Skip to content

WIP - dmg/exe installable Letta #1394

WIP - dmg/exe installable Letta

WIP - dmg/exe installable Letta #1394

Workflow file for this run

name: Anthropic Claude Opus 3 Capabilities Test
env:
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
on:
push:
branches: [ main ]
pull_request:
branches: [ main ]
jobs:
test:
runs-on: ubuntu-latest
timeout-minutes: 15
steps:
- name: Checkout
uses: actions/checkout@v4
- name: "Setup Python, Poetry and Dependencies"
uses: packetcoders/action-setup-cache-python-poetry@main
with:
python-version: "3.12"
poetry-version: "1.8.2"
install-args: "-E dev -E external-tools"
- name: Test first message contains expected function call and inner monologue
id: test_first_message
env:
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
run: |
poetry run pytest -s -vv tests/test_model_letta_perfomance.py::test_claude_opus_3_returns_valid_first_message
echo "TEST_FIRST_MESSAGE_EXIT_CODE=$?" >> $GITHUB_ENV
continue-on-error: true
- name: Test model sends message with keyword
id: test_keyword_message
env:
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
run: |
poetry run pytest -s -vv tests/test_model_letta_perfomance.py::test_claude_opus_3_returns_keyword
echo "TEST_KEYWORD_MESSAGE_EXIT_CODE=$?" >> $GITHUB_ENV
continue-on-error: true
- name: Test model uses external tool correctly
id: test_external_tool
env:
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
run: |
poetry run pytest -s -vv tests/test_model_letta_perfomance.py::test_claude_opus_3_uses_external_tool
echo "TEST_EXTERNAL_TOOL_EXIT_CODE=$?" >> $GITHUB_ENV
continue-on-error: true
- name: Test model recalls chat memory
id: test_chat_memory
env:
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
run: |
poetry run pytest -s -vv tests/test_model_letta_perfomance.py::test_claude_opus_3_recall_chat_memory
echo "TEST_CHAT_MEMORY_EXIT_CODE=$?" >> $GITHUB_ENV
continue-on-error: true
- name: Test model uses 'archival_memory_search' to find secret
id: test_archival_memory
env:
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
run: |
poetry run pytest -s -vv tests/test_model_letta_perfomance.py::test_claude_opus_3_archival_memory_retrieval
echo "TEST_ARCHIVAL_MEMORY_EXIT_CODE=$?" >> $GITHUB_ENV
continue-on-error: true
- name: Test model can edit core memories
id: test_core_memory
env:
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
run: |
poetry run pytest -s -vv tests/test_model_letta_perfomance.py::test_claude_opus_3_edit_core_memory
echo "TEST_CORE_MEMORY_EXIT_CODE=$?" >> $GITHUB_ENV
continue-on-error: true
- name: Summarize test results
if: always()
run: |
echo "Test Results Summary:"
echo "Test first message: $([[ $TEST_FIRST_MESSAGE_EXIT_CODE -eq 0 ]] && echo ✅ || echo ❌)"
echo "Test model sends message with keyword: $([[ $TEST_KEYWORD_MESSAGE_EXIT_CODE -eq 0 ]] && echo ✅ || echo ❌)"
echo "Test model uses external tool: $([[ $TEST_EXTERNAL_TOOL_EXIT_CODE -eq 0 ]] && echo ✅ || echo ❌)"
echo "Test model recalls chat memory: $([[ $TEST_CHAT_MEMORY_EXIT_CODE -eq 0 ]] && echo ✅ || echo ❌)"
echo "Test model uses 'archival_memory_search' to find secret: $([[ $TEST_ARCHIVAL_MEMORY_EXIT_CODE -eq 0 ]] && echo ✅ || echo ❌)"
echo "Test model can edit core memories: $([[ $TEST_CORE_MEMORY_EXIT_CODE -eq 0 ]] && echo ✅ || echo ❌)"
# Check if any test failed
if [[ $TEST_FIRST_MESSAGE_EXIT_CODE -ne 0 || \
$TEST_KEYWORD_MESSAGE_EXIT_CODE -ne 0 || \
$TEST_EXTERNAL_TOOL_EXIT_CODE -ne 0 || \
$TEST_CHAT_MEMORY_EXIT_CODE -ne 0 || \
$TEST_ARCHIVAL_MEMORY_EXIT_CODE -ne 0 || \
$TEST_CORE_MEMORY_EXIT_CODE -ne 0 ]]; then
echo "Some tests failed."
exit 78
fi