v1.0.0

🎉 1.0.0

This release marks the 1.0.0 milestone for Daqifi.Core. The public API is now stable under semantic versioning — breaking changes will bump the major version. The library supports DAQiFi firmware v3.5.0+ (per ADR 0001), and it is the foundation the DAQiFi desktop application is migrating onto.

What's new

📊 Decoded per-frame sample pipeline (#279)

Core now decodes each live stream-data frame into per-channel samples and raises IChannel.SampleReceived, instead of only re-raising the raw DaqifiOutMessage. Consumers no longer have to hand-demux frames — channel ordering, calibration scaling, digital bit-unpacking, and rollover-aware timestamps all live in Core. Closes #242.

• USB pre-scaled floats (AnalogInDataFloat) are used directly; WiFi raw ADC counts (AnalogInData) are scaled per channel via IAnalogChannel.GetScaledValue. Digital bytes are bit-unpacked to per-channel high/low for input-direction channels.
• New public API: IDataSample.RawValue (raw ADC count, null on the float path), IDataSample.DeviceTimestamp (raw device tick of the frame), SampleReceivedEventArgs.Channel, and an IChannel.SetActiveSample(IDataSample) overload.
• Existing MessageReceived consumers are unchanged — the raw re-raise still fires first. Decoding runs only while IsStreaming and is best-effort per frame, so a malformed frame can't tear down the stream.

Validated end-to-end on a real Nyquist 1 over USB serial, exercising the calibration-scaling path with monotonic frame-aligned device timestamps.

🔌 Digital output, first-class (#280)

Digital output (SetDioDirection / SetDioValue) existed in Core but was undocumented, unexposed over MCP, and unverified on hardware. It's now a first-class, hardware-validated capability — including a stream decode fix: digital stream bits are now indexed by channel number (bit N is pin N, matching how the firmware snapshots the whole DIO port), instead of by position in the enabled-channel list. Previously, enabling a subset of digital channels read the wrong pins — e.g. enabling only DIO 4 read pin 0's state.

• New MCP tools: configure_digital_channels, set_digital_direction, and set_digital_output (auto-switches an input channel to output so one call drives a pin). list_channels now reports OutputValue for digital channels; get_device_status reports EnabledDigitalChannels.
• Docs: README capability row + digital-output recipe; MCP tool table in src/Daqifi.Mcp/README.md.

Validated on a Nyquist 1 over USB with physical loopbacks — raw SCPI and Core end-to-end, including streaming a strict subset of digital channels (the exact scenario the old positional decode fails).

⚡ PWM output (#281)

Adds PWM output as a first-class, hardware-validated capability, building on #280.

• Device API: SetPwmEnabled(channel, bool) and SetPwmDutyCycle(channel, 1–100) are per channel; SetPwmFrequency(6–50,000 Hz) is device-wide, because all PWM channels share one hardware timer — a per-channel frequency API would mislead. IDigitalChannel gains IsPwmCapable (populated from the firmware board mask — channels 0, 3, 4, 5, 6, 7), plus IsPwmEnabled / PwmDutyCyclePercent bookkeeping that survives status refresh.
• Client-side guardrails for confirmed firmware traps: enabling PWM on a non-capable channel throws (the firmware would otherwise freeze digital writes on that channel with disable as the only recovery — so disable is accepted on any channel), duty 0 is rejected (firmware readback would claim 0 while the old duty keeps toggling), and frequency is limited to 6–50,000 Hz (below 6 Hz the firmware's period register silently wraps).
• New MCP tools: set_pwm_output (duty + optional device-wide frequency + enable in one call) and disable_pwm. list_channels now reports PwmCapable / PwmEnabled.

Validated on a Nyquist 1 over USB: commanded duty and frequency measured through a physically wired loopback from streamed samples (exactly 40 rising edges in 4 s at 10 Hz), including a live duty change and disable → drive-low handback.

Full Changelog: v0.28.0...v1.0.0