feat: Generate test cases to evaluate each issue #81

	name: Run test harness validation

	on:
	pull_request:
	types: [opened, synchronize, labeled]

	jobs:
	validate:
	if: ${{ contains(github.event.pull_request.labels.*.name, 'validate') }}
	runs-on: SWE-Bench_Larger
	defaults:
	run:
	shell: bash -leo pipefail {0}
	steps:
	- name: Checkout
	uses: actions/checkout@v4

	- name: Install dependencies
	run: \|
	sudo apt-get update
	sudo apt-get install -y \
	texlive \
	texlive-xetex \
	dvipng \
	ghostscript \
	libfreetype-dev \
	libtiff-dev \
	libxrender1

	# Cache the conda environment
	- name: Cache conda environment
	id: cache-conda
	uses: actions/cache@v4
	with:
	path: /usr/share/miniconda/envs/swe-bench
	key: conda-${{ runner.os }}-${{ hashFiles('environment.yml') }}

	# Create conda env if cache miss happens
	- name: Create conda env
	if: steps.cache-conda.outputs.cache-hit != 'true'
	run: \|
	conda init bash
	conda env create -f environment.yml

	- name: Run validation
	run: \|
	source /usr/share/miniconda/etc/profile.d/conda.sh
	conda activate swe-bench
	export PYTHONPATH=$PYTHONPATH:$(pwd)
	mkdir -p validate_logs
	python swebench/harness/engine_validation.py \
	--instances_path "princeton-nlp/SWE-bench_Lite" \
	--log_dir "validate_logs" \
	--temp_dir "${{ runner.temp }}" \
	--num_workers 1 \
	--verbose \
	--path_conda $(conda info --base)

	- name: Upload validation logs
	uses: actions/upload-artifact@v3
	if: ${{ always() }}
	with:
	name: validation-logs-${{ github.run_id }}
	path: validate_logs

Provide feedback