10 Commits

Author SHA1 Message Date
212832e7e4 feat: expose protocol analysis, OOT export tools; harden for release
Wire up protocol analysis (parse_protocol_spec, generate_decoder_chain,
get_missing_oot_modules), signal analysis (analyze_iq_file), and OOT
export (generate_oot_skeleton, export_block_to_oot, export_from_flowgraph)
as MCP tools with integration tests.

Security fixes from Hamilton review:
- Remove `from __future__ import annotations` from tool registration
  files (breaks FastMCP schema generation)
- Add blocklist guard to evaluate_expression (was unsandboxed eval)
- Replace string interpolation with base64 encoding in Docker test
  harness (prevents code injection)
- Add try/finally cleanup for temp files and Docker containers
- Replace assert with proper ValueError in flowgraph block creation
- Log OOT auto-discovery failures instead of swallowing silently

Packaging:
- Move entry point to src/gnuradio_mcp/server.py with script entry
  point (uv run gnuradio-mcp)
- Add PyPI metadata (authors, license, classifiers, urls)
- Add MIT LICENSE file
- Rewrite README for current feature set (80+ tools)
- Document single-session limitation
2026-02-20 13:17:11 -07:00
5db7d71d2b feat: add AI-assisted block development tools
Implements complete workflow for generating GNU Radio blocks from
descriptions:

Block Generation:
- generate_sync_block, generate_basic_block, generate_interp_block,
  generate_decim_block tools for creating different block types
- Template-based code generation with customizable work logic
- Automatic validation via AST parsing and signature checking

Protocol Analysis:
- Parse protocol specifications into structured models
- Generate decoder pipelines matching modulation to demodulator blocks
- Templates for BLE, Zigbee, LoRa, POCSAG, ADS-B protocols

OOT Export:
- Export generated blocks to full OOT module structure
- Generate CMakeLists.txt, block YAML, Python modules
- gr_modtool-compatible output

Dynamic Tool Registration:
- enable_block_dev_mode/disable_block_dev_mode for context management
- Tools only registered when needed (reduces LLM context usage)

Includes comprehensive test coverage and end-to-end demo.
2026-02-09 12:36:54 -07:00
e63f6e1ba0 test: add coverage for dynamic runtime mode and client capabilities
Add integration tests for:
- TestDynamicRuntimeMode (5 tests): verify enable/disable toggle
  behavior, tool registration/removal, idempotency
- TestClientCapabilities (2 tests): verify structured capability
  output and roots listing

Total test count: 370 → 377
2026-02-02 13:08:18 -07:00
3811099623 feat: dynamic runtime tool registration to reduce context usage
Runtime tools now register on-demand rather than at startup:
- get_runtime_mode(): check mode status and available capabilities
- enable_runtime_mode(): dynamically register 36 runtime tools
- disable_runtime_mode(): remove runtime tools when not needed

At startup, only 33 design-time tools are registered. When runtime mode
is enabled, tool count increases to 69. This reduces context usage
significantly when only doing flowgraph design work.

Uses FastMCP's add_tool/remove_tool API for dynamic registration,
following MCP spec's notifications/tools/list_changed pattern.
2026-02-02 02:11:53 -07:00
e8b3600e60 feat: add LoRa SDR receiver with Docker runtime infrastructure
LoRa receiver flowgraph built programmatically via gr-mcp:
- osmosdr_source → low_pass_filter → lora_rx → message_debug
- XML-RPC server for runtime variable control (samp_rate,
  center_freq) with introspection enabled
- Qt frequency sink for spectrum visualization

Docker infrastructure:
- gnuradio-lora: gr-lora_sdr OOT module from EPFL (chirp spread spectrum)
- gnuradio-lora-runtime: combined runtime with Xvfb + gr-lora_sdr
- Compose file, entrypoint, and launch script for LoRa receiver

Also includes:
- lora_scanner.py: multi-SF LoRa scanner example
- lora_infrastructure_test.py: hardware-free pipeline validation
  (signal_source → throttle → null_sink + xmlrpc variable control)
- Integration tests for LoRa scanner flowgraph construction

End-to-end pipeline validated: launch_flowgraph → connect_to_container →
list_variables → get/set_variable all working through Docker + XML-RPC.
2026-01-30 13:55:40 -07:00
fdcbffba1a tests: add integration tests for runtime middleware
Add comprehensive integration tests that don't require Docker:

- test_xmlrpc_subprocess.py: Tests XmlRpcMiddleware against a real
  subprocess-based XML-RPC server. Covers connection, variable
  discovery, get/set operations, and flowgraph control (18 tests).

- test_mcp_runtime.py: Tests MCP runtime tools via FastMCP Client.
  Verifies connect, disconnect, list_variables, get_variable,
  set_variable, and flowgraph control work end-to-end (15 tests).

- test_fm_scanner.py: Tests signal probe functionality and FM scanner
  parsing. Includes unit tests for helper functions and integration
  tests for flowgraph construction with signal probe blocks (18 tests).

All 59 integration tests pass. These tests provide faster feedback
than Docker-based tests while still exercising real XML-RPC
communication and flowgraph construction.
2026-01-29 05:14:29 -07:00
0afb2f5b6e runtime: Phase 2 ControlPort/Thrift integration
Add ControlPort/Thrift support as an alternative transport to XML-RPC:

New middleware:
- ThriftMiddleware wrapping GNURadioControlPortClient

New MCP tools:
- connect_controlport, disconnect_controlport
- get_knobs (with regex filtering), set_knobs (atomic)
- get_knob_properties (units, min/max, description)
- get_performance_counters (throughput, timing, buffers)
- post_message (PMT injection to block ports)

Docker support:
- enable_controlport param in launch_flowgraph
- ENABLE_CONTROLPORT env in entrypoint.sh
- ControlPort config generation in ~/.gnuradio/config.conf

Models: KnobModel, KnobPropertiesModel, PerfCounterModel,
ThriftConnectionInfoModel, plus ContainerModel updates.
2026-01-28 12:05:32 -07:00
91a442cdf9 tests: add RuntimeProvider coverage (82% → 94%)
Unit tests (35):
- Mocked DockerMiddleware and XmlRpcMiddleware
- Full coverage of orchestration logic, preconditions, state tracking

Integration tests (6):
- Real Docker containers with gnuradio-runtime image
- End-to-end: launch → connect → control variables → stop
- Skip gracefully when Docker/image unavailable

Total: 112 tests, 94% coverage
2026-01-27 10:16:47 -07:00
fdc42c9bed Upgrade to FastMCP 3.0 and Python 3.14
- Bump version to 0.2.0
- Require Python 3.14+
- Update to FastMCP 3.0.0b1 with new Client API
- Update tests for FastMCP 3.0 (Pydantic models in .data)
- Pin modern dependency versions
2026-01-26 16:46:05 -07:00
Yoel bassin
664c2adedc add tests for FastMCP integration and update pre-commit 2026-01-05 20:52:48 +02:00