ramllm 0.1.0__tar.gz

ramllm-0.1.0/.github/workflows/ci.yml

@@ -0,0 +1,49 @@
+name: CI
+
+on:
+  push:
+    branches: [ "main" ]
+  pull_request:
+    branches: [ "main" ]
+
+jobs:
+  test:
+    name: Test (${{ matrix.os }}, py${{ matrix.python-version }})
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest, macos-latest, windows-latest]
+        python-version: ["3.11", "3.12", "3.13"]
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Set up uv
+      uses: astral-sh/setup-uv@v5
+      with:
+        enable-cache: true
+        python-version: ${{ matrix.python-version }}
+
+    - name: Install dependencies
+      run: uv sync --all-extras
+
+    - name: Run Ruff Lint Check
+      run: uv run ruff check .
+
+    - name: Run Ruff Format Check
+      run: uv run ruff format --check .
+
+    - name: Type Check Core & CLI
+      run: uv run mypy src tests
+
+    - name: Validate Curated Models Index Schema
+      run: uv run python -c "from ramllm.services.model_service import load_curated_index; idx = load_curated_index(); assert len(idx.models) >= 15, f'Expected >=15 models, got {len(idx.models)}'"
+
+    - name: Verify CLI Documentation Freshness
+      run: |
+        uv run python scripts/generate_cli_docs.py
+        git diff --exit-code docs/cli.md
+
+    - name: Run Tests (excluding slow/gpu)
+      run: uv run pytest -m "not slow and not gpu" -v

ramllm-0.1.0/.github/workflows/release.yml

@@ -0,0 +1,148 @@
+name: Release
+
+on:
+  push:
+    tags:
+      - "v*"
+
+permissions:
+  contents: write
+  id-token: write
+
+jobs:
+  build-package:
+    name: Build sdist and wheel
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up uv
+        uses: astral-sh/setup-uv@v5
+        with:
+          enable-cache: true
+          python-version: "3.12"
+
+      - name: Build distributions
+        run: uv build
+
+      - name: Upload dist artifacts
+        uses: actions/upload-artifact@v4
+        with:
+          name: python-dist
+          path: dist/*
+
+  build-binaries:
+    name: Build Standalone Binary (${{ matrix.os }})
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - os: ubuntu-latest
+            artifact_name: ramllm-linux-x86_64
+            binary_path: dist/ramllm
+          - os: macos-latest
+            artifact_name: ramllm-darwin-arm64
+            binary_path: dist/ramllm
+          - os: macos-13
+            artifact_name: ramllm-darwin-x86_64
+            binary_path: dist/ramllm
+          - os: windows-latest
+            artifact_name: ramllm-windows-x86_64.exe
+            binary_path: dist/ramllm.exe
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up uv
+        uses: astral-sh/setup-uv@v5
+        with:
+          enable-cache: true
+          python-version: "3.12"
+
+      - name: Install dependencies and PyInstaller
+        run: |
+          uv sync --all-extras
+          uv pip install pyinstaller
+
+      - name: Build standalone executable
+        shell: bash
+        run: |
+          uv run pyinstaller --onefile --name ramllm \
+            --add-data "src/ramllm/data/models_index.json:ramllm/data" \
+            --collect-all textual \
+            --collect-all rich \
+            src/ramllm/__main__.py
+
+      - name: Smoke test binary
+        shell: bash
+        run: |
+          ${{ matrix.binary_path }} --version
+          ${{ matrix.binary_path }} doctor --plain || true
+
+      - name: Rename and compute checksum
+        shell: bash
+        run: |
+          cp "${{ matrix.binary_path }}" "${{ matrix.artifact_name }}"
+          if command -v sha256sum >/dev/null 2>&1; then
+            sha256sum "${{ matrix.artifact_name }}" > "${{ matrix.artifact_name }}.sha256"
+          else
+            shasum -a 256 "${{ matrix.artifact_name }}" > "${{ matrix.artifact_name }}.sha256"
+          fi
+
+      - name: Upload standalone binary
+        uses: actions/upload-artifact@v4
+        with:
+          name: ${{ matrix.artifact_name }}
+          path: |
+            ${{ matrix.artifact_name }}
+            ${{ matrix.artifact_name }}.sha256
+
+  publish-release:
+    name: Create GitHub Release
+    needs: [build-package, build-binaries]
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Download all artifacts
+        uses: actions/download-artifact@v4
+        with:
+          path: release-artifacts
+
+      - name: Prepare release directory and consolidated checksums
+        shell: bash
+        run: |
+          mkdir -p dist_release
+          find release-artifacts -type f -exec cp {} dist_release/ \;
+          cd dist_release
+          sha256sum * > SHA256SUMS.txt || true
+          ls -la
+
+      - name: Create GitHub Release
+        uses: softprops/action-gh-release@v2
+        with:
+          files: dist_release/*
+          generate_release_notes: true
+          draft: false
+          prerelease: false
+
+  publish-pypi:
+    name: Publish to PyPI
+    needs: [build-package]
+    runs-on: ubuntu-latest
+    permissions:
+      id-token: write
+    steps:
+      - name: Download dist artifacts
+        uses: actions/download-artifact@v4
+        with:
+          name: python-dist
+          path: dist/
+
+      - name: Publish package to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1
+        continue-on-error: true
+        with:
+          skip-existing: true
+

ramllm-0.1.0/.github/workflows/verify_index_schedule.yml

@@ -0,0 +1,27 @@
+name: Curated Model Index Verification
+
+on:
+  schedule:
+    # Run every Sunday at midnight UTC
+    - cron: '0 0 * * 0'
+  workflow_dispatch:
+
+jobs:
+  verify:
+    name: Verify Hugging Face Links & Checksums
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Set up uv
+      uses: astral-sh/setup-uv@v5
+      with:
+        enable-cache: true
+        python-version: "3.12"
+
+    - name: Install dependencies
+      run: uv sync --all-extras
+
+    - name: Run Model Index Verification
+      run: uv run python scripts/verify_index.py

ramllm-0.1.0/.gitignore

@@ -0,0 +1,16 @@
+.venv/
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+dist/
+build/
+*.egg-info/
+.coverage
+htmlcov/
+.DS_Store
+~/.cache/ramllm/
+hardware.json

ramllm-0.1.0/.hypothesis/.gitignore

@@ -0,0 +1,9 @@
+# This .gitignore file was automatically created by Hypothesis. Hypothesis gitignores
+# .hypothesis by default, because we generally recommend that .hypothesis not be checked
+# into version control.
+#
+# If you *would* like to check .hypothesis into version control, you should delete this
+# file. Hypothesis will not re-create this .gitignore unless .hypothesis is deleted (and
+# if it does, that's a bug - please report it!)
+
+*

ramllm-0.1.0/.hypothesis/constants/004c982f914da8ee

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/profiler/disk.py
+# hypothesis_version: 6.169.1
+
+[b'\xaa', 1e-06, 2.5, 128, 1024, '1', 'Apple Fabric', 'Darwin', 'Linux', 'NVMe', 'Rotational', 'Solid State', 'diskutil', 'hdd', 'info', 'nvme', 'rb', 'ssd', 'unknown', 'wb']

ramllm-0.1.0/.hypothesis/constants/00a668a974c7cad2

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/services/config_service.py
+# hypothesis_version: 6.169.1
+
+[1024, 8080, 65535, ', ', '0', '1', '127.0.0.1', 'HF_TOKEN', 'always', 'ascii', 'auto', 'balanced', 'conservative', 'false', 'fast', 'max', 'never', 'no', 'none', 'null', 'quality', 'rb', 'true', 'wb', 'yes']

ramllm-0.1.0/.hypothesis/constants/00bd7f5a0e99674d

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/main.py
+# hypothesis_version: 6.169.1
+
+[40.0, 60.0, 128, 200, 1024, 8080, 8085, ' (Unified Memory)', ',', ', ', '-', '--context', '--cpu-only', '--explain', '--gpus', '--host', '--hw-json', '--json', '--live', '--no-kv-quant', '--plan-id', '--port', '--prefer', '--prompt', '--ram', '--refresh', '--vram', '-p', '0.0 GB', '127.0.0.1', '4096', 'Arch', 'BATCH_ONLY', 'Benchmark Results', 'Bottleneck', 'COMFORTABLE', 'CONSTRAINED', 'CPU', 'CPU Features', 'Component', 'Darwin', 'Decode Speed', 'Flash Attention', 'GPU', 'GPU Offload', 'Hardware Profile', 'INFEASIBLE', 'KV Quantization', 'Linux', 'Measured', 'Metric', 'MoE', 'MoE CPU Offload', 'Model Catalog', 'Name', 'No', 'Operating System', 'Output JSON', 'Output raw JSON', 'Path', 'Path to GGUF model', 'Peak RAM RSS', 'Port to listen', 'Predicted', 'Prefill Speed', 'Quant', 'RAM Bandwidth', 'RAM Peak', 'Rank', 'Size', 'Specification', 'Storage Drive', 'Strategy', 'System RAM', 'VRAM Peak', 'Verdict', 'Windows', 'Yes', '__main__', 'add', 'balanced', 'blue', 'bold', 'bold cyan', 'choices', 'content', 'cyan', 'dim', 'fastest', 'fit_at_any_cost', 'green', 'info', 'json', 'list', 'magenta', 'max_tokens', 'message', 'messages', 'models', 'ramllm', 'role', 'scan', 'standard', 'stream', 'user', 'white', 'yellow']

ramllm-0.1.0/.hypothesis/constants/022da3471cc9eef6

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/services/model_resolver.py
+# hypothesis_version: 6.169.1
+
+[0.5, 1.1, 180.0, 100, 1024, ', ', '.gguf', '/', '401', '403', ':', '@', 'Authorization', 'Downloading model...', 'HF_TOKEN']

ramllm-0.1.0/.hypothesis/constants/03b9f5b6a1362865

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/planner/explain.py
+# hypothesis_version: 6.169.1
+
+[1024, 'GPU VRAM', 'Host RAM', 'compute', 'cpu', 'disk_bw', 'expert_cache_miss', 'gpu', 'pcie_bw', 'ram_bw', 'vram_bw']

ramllm-0.1.0/.hypothesis/constants/04eaca8883ecb83a

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/telemetry/calibration.py
+# hypothesis_version: 6.169.1
+
+[0.5, 1.0, 2.0, '.local', 'a', 'bench.jsonl', 'ramllm', 'share', 'utf-8']

ramllm-0.1.0/.hypothesis/constants/0534f27584ed9984

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/planner/candidates.py
+# hypothesis_version: 6.169.1
+
+[0.25, 1.5, 500, 512, 1024, '--no-kv-offload', '-c', '-cmoe', '-ctk', '-ctv', '-lm', '-m', '-ngl', 'cpu', 'f16', 'gpu', 'llamacpp', 'none', 'q4_0', 'q8_0']

ramllm-0.1.0/.hypothesis/constants/0564273ed6240e13

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/services/model_resolver.py
+# hypothesis_version: 6.169.1
+
+[0.5, 1.1, 180.0, 100, 1024, ', ', '../', '..\\', './', '.\\', '.gguf', '/', '401', '403', ':', '@', 'Authorization', 'Downloading model...', 'HF_TOKEN', '\\', '~']

ramllm-0.1.0/.hypothesis/constants/05694ca355b24270

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/backends/llamacpp/adapter.py
+# hypothesis_version: 6.169.1
+
+[30.0, 8080, 'llamacpp']

ramllm-0.1.0/.hypothesis/constants/057e3bc7376007b0

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/backends/__init__.py
+# hypothesis_version: 6.169.1
+
+['BackendAdapter', 'BackendInstall', 'BackendProcess', 'Capabilities', 'LlamaCppAdapter', 'MemoryReport', 'MockBackendAdapter']

ramllm-0.1.0/.hypothesis/constants/07b44469d18d4b31

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/commands/tour_cmd.py
+# hypothesis_version: 6.169.1
+
+['1. Memory Tiers', '2. Plan Verdicts', 'Disk / Swap', 'Fast & spacious', 'How RAMLLM Uses It', 'Lightning fast', 'Slow safety net', 'Speed', 'System RAM', 'Tier', 'User Experience', 'VRAM', 'Verdict', 'bold magenta', 'cyan', 'green', 'white']

ramllm-0.1.0/.hypothesis/constants/083b5eb14051c7e4

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/commands/chat_cmd.py
+# hypothesis_version: 6.169.1
+
+[0.7, 0.9, 35.0, 100, 4096, 8085, '--attach-server', '--ctx', '--memory', '--mode', '--no-save', '--plain', '--port', '--prompt', '--resume', '--seed', '--system', '--temp', '--top-p', '-p', '/', '127.0.0.1', 'BATCH_ONLY', 'CONSTRAINED', 'INFEASIBLE', 'Launch anyway?', 'Sampling temperature', 'attached-server', 'qwen-2.5-7b', 'system', 'user']

ramllm-0.1.0/.hypothesis/constants/11146de3180031bf

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/setup/state.py
+# hypothesis_version: 6.169.1
+
+['balanced', 'conservative', 'fast', 'max', 'quality', 'setup_state.json', 'utf-8']

ramllm-0.1.0/.hypothesis/constants/124d9aaa6b564104

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/chat/repl.py
+# hypothesis_version: 6.169.1
+
+[0.7, 0.85, 0.9, 0.95, 30.0, 200, 1024, 4096, 8085, '\n[dim]Goodbye![/dim]', '"""', "'''", '... ', '/', 'INFEASIBLE', 'ansicyan bold', 'assistant', 'balanced', 'bg:#222222 #cccccc', 'bottom-toolbar', 'context', 'prompt', 'system', 'user']

ramllm-0.1.0/.hypothesis/constants/12530bd8966d1b79

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/services/model_service.py
+# hypothesis_version: 6.169.1
+
+[0.55, 0.7, 0.75, 0.8, 0.9, 0.92, 1.05, 1000000000.0, 512, 1024, 4096, '.gguf', '.part', '.tmp', '1.0.0', '5', '8', 'COMFORTABLE', 'CONSTRAINED', 'INFEASIBLE', 'Q3_K_M', 'Q4_0', 'Q4_K_M', 'Q4_K_S', 'Q5_K_M', 'Q8_0', 'data', 'dense', 'free_disk_bytes', 'green', 'model_files', 'models_dir', 'models_index.json', 'moe', 'name', 'partial_files', 'path', 'red', 'runs well', 'runs well on GPU', 'size_bytes', 'total_disk_bytes', 'total_model_bytes', 'total_partial_bytes', 'total_used_bytes', 'utf-8', 'works, but slowly', 'yellow']

ramllm-0.1.0/.hypothesis/constants/15d8e0650fc30dcf

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/telemetry/bench.py
+# hypothesis_version: 6.169.1
+
+[0.001, 0.1, 1.0, 40.0, 60.0, 2048, 8085, 'INFEASIBLE', 'POST', 'content', 'data: ', 'data: [DONE]', 'max_tokens', 'messages', 'role', 'stream', 'user']

ramllm-0.1.0/.hypothesis/constants/166191bf61872594

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/commands/backend_cmd.py
+# hypothesis_version: 6.169.1
+
+[1024, ', ', '-', '--tag', '--yes', '-y', 'Binary Path', 'CPU', 'Installed', 'SYSTEM', 'Status', 'Tag / Source', 'Version', 'binary_path', 'cyan', 'green', 'install', 'is_active', 'list', 'magenta', 'remove', 'tag', 'unknown', 'update', 'use', 'version', 'white']

ramllm-0.1.0/.hypothesis/constants/193eb30f6e0fb5ce

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/chat/__init__.py
+# hypothesis_version: 6.169.1
+
+['ChatClient', 'ChatMessage', 'ChatMetrics', 'ChatSession', 'start_chat_session']

ramllm-0.1.0/.hypothesis/constants/19898c371d1c7d28

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/chat/client.py
+# hypothesis_version: 6.169.1
+
+[0.001, 0.7, 0.9, 2.0, 120.0, 200, 204, '/', '/v1', 'POST', '[DONE]', 'choices', 'completion_tokens', 'content', 'data:', 'delta', 'max_tokens', 'messages', 'prompt_tokens', 'replace', 'role', 'seed', 'stream', 'temperature', 'top_p', 'usage', 'utf-8']

ramllm-0.1.0/.hypothesis/constants/1c4b8bf4b6f1c395

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/app.py
+# hypothesis_version: 6.169.1
+
+[40.0, 60.0, 128, 200, 1024, 8080, 8085, ' (Unified Memory)', ',', ', ', '-', '--context', '--cpu-only', '--debug', '--details', '--dry-run', '--explain', '--gpus', '--host', '--hw-json', '--json', '--live', '--max-retries', '--memory', '--mode', '--no-kv-quant', '--plain', '--plan-id', '--port', '--prefer', '--prompt', '--ram', '--refresh', '--vram', '-p', '0.0 GB', '127.0.0.1', '4096', 'BATCH_ONLY', 'Benchmark Results', 'Bottleneck', 'CHAT & RUN', 'COMFORTABLE', 'CONSTRAINED', 'CPU', 'CPU Features', 'Component', 'Decode Speed', 'GET STARTED', 'GPU', 'Hardware Profile', 'INFEASIBLE', 'Key Advantage', 'MODELS', 'Measured', 'Metric', 'Model', 'Operating System', 'Output JSON', 'Output raw JSON', 'Params', 'Path to GGUF model', 'Peak RAM RSS', 'Plan Summary', 'Port to listen', 'Predicted', 'Prefill Speed', 'Quant', 'RAM Bandwidth', 'RAM Peak', 'Rank', 'SYSTEM', 'Specification', 'Speed', 'Storage Drive', 'Strategy', 'System RAM', 'VRAM Peak', 'Verdict', 'backend', 'balanced', 'blue', 'bold', 'bold cyan', 'bold green', 'choices', 'config', 'content', 'cyan', 'dim', 'doctor', 'explain', 'fast', 'fastest', 'fit_at_any_cost', 'green', 'info', 'json', 'magenta', 'max', 'max_tokens', 'message', 'messages', 'models', 'quality', 'ramllm', 'rich', 'role', 'standard', 'stream', 'user', 'white', 'yellow']

ramllm-0.1.0/.hypothesis/constants/1c65a5618b7fbdc1

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/governor/monitor.py
+# hypothesis_version: 6.169.1
+
+[0.0001, 0.15, 0.25, 1.0, 2.0, 15.0, 120, 500, 1024, 'sin']

ramllm-0.1.0/.hypothesis/constants/1fb716436f9706ed

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/governor/enforcement/__init__.py
+# hypothesis_version: 6.169.1
+
+['EnforcementManager', 'EnforcementStatus']

ramllm-0.1.0/.hypothesis/constants/2022e9b7980e5085

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/backends/llamacpp/adapter.py
+# hypothesis_version: 6.169.1
+
+[30.0, 8080, 'llamacpp']

ramllm-0.1.0/.hypothesis/constants/209c98dca1a33d09

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/app.py
+# hypothesis_version: 6.169.1
+
+[40.0, 60.0, 128, 200, 1024, 8080, 8085, ' (Unified Memory)', ',', ', ', '-', '--context', '--cpu-only', '--debug', '--dest', '--details', '--explain', '--gpus', '--host', '--hw-json', '--json', '--live', '--max-retries', '--memory', '--mode', '--no-kv-quant', '--plain', '--plan-id', '--port', '--prefer', '--prompt', '--ram', '--refresh', '--sha256', '--vram', '-p', '0.0 GB', '127.0.0.1', '4096', 'Arch', 'BATCH_ONLY', 'Benchmark Results', 'Bottleneck', 'CHAT & RUN', 'COMFORTABLE', 'CONSTRAINED', 'CPU', 'CPU Features', 'Component', 'Decode Speed', 'GET STARTED', 'GPU', 'Hardware Profile', 'INFEASIBLE', 'Key Advantage', 'MODELS', 'Measured', 'Metric', 'MoE', 'Model', 'Model Catalog', 'Name', 'No', 'Operating System', 'Output JSON', 'Output raw JSON', 'Params', 'Path', 'Path to GGUF model', 'Peak RAM RSS', 'Plan Summary', 'Port to listen', 'Predicted', 'Prefill Speed', 'Quant', 'RAM Bandwidth', 'RAM Peak', 'Rank', 'SYSTEM', 'Size', 'Specification', 'Speed', 'Storage Drive', 'Strategy', 'System RAM', 'VRAM Peak', 'Verdict', 'Yes', 'add', 'backend', 'balanced', 'blue', 'bold', 'bold cyan', 'bold green', 'choices', 'config', 'content', 'cyan', 'dim', 'doctor', 'explain', 'fast', 'fastest', 'fit_at_any_cost', 'green', 'info', 'json', 'list', 'magenta', 'max', 'max_tokens', 'message', 'messages', 'models', 'pull', 'quality', 'ramllm', 'remove', 'rich', 'role', 'scan', 'standard', 'stream', 'user', 'verify', 'white', 'yellow']

ramllm-0.1.0/.hypothesis/constants/21dba30c08d5cdc5

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/planner/candidates.py
+# hypothesis_version: 6.169.1
+
+[0.25, 1.5, 500, 512, 1024, '--no-kv-offload', '-c', '-cmoe', '-ctk', '-ctv', '-lm', '-m', '-ngl', 'cpu', 'f16', 'gpu', 'llamacpp', 'none', 'q4_0', 'q8_0']

ramllm-0.1.0/.hypothesis/constants/25efdf23674d5c16

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/main.py
+# hypothesis_version: 6.169.1
+
+['__main__']

ramllm-0.1.0/.hypothesis/constants/27194a0932b262d5

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/services/plan_service.py
+# hypothesis_version: 6.169.1
+
+[0.5, 0.7, 0.75, 0.85, 1.0, 1.5, 2.0, 128, 256, 512, 1024, 4096, ',', 'BATCH_ONLY', 'COMFORTABLE', 'CONSTRAINED', 'INFEASIBLE', 'balanced', 'cmoe', 'conservative', 'cpu', 'fast', 'fastest', 'fit_at_any_cost', 'fraction', 'max', 'moderate', 'quality', 'ram_reserve_gb', 'very slow', 'vram_reserve_mb']

ramllm-0.1.0/.hypothesis/constants/286e2ddca3cdbaef

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/governor/recovery.py
+# hypothesis_version: 6.169.1
+
+[30.0, 8080, 'INFEASIBLE', 'cgroup memory limit', 'code -9', 'code 137', 'cuda out of memory', 'cudaerrornodevice', 'device lost', 'device unavailable', 'driver reset', 'driver shutting down', 'driver_loss', 'general_crash', 'gpu_oom', 'host_oom', 'metal device failure', 'out of memory', 'std::bad_alloc', 'timed out', 'timeout']

ramllm-0.1.0/.hypothesis/constants/28ad55a1b2e80514

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/commands/models_cmd.py
+# hypothesis_version: 6.169.1
+
+[0.85, 1.0, 1024, ', ', '--dest', '--fits', '--json', '--quant', '--sha256', '--tag', '--yes', '-q', '-t', '-y', '@', 'Arch', 'COMFORTABLE', 'CONSTRAINED', 'File', 'Fits?', 'Hugging Face File', 'ID', 'INFEASIBLE', 'Kind', 'MODELS', 'MoE', 'Model (GGUF)', 'Model Catalog', 'Model Disk Usage', 'Model Name', 'N/A', 'Name', 'No', 'Notes', 'Params', 'Path', 'Quant', 'Size', 'Tags', 'Yes', '[red]✘ heavy[/red]', 'add', 'bold', 'bold cyan', 'bold green', 'clean', 'cyan', 'dim', 'du', 'green', 'info', 'json', 'list', 'magenta', 'model_files', 'name', 'partial_files', 'pull', 'remove', 'scan', 'search', 'size_bytes', 'verify', 'white', 'yellow', '~', '✔', '✘']

ramllm-0.1.0/.hypothesis/constants/28cf0cf24899d9ab

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/commands/serve_cmd.py
+# hypothesis_version: 6.169.1
+
+[0.4, 0.8, 35.0, 100, 200, 512, 8080, '"not-needed"', ',', '--api-key', '--context', '--detach', '--engine-id', '--gpus', '--host', '--max-retries', '--port', '--ram', '--vram', '-d', '-m', '0.0.0.0', '127.0.0.1', '4096', '::', 'active', 'green', 'info', 'ramllm.cli.main', 'replace', 'serve', 'utf-8', 'w', 'win32']

ramllm-0.1.0/.hypothesis/constants/2aa29dbd39b0f103

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/commands/dashboard_cmd.py
+# hypothesis_version: 6.169.1
+
+[2.0, '--watch', '-w', 'custom-url', 'http://', 'https://', 'local-engine']

ramllm-0.1.0/.hypothesis/constants/2b074c0d9c3ae907

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/planner/recommend.py
+# hypothesis_version: 6.169.1
+
+[1.0, 1.5, 4.0, 8.0, 10.0, 12.0, 15.0, 20.0, 40.0, 60.0, 100.0, 1000000000.0, 'BATCH_ONLY', 'COMFORTABLE', 'CONSTRAINED', 'F16', 'INFEASIBLE', 'Q3', 'Q4', 'Q5', 'Q6', 'Q8']

ramllm-0.1.0/.hypothesis/constants/2d18a4b7582ff54b

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/services/backend_manager.py
+# hypothesis_version: 6.169.1
+
+[15.0, 60.0, 100, 200, 1024, '-arm64.', '-bin-', '-d', '-x64.', '.tar.gz', '.zip', '11.8', '12.2', '12.4', '12.8', '13.4', 'GET', 'RAMLLM_BACKEND_PATH', 'User-Agent', 'aarch64', 'active', 'amd64', 'android', 'arm64', 'assets', 'binary_path', 'browser_download_url', 'capabilities', 'com.apple.quarantine', 'config', 'content-length', 'cpu', 'cuda', 'cudart-', 'darwin', 'download.tmp', 'env', 'installed_at', 'installs', 'is_active', 'linux', 'llama', 'llama-', 'llama-server', 'llama-server.exe', 'llama.cpp', 'mac', 'macos', 'managed', 'name', 'openvino', 'path', 'r', 'r:*', 'registry.json', 'rocm', 's390x', 'size', 'snapdragon', 'sycl', 'tag', 'tag_name', 'ubuntu', 'ubuntu-arm64.tar.gz', 'ubuntu-x64.tar.gz', 'ui', 'unknown', 'utf-8', 'version', 'vulkan', 'wb', 'win', 'win-cpu', 'win-cuda', 'windows', 'x64', 'x86_64', 'xattr', 'xcframework']

ramllm-0.1.0/.hypothesis/constants/2e8abcd17805cf9a

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/main.py
+# hypothesis_version: 6.169.1
+
+['__main__']

ramllm-0.1.0/.hypothesis/constants/301fde9e938d5711

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/main.py
+# hypothesis_version: 6.169.1
+
+[40.0, 60.0, 128, 200, 1024, 8080, 8085, ' (Unified Memory)', ',', ', ', '-', '--context', '--cpu-only', '--dest', '--explain', '--gpus', '--host', '--hw-json', '--json', '--live', '--max-retries', '--no-kv-quant', '--plan-id', '--port', '--prefer', '--prompt', '--ram', '--refresh', '--sha256', '--vram', '-p', '0.0 GB', '127.0.0.1', '4096', 'Arch', 'BATCH_ONLY', 'Benchmark Results', 'Bottleneck', 'COMFORTABLE', 'CONSTRAINED', 'CPU', 'CPU Features', 'Component', 'Decode Speed', 'Flash Attention', 'GPU', 'GPU Offload', 'Hardware Profile', 'INFEASIBLE', 'KV Quantization', 'Key Advantage', 'Measured', 'Metric', 'MoE', 'MoE CPU Offload', 'Model', 'Model Catalog', 'Name', 'No', 'Operating System', 'Output JSON', 'Output raw JSON', 'Params', 'Path', 'Path to GGUF model', 'Peak RAM RSS', 'Port to listen', 'Predicted', 'Prefill Speed', 'Quant', 'RAM Bandwidth', 'RAM Peak', 'Rank', 'Size', 'Specification', 'Speed', 'Storage Drive', 'Strategy', 'System RAM', 'VRAM Peak', 'Verdict', 'Yes', '__main__', 'add', 'balanced', 'blue', 'bold', 'bold cyan', 'bold green', 'choices', 'content', 'cyan', 'dim', 'fastest', 'fit_at_any_cost', 'green', 'info', 'json', 'list', 'magenta', 'max_tokens', 'message', 'messages', 'models', 'pull', 'ramllm', 'remove', 'role', 'scan', 'standard', 'stream', 'user', 'verify', 'white', 'yellow']

ramllm-0.1.0/.hypothesis/constants/315c6b8b4649d1f5

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/chat/toolbar.py
+# hypothesis_version: 6.169.1
+
+[1000, 1024, 4096, ' │ ', 'ChatSession', '…']

ramllm-0.1.0/.hypothesis/constants/31a5d9fe538978dd

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/commands/chat_cmd.py
+# hypothesis_version: 6.169.1
+
+[0.7, 0.9, 35.0, 100, 4096, 8085, '--attach-server', '--ctx', '--memory', '--mode', '--no-save', '--plain', '--port', '--prompt', '--resume', '--seed', '--system', '--temp', '--top-p', '-p', '/', '127.0.0.1', 'BATCH_ONLY', 'COMFORTABLE', 'CONSTRAINED', 'INFEASIBLE', 'Launch anyway?', 'Sampling temperature', 'attached-server', 'qwen-2.5-7b', 'system', 'user']

ramllm-0.1.0/.hypothesis/constants/3401fa2e1bf8909c

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/planner/explain.py
+# hypothesis_version: 6.169.1
+
+[1024, 'GPU VRAM', 'Host RAM', 'compute', 'cpu', 'disk_bw', 'expert_cache_miss', 'gpu', 'pcie_bw', 'ram_bw', 'vram_bw']

ramllm-0.1.0/.hypothesis/constants/3645be7b2a65bcc8

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/ui/console.py
+# hypothesis_version: 6.169.1
+
+[]

ramllm-0.1.0/.hypothesis/constants/37d53b76366447eb

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/ui/glossary.py
+# hypothesis_version: 6.169.1
+
+['Context Window', 'Feasibility Verdicts', 'Model Layers', 'System RAM', 'Video RAM (VRAM)', 'bottleneck', 'context', 'explanation', 'kv-cache', 'layers', 'moe', 'quantization', 'ram', 'ram-e001', 'ram-e002', 'ram-e005', 'ram-e006', 'ram-e007', 'ram-e008', 'ram-e009', 'ram-e010', 'ram-e011', 'ram-e012', 'ram-e013', 'ram-e014', 'ram-e015', 'tip', 'title', 'verdicts', 'vram']

ramllm-0.1.0/.hypothesis/constants/37e8925d3a8d15b0

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/ui/console.py
+# hypothesis_version: 6.169.1
+
+[]

ramllm-0.1.0/.hypothesis/constants/389870ed1fd87ac1

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/governor/enforcement/factory.py
+# hypothesis_version: 6.169.1
+
+['darwin', 'linux', 'win32']

ramllm-0.1.0/.hypothesis/constants/399e77c8ff4258e6

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/app.py
+# hypothesis_version: 6.169.1
+
+[40.0, 60.0, 128, 200, 1024, 8080, 8085, ' (Unified Memory)', ',', ', ', '-', '--context', '--cpu-only', '--debug', '--dest', '--explain', '--gpus', '--host', '--hw-json', '--json', '--live', '--max-retries', '--no-kv-quant', '--plain', '--plan-id', '--port', '--prefer', '--prompt', '--ram', '--refresh', '--sha256', '--vram', '-p', '0.0 GB', '127.0.0.1', '4096', 'Arch', 'BATCH_ONLY', 'Benchmark Results', 'Bottleneck', 'CHAT & RUN', 'COMFORTABLE', 'CONSTRAINED', 'CPU', 'CPU Features', 'Component', 'Decode Speed', 'GET STARTED', 'GPU', 'Hardware Profile', 'INFEASIBLE', 'Key Advantage', 'MODELS', 'Measured', 'Metric', 'MoE', 'Model', 'Model Catalog', 'Name', 'No', 'Operating System', 'Output JSON', 'Output raw JSON', 'Params', 'Path', 'Path to GGUF model', 'Peak RAM RSS', 'Port to listen', 'Predicted', 'Prefill Speed', 'Quant', 'RAM Bandwidth', 'RAM Peak', 'Rank', 'SYSTEM', 'Size', 'Specification', 'Speed', 'Storage Drive', 'Strategy', 'System RAM', 'VRAM Peak', 'Verdict', 'Yes', 'add', 'backend', 'balanced', 'blue', 'bold', 'bold cyan', 'bold green', 'choices', 'content', 'cyan', 'dim', 'doctor', 'explain', 'fastest', 'fit_at_any_cost', 'green', 'info', 'json', 'list', 'magenta', 'max_tokens', 'message', 'messages', 'models', 'pull', 'ramllm', 'remove', 'rich', 'role', 'scan', 'standard', 'stream', 'user', 'verify', 'white', 'yellow']

ramllm-0.1.0/.hypothesis/constants/39d730a575ffcb6a

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/profiler/memory.py
+# hypothesis_version: 6.169.1
+
+[1e-06, 30.0, 1024]

ramllm-0.1.0/.hypothesis/constants/3b5c2c70b9b055aa

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/core/planner/candidates.py
+# hypothesis_version: 6.169.1
+
+[0.25, 0.5, 1.0, 1.5, 300.0, 500, 512, 1024, ',', '--no-kv-offload', '-c', '-cmoe', '-ctk', '-ctv', '-lm', '-m', '-mg', '-ncmoe', '-ngl', '-sm', '-ts', 'apple', 'cpu', 'f16', 'gpu', 'layer', 'llamacpp', 'macos', 'none', 'q4_0', 'q8_0']

ramllm-0.1.0/.hypothesis/constants/3ea55bb8ee73a085

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/api/server.py
+# hypothesis_version: 6.169.1
+
+[40.0, 120.0, 500, 503, 4096, 18080, 1700000000, '/healthz', '/ramllm/metrics', '/ramllm/plan', '/ramllm/replan', '/v1/chat/completions', '/v1/completions', '/v1/models', '0.1.0', 'INFEASIBLE', 'POST', 'RAMLLM Server', 'created', 'data', 'id', 'list', 'model', 'object', 'ok', 'owned_by', 'ramllm', 'ramllm-model', 'status', 'stream', 'text/event-stream']

ramllm-0.1.0/.hypothesis/constants/3eb5a1362f5deaa9

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/main.py
+# hypothesis_version: 6.169.1
+
+[40.0, 60.0, 128, 200, 1024, 8080, 8085, ' (Unified Memory)', ',', ', ', '-', '--context', '--cpu-only', '--dest', '--explain', '--gpus', '--host', '--hw-json', '--json', '--live', '--no-kv-quant', '--plan-id', '--port', '--prefer', '--prompt', '--ram', '--refresh', '--sha256', '--vram', '-p', '0.0 GB', '127.0.0.1', '4096', 'Arch', 'BATCH_ONLY', 'Benchmark Results', 'Bottleneck', 'COMFORTABLE', 'CONSTRAINED', 'CPU', 'CPU Features', 'Component', 'Decode Speed', 'Flash Attention', 'GPU', 'GPU Offload', 'Hardware Profile', 'INFEASIBLE', 'KV Quantization', 'Key Advantage', 'Measured', 'Metric', 'MoE', 'MoE CPU Offload', 'Model', 'Model Catalog', 'Name', 'No', 'Operating System', 'Output JSON', 'Output raw JSON', 'Params', 'Path', 'Path to GGUF model', 'Peak RAM RSS', 'Port to listen', 'Predicted', 'Prefill Speed', 'Quant', 'RAM Bandwidth', 'RAM Peak', 'Rank', 'Size', 'Specification', 'Speed', 'Storage Drive', 'Strategy', 'System RAM', 'VRAM Peak', 'Verdict', 'Yes', '__main__', 'add', 'balanced', 'blue', 'bold', 'bold cyan', 'bold green', 'choices', 'content', 'cyan', 'dim', 'fastest', 'fit_at_any_cost', 'green', 'info', 'json', 'list', 'magenta', 'max_tokens', 'message', 'messages', 'models', 'pull', 'ramllm', 'remove', 'role', 'scan', 'standard', 'stream', 'user', 'verify', 'white', 'yellow']

ramllm-0.1.0/.hypothesis/constants/407373161881ba25

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/commands/explain_cmd.py
+# hypothesis_version: 6.169.1
+
+['Available terms:', 'tip']

ramllm-0.1.0/.hypothesis/constants/43780037291daff2

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/chat/slash.py
+# hypothesis_version: 6.169.1
+
+[1024, 4096, '-selection', '/', '/budget', '/clear', '/copy', '/ctx', '/exit', '/help', '/load', '/mem', '/mode', '/model', '/plan', '/quit', '/retry', '/save', '/seed', '/serve', '/stats', '/system', '/temp', '/toolbar', '/top_p', '/undo', 'ChatSession', 'Command', 'Description', 'assistant', 'auto', 'balanced', 'bold cyan', 'clip', 'clipboard', 'cyan', 'darwin', 'enabled', 'fast', 'hidden', 'max', 'off', 'on', 'pbcopy', 'quality', 'user', 'utf-8', 'white', 'windows', 'xclip']

ramllm-0.1.0/.hypothesis/constants/4421a3e2a3920769

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/backends/llamacpp/launcher.py
+# hypothesis_version: 6.169.1
+
+[0.3, 0.5, 30.0, 200, 1024, 8080, 'GB', 'GIB', 'KB', 'KIB', 'MB', 'MIB', 'compute', 'kv', 'model']

ramllm-0.1.0/.hypothesis/constants/445382ac9051d875

@@ -0,0 +1,4 @@
+# file: /Users/cassain/Projects/RAMLLM/src/ramllm/cli/commands/config_cmd.py
+# hypothesis_version: 6.169.1
+
+['(none)', 'Description', 'EDITOR', 'SYSTEM', 'Setting', 'VISUAL', 'Value', 'bold green', 'cyan', 'edit', 'get', 'list', 'nano', 'notepad', 'nt', 'path', 'reset', 'set', 'white']