openadapt-ml 0.1.0__tar.gz

openadapt_ml-0.1.0/.env.example

@@ -0,0 +1,132 @@
+# API Keys for VLM backends
+ANTHROPIC_API_KEY=your-anthropic-api-key-here
+OPENAI_API_KEY=your-openai-api-key-here
+
+# =============================================================================
+# Lambda Labs (Cloud GPU for fast training)
+# =============================================================================
+# Get your API key at: https://cloud.lambdalabs.com/api-keys
+#
+# Usage:
+#   python -m openadapt_ml.cloud.lambda_labs list      # See available GPUs
+#   python -m openadapt_ml.cloud.lambda_labs launch    # Launch A100 instance
+#   python -m openadapt_ml.cloud.lambda_labs terminate <id>  # Stop billing!
+#
+LAMBDA_API_KEY=your-lambda-api-key-here
+
+# Google Gemini API (for GeminiGrounder element detection)
+# Get your key via one of these methods:
+#
+# Option 1: Google AI Studio (easiest)
+#   1. Go to https://aistudio.google.com/apikey
+#   2. Click "Create API Key"
+#   3. Select or create a Google Cloud project
+#   4. Copy the key
+#
+# Option 2: Google Cloud Console
+#   1. Go to https://console.cloud.google.com/apis/credentials
+#   2. Select your project
+#   3. Click "Create Credentials" → "API Key"
+#   4. Enable "Generative Language API" at:
+#      https://console.cloud.google.com/apis/library/generativelanguage.googleapis.com
+#
+GOOGLE_API_KEY=your-google-api-key-here
+
+# =============================================================================
+# Azure Credentials (for WAA benchmark evaluation on Azure VMs)
+# =============================================================================
+#
+# EASY SETUP: Run the automated setup script:
+#   python scripts/setup_azure.py
+#
+# This script will:
+#   - Check Azure CLI installation
+#   - Log you in (opens browser)
+#   - Create service principal, resource group, and ML workspace
+#   - Write credentials to this file
+#
+# MANUAL SETUP: Follow the steps below if you prefer manual control.
+#
+# Azure authentication uses Service Principal credentials. To create them:
+#
+# Step 1: Install Azure CLI
+#   macOS:   brew install azure-cli
+#   Windows: winget install Microsoft.AzureCLI
+#   Linux:   curl -sL https://aka.ms/InstallAzureCLIDeb | sudo bash
+#
+# Step 2: Login to Azure
+#   az login
+#
+# Step 3: Get your subscription ID
+#   az account show --query id -o tsv
+#
+# Step 4: Create a Service Principal (replace <subscription-id>)
+#   az ad sp create-for-rbac \
+#     --name "openadapt-ml-waa" \
+#     --role "Contributor" \
+#     --scopes "/subscriptions/<subscription-id>" \
+#     --sdk-auth
+#
+#   This outputs JSON with clientId, clientSecret, tenantId - copy these below.
+#
+# Step 5: Create Azure ML Workspace (if you don't have one)
+#   az ml workspace create \
+#     --name agents_ml \
+#     --resource-group agents \
+#     --location eastus
+#
+# Step 6: Request vCPU quota increase (if needed)
+#   Go to: https://portal.azure.com/#view/Microsoft_Azure_Capacity/QuotaMenuBlade
+#   Select your subscription → Compute → Standard Dv3 Family
+#   Request increase to at least 320 vCPUs (for 40 workers × 8 vCPUs each)
+#
+# Alternative: Use Azure CLI login (no service principal needed)
+#   If AZURE_CLIENT_ID is not set, DefaultAzureCredential will try:
+#   1. Environment variables (service principal)
+#   2. Azure CLI credentials (az login)
+#   3. Managed Identity (when running on Azure)
+#
+# Service Principal credentials (from Step 4 output)
+AZURE_CLIENT_ID=your-client-id-here
+AZURE_CLIENT_SECRET=your-client-secret-here
+AZURE_TENANT_ID=your-tenant-id-here
+
+# Azure ML Workspace config
+AZURE_SUBSCRIPTION_ID=your-subscription-id-here
+AZURE_ML_RESOURCE_GROUP=agents
+AZURE_ML_WORKSPACE_NAME=agents_ml
+
+# Optional: Override default VM settings
+# AZURE_VM_SIZE=Standard_D8_v3
+# AZURE_DOCKER_IMAGE=ghcr.io/microsoft/windowsagentarena:latest
+
+# =============================================================================
+# Azure Storage for Async Inference Queue (Phase 2)
+# =============================================================================
+#
+# These settings are auto-configured by setup_azure.py (steps 11-12).
+# They enable live inference during training:
+# - Training uploads checkpoints to blob storage
+# - Inference worker polls queue for new checkpoints
+# - Results uploaded to blob for dashboard to display
+#
+# Setup: Run 'python scripts/setup_azure.py' to create storage account and queue
+#
+# Manual setup (advanced):
+#   1. Create storage account:
+#      az storage account create --name openadaptmlstorage \
+#        --resource-group agents --location eastus --sku Standard_LRS
+#
+#   2. Get connection string:
+#      az storage account show-connection-string \
+#        --name openadaptmlstorage -o tsv
+#
+#   3. Create queue and containers:
+#      az storage queue create --name inference-jobs --connection-string <conn-str>
+#      az storage container create --name checkpoints --connection-string <conn-str>
+#      az storage container create --name comparisons --connection-string <conn-str>
+#
+AZURE_STORAGE_CONNECTION_STRING=your-storage-connection-string-here
+AZURE_INFERENCE_QUEUE_NAME=inference-jobs
+AZURE_CHECKPOINTS_CONTAINER=checkpoints
+AZURE_COMPARISONS_CONTAINER=comparisons

openadapt_ml-0.1.0/.github/workflows/publish.yml

@@ -0,0 +1,30 @@
+name: Publish to PyPI
+
+on:
+  push:
+    tags:
+      - 'v*'
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    permissions:
+      id-token: write  # Required for trusted publishing
+      contents: read   # Required for checkout
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v4
+        with:
+          version: "latest"
+
+      - name: Set up Python
+        run: uv python install 3.12
+
+      - name: Build package
+        run: uv build
+
+      - name: Publish to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1

openadapt_ml-0.1.0/.gitignore

@@ -0,0 +1,41 @@
+# Python-generated files
+__pycache__/
+*.py[oc]
+build/
+dist/
+wheels/
+*.egg-info
+
+# Virtual environments
+.venv
+local_context_openadapt_ml_internal.md
+
+# Environment variables
+.env
+
+# Ephemeral synthetic assets (frames, debug sessions, etc.)
+synthetic/
+synthetic_train/
+synthetic_train_dev/
+
+# Local OS cruft
+.DS_Store
+
+# Editor files
+*.swp
+*.swo
+*.swn
+*~
+.*.sw?
+
+# Ephemeral training/eval artifacts
+logs/
+checkpoints/
+checkpoints_*/
+/plots/
+eval_*.json
+benchmark_results/
+debug_*/
+
+# Internal documentation (not for public repo)
+docs/internal/

openadapt_ml-0.1.0/.gitmodules

@@ -0,0 +1,3 @@
+[submodule "vendor/WindowsAgentArena"]
+	path = vendor/WindowsAgentArena
+	url = https://github.com/microsoft/WindowsAgentArena.git

openadapt_ml-0.1.0/.python-version

@@ -0,0 +1 @@
+3.12