Boostrap new project

.gitignore

@@ -9,3 +9,4 @@ __pycache__/
 *.egg-info/
 dist/
 build/
+.idea/

Makefile

@@ -1,6 +1,8 @@
 PYTHON ?= .venv/bin/python
+# override PYTHON to use an active interpreter, e.g. `PYTHON=$(which python) make ...`
 SYSTEM_PYTHON ?= python3
 PROJECT ?= glioblastoma_perplexity_manual
+SKIP_BAD ?= 1
 MPLCONFIGDIR ?= $(PWD)/.mpl-cache
 XDG_CACHE_HOME ?= $(PWD)/.cache
 ENV_VARS = MPLCONFIGDIR=$(MPLCONFIGDIR) XDG_CACHE_HOME=$(XDG_CACHE_HOME)
@@ -16,7 +18,7 @@ cache_dirs:
 
 .PHONY: data
 data: cache_dirs
-	$(ENV_VARS) $(PYTHON) -m src.process_deepsearch --project $(PROJECT)
+	$(ENV_VARS) $(PYTHON) -m src.process_deepsearch --project $(PROJECT) $(if $(filter 1 true True,$(SKIP_BAD)),--skip-bad,)
 
 .PHONY: comparisons
 comparisons: data cache_dirs

README.md

@@ -17,6 +17,15 @@ Pipeline for parsing Perplexity DeepSearch outputs, comparing pseudo-enrichment
 PROJECT=glioblastoma_perplexity_manual make master_report
 make test         # run pytest
 make coverage     # pytest with coverage report
+
+# New project scaffold
+python -m src.init_project --project my_project
+# then fill projects/my_project/description.md and geneset_folder_mapping.csv,
+# add runs under deepsearch/my_project/, comparisons under Comparisons/my_project/,
+# and run with PROJECT=my_project make master_report
+
+# If run markdowns are malformed, you can skip bad ones:
+PROJECT=my_project make data SKIP_BAD=1  # or python -m src.process_deepsearch --project my_project --skip-bad
 ```
 Targets: `data` (parse runs), `comparisons` (parse GO tables), `figures`, `heatmaps`, `run_reports`, `master_report`. Environment variables `MPLCONFIGDIR` and `XDG_CACHE_HOME` default to repo-local caches.