make makefile to run test locally

Author: ollmer

.github/workflows/darglint.yml

@@ -31,4 +31,4 @@ jobs:
         run: pip list
 
       - name: Darglint checks
-        run: darglint -v 2 -z short .
+        run: darglint -v 2 -z short src/

.gitignore

@@ -168,4 +168,6 @@ results/
 # gradio
 .gradio/
 
-outputs/
+outputs/
+miniwob-plusplus/
+.miniwob-server.pid

Makefile

@@ -0,0 +1,22 @@
+.PHONY: test setup miniwob lint
+setup:
+	@pip install darglint black
+	@pip install -e .
+	@playwright install chromium --with-deps
+	@python -c 'import nltk; nltk.download("punkt_tab")'
+
+miniwob:
+	@kill -9 `cat .miniwob-server.pid` || true
+	@git clone https://github.com/Farama-Foundation/miniwob-plusplus.git || true
+	@cd miniwob-plusplus && git checkout 7fd85d71a4b60325c6585396ec4f48377d049838
+	@python -m http.server 8080 --directory miniwob-plusplus/miniwob/html & echo $$! > .miniwob-server.pid
+	@sleep 2
+	@echo "MiniWob server started on port 8080"
+	@echo "To stop the server: kill \`cat .miniwob-server.pid\`"
+
+test: setup miniwob
+	@MINIWOB_URL="http://localhost:8080/miniwob/" pytest -n 5 --durations=10 -m 'not pricy' -v tests/
+
+lint:
+	black src/ --check --diff
+	darglint -v 2 -z short src/

src/agentlab/experiments/loop.py

@@ -924,10 +924,13 @@ def _flatten_dict(d, parent_key="", sep="."):
     return dict(items)
 
 
-def as_tape(steps_info: list) -> Tape:
+def as_tape(steps_info: list[StepInfo]) -> Tape:
     """
     Create a Tape object from the steps info.
 
+    Args:
+        steps_info: list of StepInfo objects.
+
     Returns:
         Tape: a Tape object containing the steps and metadata.
     """