with Optuna: 3-fold CV tune + full data train
## 用于最终模型训练：3-fold 调参后用全量数据训练（无 early-stop） ## 用于最终模型训练：3-fold 调参后用全量数据训练（无 early-stop）
## 默认加载 models/pretrain_delivery.pt 预训练权重，使用 NO_PRETRAIN=1 禁用
## 使用示例: make final_optuna DEVICE=cuda N_TRIALS=30 USE_SWA=1 ## 使用示例: make final_optuna DEVICE=cuda N_TRIALS=30 USE_SWA=1
.PHONY: final_optuna .PHONY: final_optuna
final_optuna: requirements final_optuna: requirements
@ -202,15 +195,9 @@ final_optuna: requirements
$(DEVICE_FLAG) $(MPNN_FLAG) $(SEED_FLAG) $(INIT_PRETRAIN_FLAG) \ $(DEVICE_FLAG) $(MPNN_FLAG) $(SEED_FLAG) $(INIT_PRETRAIN_FLAG) \
$(N_TRIALS_FLAG) $(EPOCHS_PER_TRIAL_FLAG) $(MIN_STRATUM_FLAG) $(OUTPUT_DIR_FLAG) $(USE_SWA_FLAG) $(N_TRIALS_FLAG) $(EPOCHS_PER_TRIAL_FLAG) $(MIN_STRATUM_FLAG) $(OUTPUT_DIR_FLAG) $(USE_SWA_FLAG)
## Run predictions #################################################################################
.PHONY: predict # SERVING & DEPLOYMENT #
predict: requirements #################################################################################
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.predict $(DEVICE_FLAG)
## Test model on test set (with detailed metrics, auto-detects MPNN from checkpoint)
.PHONY: test
test: requirements
$(PYTHON_INTERPRETER) -m lnp_ml.modeling.predict test $(DEVICE_FLAG)
## Formulation optimization: find optimal LNP formulation for target organ ## Formulation optimization: find optimal LNP formulation for target organ
## Usage: make optimize SMILES="CC(C)..." ORGAN=liver ## Usage: make optimize SMILES="CC(C)..." ORGAN=liver
@ -237,9 +224,8 @@ serve:
@echo "" @echo ""
@echo "然后访问: http://localhost:8501" @echo "然后访问: http://localhost:8501"
################################################################################# #################################################################################
# DOCKER COMMANDS # # DOCKER #
################################################################################# #################################################################################
## Build Docker images ## Build Docker images
@ -279,7 +265,6 @@ docker-clean:
docker compose down -v --rmi local docker compose down -v --rmi local
docker system prune -f docker system prune -f
################################################################################# #################################################################################
# Self Documenting Commands # # Self Documenting Commands #
################################################################################# #################################################################################