API 参考¶

本页列出 wavekit 的主要公共 API。API 的类名、函数名和签名保持代码中的英文形式，便于与实际调用对应；具体 API 说明由源代码自动生成。

Waveform¶

Waveform 是波形分析的核心对象，包含采样值以及对应的周期和时间轴。

`Waveform(value, clock, time, width=None, signed=False, signal=None)` ¶

A clock-synchronised, numpy-backed time series of a hardware signal.

Every Waveform is a triple of parallel numpy arrays of equal length:

value — signal values sampled on every clock edge.
clock — absolute sampling-edge number counted from the first selected edge in the waveform file (which is cycle 0).
time — simulation timestamp (in the file's native time unit) of each sample.

These three arrays are always kept in sync; any operation that filters or transforms values also transforms the corresponding clock and time entries so positional alignment is preserved.

The convenience property data returns them as a single numpy record array with fields ("time", "clock", "value") for easy pandas / numpy interop.

width and signed are direct fields on Waveform. signal is an optional Signal that is set only when the waveform was loaded directly from a reader.

Bit-width rules

Widths ≤ 64 bits are stored as np.int64 (signed) or np.uint64 (unsigned).
Widths > 64 bits are stored as Python object arrays (arbitrary precision integers).
Arithmetic operators automatically infer a result width (e.g. addition widens by 1 bit; multiplication sums both widths). Width inference is capped at 64 bits for integer types.
Two Waveform operands must have the same signedness; mixing signed and unsigned raises ValueError.

Typical usage

Waveforms are normally created by a Reader, not constructed directly.

with VcdReader("sim.vcd") as r:
    data = r.load_waveform("tb.dut.data[7:0]", clock="tb.clk")
    valid = r.load_waveform("tb.dut.valid", clock="tb.clk")

valid_data = data.mask(valid == 1)
print(valid_data.value)

`data` `property` ¶

Return value/clock/time as a single numpy record array.

Fields: "time" (simulation timestamp), "clock" (edge counter), "value" (signal value). Useful for pandas conversion or bulk numpy operations that need all three columns together.

`ahead(n=1, pad='repeat', pad_value=None)` ¶

Return a new Waveform looking n cycles into the future.

Convenience wrapper around relative() for positive offsets.

Parameters:

Name	Type	Description	Default
`n`	`int`	Number of cycles to look ahead. Default is 1.	`1`
`pad`	`Literal['repeat', 'value']`	See `relative()` for the padding options.	`'repeat'`
`pad_value`	`Any`	See `relative()` for an example.	`None`

Returns:

Type	Description
`Waveform`	A new waveform shifted forward by n cycles.

Example

# Rising edge detection
rising = (wave == 0) & wave.ahead()

`any_edge()` ¶

Detect either edge in a 1-bit waveform.

Returns a new unsigned 1-bit Waveform that is true for both 0 -> 1 and 1 -> 0 transitions. The first sample is always false.

Raises:

Type	Description
`ValueError`	If the source waveform is not one bit wide.

`as_signed()` ¶

Reinterpret the unsigned bit pattern as a two's-complement signed integer.

Raises ValueError if width is unknown (None). Returns a copy if the waveform is already signed.

`as_unsigned()` ¶

Reinterpret the signed value as an unsigned bit pattern.

Raises ValueError if width is unknown (None). Returns a copy if the waveform is already unsigned.

`back(n=1, pad='repeat', pad_value=None)` ¶

Return a new Waveform looking n cycles into the past.

Convenience wrapper around relative() for negative offsets.

Parameters:

Name	Type	Description	Default
`n`	`int`	Number of cycles to look back. Default is 1.	`1`
`pad`	`Literal['repeat', 'value']`	See `relative()` for the padding options.	`'repeat'`
`pad_value`	`Any`	See `relative()` for an example.	`None`

Returns:

Type	Description
`Waveform`	A new waveform shifted backward by n cycles.

Example

# Check if current value equals previous
same = wave == wave.back()

`bit_count()` ¶

Count the number of set bits (population count) in each sample value.

Returns a new unsigned Waveform with width=64 where each value is the popcount of the corresponding source sample. Supports arbitrarily wide signals (> 64 bits) by chunking.

Raises:

Type	Description
`ValueError:`	If `self.width` is `None`.

`changed()` ¶

Detect changes between adjacent samples.

Returns a new unsigned 1-bit Waveform where value[i] is true when the current sample differs from the previous sample. The first sample is always false.

`compress()` ¶

Compact a waveform while preserving value changes and the final sample.

This removes redundant samples inside each stable run, but keeps the first sample of each value run and always keeps the final sample. The result still records both where values change and where the original sampled window ends.

Use this when reducing waveform size for display, export, or quick inspection without losing the final sample.

Returns:

Type	Description
`Waveform`	A compacted waveform. It may still contain one repeated final value when the last stable run has more than one sample, because the final sample is preserved intentionally.

`concatenate(waves)` `staticmethod` ¶

Concatenate multiple waveforms into a single wider waveform.

Bits are joined so that the last element in waves becomes the MSB group and the first becomes the LSB group — this is the inverse of split_bits().

All waveforms in waves must be unsigned and have the same length. The result width is the sum of all input widths.

Parameters:

Name	Type	Description	Default
`waves`	`list[Waveform]`	List of waveforms to concatenate; at least one element required.	required

Raises:

Type	Description
`Exception:`	If any waveform is signed.
`ValueError:`	If any waveform has `width=None`.

Example

# Recombine four 8-bit bytes into one 32-bit value (byte3 = MSB)
bus32 = Waveform.concatenate([byte0, byte1, byte2, byte3])

`copy()` ¶

Return a deep copy of this waveform.

Returns:

Type	Description
`Waveform`	New waveform with copied `value`, `clock`, and `time` arrays and the same `width` / `signed` metadata.

`cycle_slice(begin_cycle=None, end_cycle=None, include_end=False)` ¶

Return a new Waveform trimmed to the given absolute clock cycle range.

Uses binary search on the sorted clock array so the operation is O(log n) regardless of waveform length. The clock values are absolute cycle numbers from the start of simulation (not relative to this waveform's window), so the same cycle number means the same simulation instant across different waveforms.

Parameters:

Name	Type	Description	Default
`begin_cycle`	`int \| None`	First clock cycle to include (inclusive). Defaults to the first sample's cycle number.	`None`
`end_cycle`	`int \| None`	Last clock cycle. Exclusive by default; set `include_end=True` to make it inclusive.	`None`
`include_end`	`bool`	If `True`, samples exactly at end_cycle are included.	`False`

See Also

time_slice : slice by simulation timestamp instead of cycle number.

Example

# Analyse cycles 100 to 199 (exclusive end)
window = wave.cycle_slice(100, 200)

`downsample(chunk_size, func=np.mean)` ¶

Reduce the sample rate by aggregating consecutive chunks.

Splits value, clock, and time into non-overlapping windows of chunk_size and applies func to each window. The result length is ceil(len / chunk_size).

Parameters:

Name	Type	Description	Default
`chunk_size`	`int`	Number of consecutive samples to aggregate into one.	required
`func`	`Callable[[NDArray[Any]], float]`	Aggregation function applied to each value chunk. Defaults to `np.mean`. `clock` and `time` chunks are always averaged.	`mean`

Example

# Average occupancy in 100-cycle windows
avg = occupancy.downsample(100, np.mean)

`falling_edge()` ¶

Detect 1→0 transitions in a 1-bit waveform.

Returns a new 1-bit Waveform where value[i] == True if and only if self.value[i-1] == 1 and self.value[i] == 0. The first sample is always False.

Raises:

Type	Description
`ValueError`	If the source waveform is not one bit wide.

See Also

rising_edge : detect 0→1 transitions. any_edge : detect either transition.

`filter(condition)` ¶

Return a new Waveform keeping only the samples that satisfy condition.

condition is called once per sample value (scalar, not vectorized). For large waveforms prefer vectorized_filter().

Parameters:

Name	Type	Description	Default
`condition`	`Callable[[Any], bool]`	A callable that accepts a single value and returns `bool`.	required

Example

non_zero = wave.filter(lambda v: v != 0)

`map(func, width=None, signed=None)` ¶

Apply a scalar function element-wise and return a new Waveform.

Internally wraps func with np.vectorize. For large waveforms prefer vectorized_map() with a native numpy operation.

Parameters:

Name	Type	Description	Default
`func`	`Callable[[Any], Any]`	Callable applied to each value element individually.	required
`width`	`int \| None`	Bit-width of the result. `None` if not known (default).	`None`
`signed`	`bool \| None`	Signedness of the result. Defaults to `False`.	`None`

Example

upper_nibble = wave.map(lambda v: (v >> 4) & 0xF, width=4, signed=False)

`mask(mask)` ¶

Return a new Waveform keeping only the samples where mask is True.

Parameters:

Name	Type	Description	Default
`mask`	`NDArray[bool_] \| Waveform`	Either a boolean `np.ndarray` or a 1-bit `Waveform` (`width == 1` or `dtype == bool`). Must have the same length as `self`.	required

Raises:

Type	Description
`TypeError:`	If mask is a Waveform with width != 1, or is not a boolean array.

Example

valid_data = data.mask(valid == 1)   # keep only cycles where valid is high

`merge(waves, func, width, signed)` `staticmethod` ¶

Combine multiple same-length waveforms into one using a custom function.

At each sample index i, func is called with [w.value[i] for w in waves] and the returned value becomes the result sample. clock and time are taken from waves[0].

All waveforms must have the same number of samples.

Parameters:

Name	Type	Description	Default
`waves`	`list[Waveform]`	Input waveforms; must be non-empty and all equal in length.	required
`func`	`Callable[[list[Any]], Any]`	Callable `(list[scalar]) -> scalar` applied per sample.	required
`width`	`int`	Bit-width to assign to the result.	required
`signed`	`bool`	Signedness of the result.	required

Example

# Compute bitwise majority across three 1-bit signals
majority = Waveform.merge(
    [a, b, c], lambda vs: int(sum(vs) >= 2), width=1, signed=False
)

`relative(offset, pad='repeat', pad_value=None)` ¶

Return a new Waveform shifted by offset cycles.

This is the core method for relative time access. Use ahead() and back() for more readable positive/negative offsets.

Parameters:

Name	Type	Description	Default
`offset`	`int`	Number of cycles to shift. Positive looks forward (future), negative looks backward (past).	required
`pad`	`Literal['repeat', 'value']`	Boundary handling strategy: `'repeat'` (default): pad with boundary value (first/last element). `'value'`: pad with pad_value (must be provided).	`'repeat'`
`pad_value`	`Any`	Value to use when `pad='value'`. Ignored otherwise.	`None`

Returns:

Type	Description
`Waveform`	A new waveform shifted by offset cycles. `clock` and `time` arrays are always preserved unchanged.

Raises:

Type	Description
`ValueError:`	If `pad='value'` but pad_value is not provided.
`ValueError:`	If pad is not one of `'repeat'`, `'value'`.

Example

# Rising edge detection
rising = (wave == 0) & wave.ahead()

# Look back 3 cycles
past = wave.relative(-3)

See Also

ahead : Shift forward (positive offset). back : Shift backward (negative offset).

`rising_edge()` ¶

Detect 0→1 transitions in a 1-bit waveform.

Returns a new 1-bit Waveform where value[i] == True if and only if self.value[i-1] == 0 and self.value[i] == 1. The first sample is always False.

Raises:

Type	Description
`ValueError`	If the source waveform is not one bit wide.

See Also

falling_edge : detect 1→0 transitions. any_edge : detect either transition.

`slice(begin_idx, end_idx, include_end=False)` ¶

Return a new Waveform trimmed to the given sample index range.

Parameters:

Name	Type	Description	Default
`begin_idx`	`int`	First sample index to include (inclusive).	required
`end_idx`	`int`	Last sample index. Exclusive by default; set `include_end=True` to make it inclusive.	required
`include_end`	`bool`	If `True`, the sample at end_idx is included.	`False`

See Also

time_slice : slice by simulation timestamp instead of array index.

`split_bits(bit_group_size, padding=False)` ¶

Split the waveform into multiple narrower waveforms by bit groups.

Parameters:

Name	Type	Description	Default
`bit_group_size`	`int \| list[int]`	`int` — split into equal-sized groups of this many bits, starting from bit 0 (LSB). `width` must be a multiple of bit_group_size unless `padding=True`. `list[int]` — explicit widths for each group (LSB-first). The values must sum to `self.width`; `padding` is ignored.	required
`padding`	`bool`	If `True` and an integer bit_group_size is given, the last group may be narrower than bit_group_size.	`False`

Returns:

Type	Description
`list[Waveform]:`	Waveforms ordered from LSB group to MSB group, each unsigned.

Raises:

Type	Description
`ValueError:`	If `self.width` is `None`.
`Exception:`	If width is not divisible by bit_group_size when `padding=False`, or if list sizes do not sum to `self.width`.

Example

# Split a 32-bit bus into four 8-bit bytes (byte0 = bits[7:0])
bytes_ = bus32.split_bits(8)

`take(indices)` ¶

Return a new Waveform selecting samples at the given integer positions.

Parameters:

Name	Type	Description	Default
`indices`	`NDArray[integer] \| list[int] \| Waveform`	Integer index array, list of ints, or a `Waveform` whose `value` array contains integer indices (e.g. the result of `np.where`). Boolean arrays are not accepted; use `mask()` instead.	required

Raises:

Type	Description
`TypeError:`	If indices contains booleans or non-integer values.

Example

# Keep every other sample
even = wave.take(list(range(0, len(wave.value), 2)))

`time_slice(begin_time=None, end_time=None, include_end=False)` ¶

Return a new Waveform trimmed to the given simulation time range.

Uses binary search on the sorted time array so the operation is O(log n) regardless of waveform length.

Parameters:

Name	Type	Description	Default
`begin_time`	`int \| None`	Start of the time window (inclusive). Defaults to the first sample's timestamp.	`None`
`end_time`	`int \| None`	End of the time window. Exclusive by default; set `include_end=True` to make it inclusive.	`None`
`include_end`	`bool`	If `True`, samples exactly at end_time are included.	`False`

Example

# Analyse only the first 1000 simulation time units
early = wave.time_slice(0, 1000)

`unique_consecutive()` ¶

Remove consecutive duplicate values.

Equivalent to run-length deduplication: each run of equal consecutive values is represented by its first sample. This matches the usual unique_consecutive semantics and is useful when only the sequence of observed values matters.

Returns:

Type	Description
`Waveform`	A new waveform with consecutive duplicate values removed. The final timestamp may be dropped when the last value spans multiple samples.

`vectorized_filter(func)` ¶

Return a new Waveform keeping only the samples where func returns True.

func receives the entire value array at once and must return a boolean array of the same length. Prefer this over filter() for performance-critical paths.

Parameters:

Name	Type	Description	Default
`func`	`Callable[[NDArray[Any]], NDArray[bool_]]`	Vectorized callable: `(NDArray) -> NDArray[bool]`.	required

`vectorized_map(func, width=None, signed=None)` ¶

Apply a vectorized function to the value array and return a new Waveform.

func receives the entire value ndarray and must return an ndarray of the same length. clock and time are deep-copied unchanged.

Parameters:

Name	Type	Description	Default
`func`	`Callable[[NDArray[Any]], NDArray[Any]]`	Vectorized callable: `(NDArray) -> NDArray`.	required
`width`	`int \| None`	Bit-width of the result. `None` if not known (default).	`None`
`signed`	`bool \| None`	Signedness of the result. Defaults to `False`.	`None`

Reader¶

不同格式的 Reader 使用同一套加载、查询和表达式求值 API。

`VcdReader(file)` ¶

Bases: Reader

Read VCD waveform files via vcdvcd.

Supports the common Reader APIs, including hierarchy traversal, expression evaluation, and clock-synchronised waveform loading.

`begin_time` `property` ¶

Return the first timestamp stored in the VCD file.

`end_time` `property` ¶

Return the last timestamp stored in the VCD file.

`top_scopes` `cached` `property` ¶

Return immutable top-level scopes in the VCD hierarchy.

`close()` ¶

Close this VCD reader.

The underlying vcdvcd reader keeps data in memory and does not expose an explicit close operation, so this method is a no-op.

`eval(expr, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, mode='single', root_scope=None)` ¶

Evaluate a waveform expression containing physical signal paths.

Parameters:

Name	Type	Description	Default
`expr`	`str`	Expression string. Signal paths may be used as operands or as arguments to registered expression functions.	required
`clock`	`str`	Clock signal used for all waveform loads.	required
`xz_value`	`int`		`0`
`signed`	`int`		`0`
`sample_on_posedge`	`int`		`0`
`begin_time`	`int`		`0`
`end_time`	`int`		`0`
`begin_cycle`	`int`		`0`
`end_cycle`	`int \| None`	Forwarded to `load_matched_waveforms` for every path.	`None`
`mode`	`Literal['single', 'zip']`	`'single'` requires every path to match one signal. `'zip'` evaluates once per shared multi-match key and broadcasts singleton paths.	`'single'`
`root_scope`	`Scope \| None`	If provided, resolve paths within this scope.	`None`

Returns:

Type	Description
`Waveform or dict[tuple[Capture, ...], Waveform]`	The evaluated waveform, or one waveform per zip key.

`get_matched_scopes(path, root_scope=None)` ¶

Return all scopes whose paths match path, keyed by captures.

Similar to get_matched_signals but stops at the scope level — the last component of path must match a scope name, not a signal. Useful for enumerating module instances before loading their signals.

Parameters:

Name	Type	Description	Default
`path`	`str`	Scope query path using the same syntax as signal paths. The last component must match a scope (module) name, e.g. `"tb.dut.fifo_{0..3}"` or `r"tb./([a-z]+)_core/"`.	required
`root_scope`	`Scope \| None`	If provided, search only within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Scope]:`	Maps each capture key to the matched `Scope`. Ordinary exact-name matches are omitted from the key, so a query without binding matchers uses `()`.

Raises:

Type	Description
`ValueError:`	If two different scopes resolve to the same key, or if using module matchers on a backend without `definition` support (VCD/FST), or if the path contains a terminal signal bit-range suffix.

`get_matched_signals(path, root_scope=None)` ¶

Return all signals whose paths match path, keyed by captures.

Traverses the scope tree starting from root_scope (or the file's top-level scopes if root_scope is None) and applies the query path to each level. See the class docstring for query path syntax.

Parameters:

Name	Type	Description	Default
`path`	`str`	Signal query path, e.g. `"tb.dut.fifo_{0..3}.w_ptr[2:0]"` or `r"tb.dut./([a-z]+)_valid/"`.	required
`root_scope`	`Scope \| None`	If provided, search only within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Signal]:`	Maps each capture key to the matched `Signal` object (carrying name, width, range, signed). Ordinary exact-name matches are omitted from the key, so a query without binding matchers uses `()`.

Raises:

Type	Description
`ValueError:`	If two different signals resolve to the same key, or if using module matchers on a backend without `definition` support (VCD/FST).

`get_scope(path, root_scope=None)` ¶

Return the scope at an exact hierarchy path.

Parameters:

Name	Type	Description	Default
`path`	`str`	Exact dotted scope path. Matcher expressions and terminal signal ranges are not accepted.	required
`root_scope`	`Scope \| None`	If provided, resolve path within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`Scope`	The exact matched scope.

Raises:

Type	Description
`ValueError`	If path contains matcher syntax, contains a terminal range, or the scope does not exist.

`get_signal(path, root_scope=None)` ¶

Return the signal at an exact hierarchy path.

Parameters:

Name	Type	Description	Default
`path`	`str`	Exact dotted signal path, with an optional terminal bit selection or range. Matcher expressions are not accepted.	required
`root_scope`	`Scope \| None`	If provided, resolve path within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`Signal`	The exact matched signal, including any requested range view.

Raises:

Type	Description
`ValueError`	If path contains matcher syntax or the signal does not exist.

`load_matched_unknown_masks(signal_path, clock_path, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

Batch-load X/Z mask waveforms for all signals matching signal_path.

Clock assignment follows load_matched_waveforms: a single matched clock is broadcast to all signals; otherwise the longest-prefix clock key is selected for each signal key.

Parameters:

Name	Type	Description	Default
`signal_path`	`str`	Signal query path. See class docstring.	required
`clock_path`	`str`	Clock signal query path. Must match at least one signal.	required
`include_x`	`bool`	If `True` (default), mark source `X`/`x` bits.	`True`
`include_z`	`bool`	If `True` (default), mark source `Z`/`z` bits.	`True`
`sample_on_posedge`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`root_scope`	`Scope \| None`	If provided, both signal_path and clock_path are searched within this scope instead of the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Waveform]:`	Same keys as `get_matched_signals` on signal_path.

`load_matched_waveforms(signal_path, clock_path, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

Batch-load all signals matching signal_path, each paired with its clock.

Internally calls get_matched_signals for both signal_path and clock_path, then dispatches load_waveform for every match.

Clock assignment rules:

Single clock — if clock_path matches exactly one signal, that clock is broadcast to all matched signals.
Multiple clocks — for each signal key, the clock whose key is the longest prefix of the signal key is selected. If no clock key is a prefix, raises ValueError.

Parameters:

Name	Type	Description	Default
`signal_path`	`str`	Signal query path. See class docstring.	required
`clock_path`	`str`	Clock signal query path. Must match at least one signal.	required
`xz_value`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`signed`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`sample_on_posedge`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`begin_time`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`end_time`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`begin_cycle`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`end_cycle`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`root_scope`	`Scope \| None`	If provided, both signal_path and clock_path are searched within this scope instead of the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Waveform]:`	Same keys as `get_matched_signals` on signal_path.

Raises:

Type	Description
`ValueError:`	If clock_path matches no signals, or if no clock key is a prefix of a signal key.

`load_unknown_mask(signal, clock, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

Load source X/Z presence as an unsigned bitmask waveform.

The returned Waveform is sampled on the same clock edges and supports the same time/cycle windowing as load_waveform, but its values are masks instead of substituted two-state signal values. A mask bit is 1 when the corresponding source bit is selected by include_x and/or include_z.

Parameters:

Name	Type	Description	Default
`signal`	`Signal \| str`	Full dotted signal path or `Signal` object.	required
`clock`	`Signal \| str`	Clock signal path or `Signal` object.	required
`include_x`	`bool`	If `True` (default), mark source `X`/`x` bits.	`True`
`include_z`	`bool`	If `True` (default), mark source `Z`/`z` bits.	`True`
`sample_on_posedge`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`

Returns:

Name	Type	Description
`Waveform`	`Waveform`	Unsigned mask waveform.

`load_waveform(signal, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

Load a single signal as a clock-synchronised Waveform.

The signal is sampled on every negedge of clock by default (i.e. the value is captured at each falling edge of the clock, which reflects the value that was stable during the preceding high phase). Set sample_on_posedge=True to sample on rising edges instead.

Parameters:

Name	Type	Description	Default
`signal`	`Signal \| str`	Full dotted path of the signal as a `Signal` object or a string. When a `Signal` is passed, `signal.full_name` is used as the path (which may include bit-range suffixes). When a string is passed, the value is used verbatim as the full hierarchical path, e.g. `"tb.dut.data[7:0]"` or `"tb.dut.data"`.	required
`clock`	`Signal \| str`	Clock signal as a `Signal` or full dotted path string, e.g. `"tb.clk"`.	required
`xz_value`	`int`	Integer substituted for `X` and `Z` values in the file. Defaults to `0`.	`0`
`signed`	`bool`	If `True`, the loaded values are interpreted as two's-complement signed integers.	`False`
`sample_on_posedge`	`bool`	If `True`, sample on rising clock edges; otherwise on falling edges (default).	`False`
`begin_time`	`int \| None`	Simulation time to start loading from (inclusive). `None` means start of simulation. Mutually exclusive with begin_cycle.	`None`
`end_time`	`int \| None`	Simulation time to stop loading at (exclusive). `None` means end of simulation. Mutually exclusive with end_cycle.	`None`
`begin_cycle`	`int \| None`	Absolute clock cycle number to start loading from (inclusive). `None` means start of simulation. Mutually exclusive with begin_time. The clock is always loaded from time 0 so cycle numbers are absolute and comparable across different waveforms.	`None`
`end_cycle`	`int \| None`	Absolute clock cycle number to stop loading at (exclusive). `None` means end of simulation. Mutually exclusive with end_time.	`None`

Returns:

Name	Type	Description
`Waveform`	`Waveform`	One sample per clock edge within the requested window. The `.clock` array contains absolute cycle numbers from the start of simulation. `waveform.signal.full_name` records the resolved signal path when source metadata is available.

Raises:

Type	Description
`ValueError:`	If both begin_time and begin_cycle (or both end_time and end_cycle) are provided simultaneously.

`FstReader(file)` ¶

Bases: Reader

Read FST waveform files via pylibfst.

Supports the same high-level APIs as VcdReader, including context-manager usage, hierarchy traversal, pattern matching, expression evaluation, and clock-synchronised load_waveform sampling.

`begin_time` `cached` `property` ¶

Return the first timestamp stored in the FST file.

`end_time` `cached` `property` ¶

Return the last timestamp stored in the FST file.

`top_scopes` `property` ¶

Return immutable top-level scopes in the FST hierarchy.

`close()` ¶

Close the underlying FST reader handle.

`eval(expr, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, mode='single', root_scope=None)` ¶

Evaluate a waveform expression containing physical signal paths.

Parameters:

Name	Type	Description	Default
`expr`	`str`	Expression string. Signal paths may be used as operands or as arguments to registered expression functions.	required
`clock`	`str`	Clock signal used for all waveform loads.	required
`xz_value`	`int`		`0`
`signed`	`int`		`0`
`sample_on_posedge`	`int`		`0`
`begin_time`	`int`		`0`
`end_time`	`int`		`0`
`begin_cycle`	`int`		`0`
`end_cycle`	`int \| None`	Forwarded to `load_matched_waveforms` for every path.	`None`
`mode`	`Literal['single', 'zip']`	`'single'` requires every path to match one signal. `'zip'` evaluates once per shared multi-match key and broadcasts singleton paths.	`'single'`
`root_scope`	`Scope \| None`	If provided, resolve paths within this scope.	`None`

Returns:

Type	Description
`Waveform or dict[tuple[Capture, ...], Waveform]`	The evaluated waveform, or one waveform per zip key.

`get_matched_scopes(path, root_scope=None)` ¶

Return all scopes whose paths match path, keyed by captures.

Similar to get_matched_signals but stops at the scope level — the last component of path must match a scope name, not a signal. Useful for enumerating module instances before loading their signals.

Parameters:

Name	Type	Description	Default
`path`	`str`	Scope query path using the same syntax as signal paths. The last component must match a scope (module) name, e.g. `"tb.dut.fifo_{0..3}"` or `r"tb./([a-z]+)_core/"`.	required
`root_scope`	`Scope \| None`	If provided, search only within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Scope]:`	Maps each capture key to the matched `Scope`. Ordinary exact-name matches are omitted from the key, so a query without binding matchers uses `()`.

Raises:

Type	Description
`ValueError:`	If two different scopes resolve to the same key, or if using module matchers on a backend without `definition` support (VCD/FST), or if the path contains a terminal signal bit-range suffix.

`get_matched_signals(path, root_scope=None)` ¶

Return all signals whose paths match path, keyed by captures.

Traverses the scope tree starting from root_scope (or the file's top-level scopes if root_scope is None) and applies the query path to each level. See the class docstring for query path syntax.

Parameters:

Name	Type	Description	Default
`path`	`str`	Signal query path, e.g. `"tb.dut.fifo_{0..3}.w_ptr[2:0]"` or `r"tb.dut./([a-z]+)_valid/"`.	required
`root_scope`	`Scope \| None`	If provided, search only within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Signal]:`	Maps each capture key to the matched `Signal` object (carrying name, width, range, signed). Ordinary exact-name matches are omitted from the key, so a query without binding matchers uses `()`.

Raises:

Type	Description
`ValueError:`	If two different signals resolve to the same key, or if using module matchers on a backend without `definition` support (VCD/FST).

`get_scope(path, root_scope=None)` ¶

Return the scope at an exact hierarchy path.

Parameters:

Name	Type	Description	Default
`path`	`str`	Exact dotted scope path. Matcher expressions and terminal signal ranges are not accepted.	required
`root_scope`	`Scope \| None`	If provided, resolve path within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`Scope`	The exact matched scope.

Raises:

Type	Description
`ValueError`	If path contains matcher syntax, contains a terminal range, or the scope does not exist.

`get_signal(path, root_scope=None)` ¶

Return the signal at an exact hierarchy path.

Parameters:

Name	Type	Description	Default
`path`	`str`	Exact dotted signal path, with an optional terminal bit selection or range. Matcher expressions are not accepted.	required
`root_scope`	`Scope \| None`	If provided, resolve path within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`Signal`	The exact matched signal, including any requested range view.

Raises:

Type	Description
`ValueError`	If path contains matcher syntax or the signal does not exist.

`load_matched_unknown_masks(signal_path, clock_path, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

Batch-load X/Z mask waveforms for all signals matching signal_path.

Clock assignment follows load_matched_waveforms: a single matched clock is broadcast to all signals; otherwise the longest-prefix clock key is selected for each signal key.

Parameters:

Name	Type	Description	Default
`signal_path`	`str`	Signal query path. See class docstring.	required
`clock_path`	`str`	Clock signal query path. Must match at least one signal.	required
`include_x`	`bool`	If `True` (default), mark source `X`/`x` bits.	`True`
`include_z`	`bool`	If `True` (default), mark source `Z`/`z` bits.	`True`
`sample_on_posedge`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`root_scope`	`Scope \| None`	If provided, both signal_path and clock_path are searched within this scope instead of the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Waveform]:`	Same keys as `get_matched_signals` on signal_path.

`load_matched_waveforms(signal_path, clock_path, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

Batch-load all signals matching signal_path, each paired with its clock.

Internally calls get_matched_signals for both signal_path and clock_path, then dispatches load_waveform for every match.

Clock assignment rules:

Single clock — if clock_path matches exactly one signal, that clock is broadcast to all matched signals.
Multiple clocks — for each signal key, the clock whose key is the longest prefix of the signal key is selected. If no clock key is a prefix, raises ValueError.

Parameters:

Name	Type	Description	Default
`signal_path`	`str`	Signal query path. See class docstring.	required
`clock_path`	`str`	Clock signal query path. Must match at least one signal.	required
`xz_value`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`signed`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`sample_on_posedge`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`begin_time`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`end_time`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`begin_cycle`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`end_cycle`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`root_scope`	`Scope \| None`	If provided, both signal_path and clock_path are searched within this scope instead of the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Waveform]:`	Same keys as `get_matched_signals` on signal_path.

Raises:

Type	Description
`ValueError:`	If clock_path matches no signals, or if no clock key is a prefix of a signal key.

`load_unknown_mask(signal, clock, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

Load source X/Z presence as an unsigned bitmask waveform.

The returned Waveform is sampled on the same clock edges and supports the same time/cycle windowing as load_waveform, but its values are masks instead of substituted two-state signal values. A mask bit is 1 when the corresponding source bit is selected by include_x and/or include_z.

Parameters:

Name	Type	Description	Default
`signal`	`Signal \| str`	Full dotted signal path or `Signal` object.	required
`clock`	`Signal \| str`	Clock signal path or `Signal` object.	required
`include_x`	`bool`	If `True` (default), mark source `X`/`x` bits.	`True`
`include_z`	`bool`	If `True` (default), mark source `Z`/`z` bits.	`True`
`sample_on_posedge`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`

Returns:

Name	Type	Description
`Waveform`	`Waveform`	Unsigned mask waveform.

`load_waveform(signal, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

Load a single signal as a clock-synchronised Waveform.

The signal is sampled on every negedge of clock by default (i.e. the value is captured at each falling edge of the clock, which reflects the value that was stable during the preceding high phase). Set sample_on_posedge=True to sample on rising edges instead.

Parameters:

Name	Type	Description	Default
`signal`	`Signal \| str`	Full dotted path of the signal as a `Signal` object or a string. When a `Signal` is passed, `signal.full_name` is used as the path (which may include bit-range suffixes). When a string is passed, the value is used verbatim as the full hierarchical path, e.g. `"tb.dut.data[7:0]"` or `"tb.dut.data"`.	required
`clock`	`Signal \| str`	Clock signal as a `Signal` or full dotted path string, e.g. `"tb.clk"`.	required
`xz_value`	`int`	Integer substituted for `X` and `Z` values in the file. Defaults to `0`.	`0`
`signed`	`bool`	If `True`, the loaded values are interpreted as two's-complement signed integers.	`False`
`sample_on_posedge`	`bool`	If `True`, sample on rising clock edges; otherwise on falling edges (default).	`False`
`begin_time`	`int \| None`	Simulation time to start loading from (inclusive). `None` means start of simulation. Mutually exclusive with begin_cycle.	`None`
`end_time`	`int \| None`	Simulation time to stop loading at (exclusive). `None` means end of simulation. Mutually exclusive with end_cycle.	`None`
`begin_cycle`	`int \| None`	Absolute clock cycle number to start loading from (inclusive). `None` means start of simulation. Mutually exclusive with begin_time. The clock is always loaded from time 0 so cycle numbers are absolute and comparable across different waveforms.	`None`
`end_cycle`	`int \| None`	Absolute clock cycle number to stop loading at (exclusive). `None` means end of simulation. Mutually exclusive with end_time.	`None`

Returns:

Name	Type	Description
`Waveform`	`Waveform`	One sample per clock edge within the requested window. The `.clock` array contains absolute cycle numbers from the start of simulation. `waveform.signal.full_name` records the resolved signal path when source metadata is available.

Raises:

Type	Description
`ValueError:`	If both begin_time and begin_cycle (or both end_time and end_cycle) are provided simultaneously.

`FsdbReader(file, *, quiet=True)` ¶

Bases: Reader

Read FSDB waveform files through the Verdi NPI runtime.

FsdbReader requires the Verdi runtime library (libNPI.so). Configure it with WAVEKIT_NPI_LIB, VERDI_HOME, or LD_LIBRARY_PATH before opening FSDB files.

quiet (default True) suppresses the NPI console banner via the -quiet initialization argument. Pass quiet=False to keep the banner. NPI initialization is process-global, so quiet only takes effect on the first FsdbReader created in the process; later readers cannot reliably re-enable the banner.

`begin_time` `property` ¶

Return the first timestamp stored in the FSDB file.

`end_time` `property` ¶

Return the last timestamp stored in the FSDB file.

`top_scopes` `cached` `property` ¶

Return immutable top-level scopes in the FSDB hierarchy.

`close()` ¶

Close the underlying FSDB/NPI reader handle.

`eval(expr, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, mode='single', root_scope=None)` ¶

Evaluate a waveform expression containing physical signal paths.

Parameters:

Name	Type	Description	Default
`expr`	`str`	Expression string. Signal paths may be used as operands or as arguments to registered expression functions.	required
`clock`	`str`	Clock signal used for all waveform loads.	required
`xz_value`	`int`		`0`
`signed`	`int`		`0`
`sample_on_posedge`	`int`		`0`
`begin_time`	`int`		`0`
`end_time`	`int`		`0`
`begin_cycle`	`int`		`0`
`end_cycle`	`int \| None`	Forwarded to `load_matched_waveforms` for every path.	`None`
`mode`	`Literal['single', 'zip']`	`'single'` requires every path to match one signal. `'zip'` evaluates once per shared multi-match key and broadcasts singleton paths.	`'single'`
`root_scope`	`Scope \| None`	If provided, resolve paths within this scope.	`None`

Returns:

Type	Description
`Waveform or dict[tuple[Capture, ...], Waveform]`	The evaluated waveform, or one waveform per zip key.

`get_matched_scopes(path, root_scope=None)` ¶

Return all scopes whose paths match path, keyed by captures.

Similar to get_matched_signals but stops at the scope level — the last component of path must match a scope name, not a signal. Useful for enumerating module instances before loading their signals.

Parameters:

Name	Type	Description	Default
`path`	`str`	Scope query path using the same syntax as signal paths. The last component must match a scope (module) name, e.g. `"tb.dut.fifo_{0..3}"` or `r"tb./([a-z]+)_core/"`.	required
`root_scope`	`Scope \| None`	If provided, search only within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Scope]:`	Maps each capture key to the matched `Scope`. Ordinary exact-name matches are omitted from the key, so a query without binding matchers uses `()`.

Raises:

Type	Description
`ValueError:`	If two different scopes resolve to the same key, or if using module matchers on a backend without `definition` support (VCD/FST), or if the path contains a terminal signal bit-range suffix.

`get_matched_signals(path, root_scope=None)` ¶

Return all signals whose paths match path, keyed by captures.

Traverses the scope tree starting from root_scope (or the file's top-level scopes if root_scope is None) and applies the query path to each level. See the class docstring for query path syntax.

Parameters:

Name	Type	Description	Default
`path`	`str`	Signal query path, e.g. `"tb.dut.fifo_{0..3}.w_ptr[2:0]"` or `r"tb.dut./([a-z]+)_valid/"`.	required
`root_scope`	`Scope \| None`	If provided, search only within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Signal]:`	Maps each capture key to the matched `Signal` object (carrying name, width, range, signed). Ordinary exact-name matches are omitted from the key, so a query without binding matchers uses `()`.

Raises:

Type	Description
`ValueError:`	If two different signals resolve to the same key, or if using module matchers on a backend without `definition` support (VCD/FST).

`get_scope(path, root_scope=None)` ¶

Return the scope at an exact hierarchy path.

Parameters:

Name	Type	Description	Default
`path`	`str`	Exact dotted scope path. Matcher expressions and terminal signal ranges are not accepted.	required
`root_scope`	`Scope \| None`	If provided, resolve path within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`Scope`	The exact matched scope.

Raises:

Type	Description
`ValueError`	If path contains matcher syntax, contains a terminal range, or the scope does not exist.

`get_signal(path, root_scope=None)` ¶

Return the signal at an exact hierarchy path.

Parameters:

Name	Type	Description	Default
`path`	`str`	Exact dotted signal path, with an optional terminal bit selection or range. Matcher expressions are not accepted.	required
`root_scope`	`Scope \| None`	If provided, resolve path within this scope instead of starting from the file's top-level scopes.	`None`

Returns:

Type	Description
`Signal`	The exact matched signal, including any requested range view.

Raises:

Type	Description
`ValueError`	If path contains matcher syntax or the signal does not exist.

`load_matched_unknown_masks(signal_path, clock_path, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

Batch-load X/Z mask waveforms for all signals matching signal_path.

Clock assignment follows load_matched_waveforms: a single matched clock is broadcast to all signals; otherwise the longest-prefix clock key is selected for each signal key.

Parameters:

Name	Type	Description	Default
`signal_path`	`str`	Signal query path. See class docstring.	required
`clock_path`	`str`	Clock signal query path. Must match at least one signal.	required
`include_x`	`bool`	If `True` (default), mark source `X`/`x` bits.	`True`
`include_z`	`bool`	If `True` (default), mark source `Z`/`z` bits.	`True`
`sample_on_posedge`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`root_scope`	`Scope \| None`	If provided, both signal_path and clock_path are searched within this scope instead of the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Waveform]:`	Same keys as `get_matched_signals` on signal_path.

`load_matched_waveforms(signal_path, clock_path, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

Batch-load all signals matching signal_path, each paired with its clock.

Internally calls get_matched_signals for both signal_path and clock_path, then dispatches load_waveform for every match.

Clock assignment rules:

Single clock — if clock_path matches exactly one signal, that clock is broadcast to all matched signals.
Multiple clocks — for each signal key, the clock whose key is the longest prefix of the signal key is selected. If no clock key is a prefix, raises ValueError.

Parameters:

Name	Type	Description	Default
`signal_path`	`str`	Signal query path. See class docstring.	required
`clock_path`	`str`	Clock signal query path. Must match at least one signal.	required
`xz_value`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`signed`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`sample_on_posedge`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`begin_time`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`end_time`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`begin_cycle`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`end_cycle`	`int`	Forwarded to `load_waveform` for every loaded signal.	`0`
`root_scope`	`Scope \| None`	If provided, both signal_path and clock_path are searched within this scope instead of the file's top-level scopes.	`None`

Returns:

Type	Description
`dict[tuple[Capture, ...], Waveform]:`	Same keys as `get_matched_signals` on signal_path.

Raises:

Type	Description
`ValueError:`	If clock_path matches no signals, or if no clock key is a prefix of a signal key.

`load_unknown_mask(signal, clock, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

Load source X/Z presence as an unsigned bitmask waveform.

The returned Waveform is sampled on the same clock edges and supports the same time/cycle windowing as load_waveform, but its values are masks instead of substituted two-state signal values. A mask bit is 1 when the corresponding source bit is selected by include_x and/or include_z.

Parameters:

Name	Type	Description	Default
`signal`	`Signal \| str`	Full dotted signal path or `Signal` object.	required
`clock`	`Signal \| str`	Clock signal path or `Signal` object.	required
`include_x`	`bool`	If `True` (default), mark source `X`/`x` bits.	`True`
`include_z`	`bool`	If `True` (default), mark source `Z`/`z` bits.	`True`
`sample_on_posedge`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_time`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`begin_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`
`end_cycle`	`bool`	Same sampling/windowing semantics as `load_waveform`.	`False`

Returns:

Name	Type	Description
`Waveform`	`Waveform`	Unsigned mask waveform.

`load_waveform(signal, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

Load a single signal as a clock-synchronised Waveform.

The signal is sampled on every negedge of clock by default (i.e. the value is captured at each falling edge of the clock, which reflects the value that was stable during the preceding high phase). Set sample_on_posedge=True to sample on rising edges instead.

Parameters:

Name	Type	Description	Default
`signal`	`Signal \| str`	Full dotted path of the signal as a `Signal` object or a string. When a `Signal` is passed, `signal.full_name` is used as the path (which may include bit-range suffixes). When a string is passed, the value is used verbatim as the full hierarchical path, e.g. `"tb.dut.data[7:0]"` or `"tb.dut.data"`.	required
`clock`	`Signal \| str`	Clock signal as a `Signal` or full dotted path string, e.g. `"tb.clk"`.	required
`xz_value`	`int`	Integer substituted for `X` and `Z` values in the file. Defaults to `0`.	`0`
`signed`	`bool`	If `True`, the loaded values are interpreted as two's-complement signed integers.	`False`
`sample_on_posedge`	`bool`	If `True`, sample on rising clock edges; otherwise on falling edges (default).	`False`
`begin_time`	`int \| None`	Simulation time to start loading from (inclusive). `None` means start of simulation. Mutually exclusive with begin_cycle.	`None`
`end_time`	`int \| None`	Simulation time to stop loading at (exclusive). `None` means end of simulation. Mutually exclusive with end_cycle.	`None`
`begin_cycle`	`int \| None`	Absolute clock cycle number to start loading from (inclusive). `None` means start of simulation. Mutually exclusive with begin_time. The clock is always loaded from time 0 so cycle numbers are absolute and comparable across different waveforms.	`None`
`end_cycle`	`int \| None`	Absolute clock cycle number to stop loading at (exclusive). `None` means end of simulation. Mutually exclusive with end_time.	`None`

Returns:

Name	Type	Description
`Waveform`	`Waveform`	One sample per clock edge within the requested window. The `.clock` array contains absolute cycle numbers from the start of simulation. `waveform.signal.full_name` records the resolved signal path when source metadata is available.

Raises:

Type	Description
`ValueError:`	If both begin_time and begin_cycle (or both end_time and end_cycle) are provided simultaneously.

`has_fsdb_support()` ¶

Check whether the Verdi FSDB runtime library is installed.

模式匹配¶

模式匹配 API 用于描述跨多个时钟周期的信号关系，并执行事务级分析。

`Pattern()` ¶

Declarative temporal pattern builder over waveform signals.

Build steps with Pattern().wait(...).capture(...) and execute with module-level match. Pattern stores only the declarative step AST; programmable checking/extraction uses match(body) or collect(body).

The first blocking step selects candidate start cycles. Later blocking steps wait within a matched transaction.

Declarative callbacks receive (index, captures). index is the current sample index into waveform arrays, not a cycle number and not rebased by start_cycle. captures is the current match's capture dict.

`branch(cond, true_body=None, false_body=None)` ¶

Run one of two epsilon bodies based on cond at the current cycle.

Parameters:

Name	Type	Description	Default
`cond`	`Waveform \| Callable[[int, dict[str, Any]], bool] \| bool`	`Waveform`, `bool`, or `callable(index, captures) -> bool`. Callback `index` is the current waveform-array sample index, not a cycle number. `captures` is the current match's capture dict.	required
`true_body`	`Pattern \| None`	Optional body run when `cond` is true.	`None`
`false_body`	`Pattern \| None`	Optional body run when `cond` is false.	`None`

Returns:

Type	Description
`Pattern`	This pattern, for chaining.

`capture(name, signal, *, mode='last')` ¶

Record a value into captures[name] at the current cycle.

Parameters:

Name	Type	Description	Default
`name`	`str`	Capture key.	required
`signal`	`Waveform \| Callable[[int, dict[str, Any]], Any]`	`Waveform` or `callable(index, captures) -> Any`. Waveforms are read as `waveform.value[index]` at the current cycle. Callback `index` is the current waveform-array sample index, not a cycle number. `captures` is the current match's capture dict.	required
`mode`	`CaptureMode`	`'last'` overwrites existing values, `'first'` keeps the first value, and `'list'` appends each captured value to a list.	`'last'`

Returns:

Type	Description
`Pattern`	This pattern, for chaining.

`consume(cond, channel=None, *, require=None, require_message=None)` ¶

Wait for cond and exclusively claim (channel, cycle).

Parameters:

Name	Type	Description	Default
`cond`	`Waveform \| Callable[[int, dict[str, Any]], bool] \| bool`	`Waveform`, `bool`, or `callable(index, captures) -> bool`. When true, this step tries to consume the current event cycle. Callback `index` is the current waveform-array sample index, not a cycle number. `captures` is the current match's capture dict.	required
`channel`	`Channel \| Hashable \| Callable[[int, dict[str, Any]], Channel \| Hashable] \| None`	`Channel`, hashable key, or `callable(index, captures) -> Channel \| Hashable`. If omitted, this consume step receives a private channel created when the step is declared. Channel callbacks use the same `(index, captures)` arguments. The resolved `(channel, cycle)` can be claimed by at most one match; earlier start cycles win. This does not reserve the channel while `cond` is false.	`None`
`require`	`Waveform \| Callable[[int, dict[str, Any]], bool] \| bool \| None`	Optional condition checked while `cond` is false, or while `cond` is true but the current `(channel, cycle)` was already claimed. It is not checked on the successful consume cycle.	`None`
`require_message`	`str \| Callable[[int, dict[str, Any]], str] \| None`	Optional message for `MatchStatus.RequireViolated`. A callable uses the same `(index, captures)` arguments and is evaluated only on failure.	`None`

Returns:

Type	Description
`Pattern`	This pattern, for chaining.

`delay(n, *, require=None, require_message=None)` ¶

Wait exactly n cycles; delay(0) is an epsilon no-op.

Parameters:

Name	Type	Description	Default
`n`	`int \| Callable[[int, dict[str, Any]], int]`	Non-negative `int` or `callable(index, captures) -> int`. Callback `index` is the current waveform-array sample index, not a cycle number. `captures` is the current match's capture dict.	required
`require`	`Waveform \| Callable[[int, dict[str, Any]], bool] \| bool \| None`	Optional condition checked before each cycle advance during the delay. Failure records `MatchStatus.RequireViolated(require_message)`.	`None`
`require_message`	`str \| Callable[[int, dict[str, Any]], str] \| None`	Optional message for `MatchStatus.RequireViolated`. A callable uses the same `(index, captures)` arguments and is evaluated only on failure.	`None`

Returns:

Type	Description
`Pattern`	This pattern, for chaining.

`loop(body, *, until=None, when=None)` ¶

Run body as a do-while (until) or while (when) loop.

Parameters:

Name	Type	Description	Default
`body`	`Pattern`	Nested declarative pattern body.	required
`until`	`Waveform \| Callable[[int, dict[str, Any]], bool] \| bool \| None`	Optional do-while exit condition. The body runs first; the loop exits when `until` becomes true. Callable conditions use `(index, captures)`.	`None`
`when`	`Waveform \| Callable[[int, dict[str, Any]], bool] \| bool \| None`	Optional while condition. The condition is checked before each iteration; the loop exits when `when` becomes false. Callable conditions use `(index, captures)`.	`None`

Returns:

Type	Description
`Pattern`	This pattern, for chaining.

Notes

Exactly one of until and when is required. Conditions use the same Waveform / bool / callable forms as wait.

`repeat(body, n)` ¶

Run body exactly n times.

Parameters:

Name	Type	Description	Default
`body`	`Pattern`	Nested declarative pattern body.	required
`n`	`int \| Callable[[int, dict[str, Any]], int]`	Non-negative `int` or `callable(index, captures) -> int`. Callback `index` is the current waveform-array sample index, not a cycle number. `captures` is the current match's capture dict.	required

Returns:

Type	Description
`Pattern`	This pattern, for chaining.

`require(cond, *, message=None)` ¶

Assert cond at the current cycle, else record RequireViolated.

Parameters:

Name	Type	Description	Default
`cond`	`Waveform \| Callable[[int, dict[str, Any]], bool] \| bool`	`Waveform`, `bool`, or `callable(index, captures) -> bool`. Callback `index` is the current waveform-array sample index, not a cycle number. `captures` is the current match's capture dict.	required
`message`	`str \| Callable[[int, dict[str, Any]], str] \| None`	Optional failure message. A callable uses the same `(index, captures)` arguments and is evaluated only on failure.	`None`

Returns:

Type	Description
`Pattern`	This pattern, for chaining.

`wait(cond, *, require=None, require_message=None)` ¶

Observe cycles until cond becomes true, without consuming the event.

Parameters:

Name	Type	Description	Default
`cond`	`Waveform \| Callable[[int, dict[str, Any]], bool] \| bool`	`Waveform`, `bool`, or `callable(index, captures) -> bool`. When true at the current cycle, the step completes at that cycle. Callback `index` is the current waveform-array sample index, not a cycle number. `captures` is the current match's capture dict.	required
`require`	`Waveform \| Callable[[int, dict[str, Any]], bool] \| bool \| None`	Optional condition checked only while `cond` is false. It is not checked on the cycle where `cond` becomes true. Failure records `MatchStatus.RequireViolated(require_message)`.	`None`
`require_message`	`str \| Callable[[int, dict[str, Any]], str] \| None`	Optional message for `MatchStatus.RequireViolated`. A callable uses the same `(index, captures)` arguments and is evaluated only on failure.	`None`

Returns:

Type	Description
`Pattern`	This pattern, for chaining.

`match(body, *, axis=None, timeout=None, timeout_message=None, start_cycle=None, end_cycle=None)` ¶

Run a declarative pattern or programmable check body.

Parameters:

Name	Type	Description	Default
`body`	`Pattern \| Callable[[PatternContext], Any]`	Declarative `Pattern` or normal callable `body(ctx)`. Check bodies run once per scanned start cycle and must return `ctx.OK` to emit an OK row, or `None` to skip that start. Other non-`None` values are errors.	required
`axis`	`Waveform \| None`	Optional waveform that defines the scan axis, cycle numbers, and result timestamps. Declarative patterns usually infer it from observed waveforms. Pass `axis` when the body may not observe a waveform before blocking, or when using `start_cycle` / `end_cycle` without an inferable waveform.	`None`
`timeout`	`int \| None`	Optional positive integer per-start maximum duration in cycles. Exceeding it records `MatchStatus.Timeout(timeout_message)`.	`None`
`timeout_message`	`str \| None`	Optional human-readable message stored in timeout statuses.	`None`
`start_cycle`	`int \| None`	Optional absolute cycle scan window. `start_cycle` is inclusive and `end_cycle` is exclusive.	`None`
`end_cycle`	`int \| None`	Optional absolute cycle scan window. `start_cycle` is inclusive and `end_cycle` is exclusive.	`None`

Returns:

Type	Description
`MatchRecords`	Ordered batch of match records and captured columns.

Raises:

Type	Description
`PatternError`	If the body is invalid, waveform axes are incompatible, or execution cannot infer a scan axis.

`collect(body, *, axis=None, timeout=None, timeout_message=None, start_cycle=None, end_cycle=None)` ¶

Run a programmable extraction body and collect returned items.

Parameters:

Name	Type	Description	Default
`body`	`Callable[[PatternContext], Any]`	Normal callable `body(ctx)`. It runs once per scanned start cycle; each non-`None` return value is appended to the output list. Declarative `Pattern` objects are intentionally unsupported.	required
`axis`	`Waveform \| None`	Optional waveform that defines the scan axis. Pass it when the body may not observe a waveform before blocking, or when using a scan window.	`None`
`timeout`	`int \| None`	Optional positive integer per-start maximum duration in cycles. Timeout raises `PatternError` instead of returning a status row.	`None`
`timeout_message`	`str \| None`	Optional human-readable timeout message.	`None`
`start_cycle`	`int \| None`	Optional absolute cycle scan window. `start_cycle` is inclusive and `end_cycle` is exclusive.	`None`
`end_cycle`	`int \| None`	Optional absolute cycle scan window. `start_cycle` is inclusive and `end_cycle` is exclusive.	`None`

Returns:

Type	Description
`list[Any]`	Non-`None` values returned by the extraction body.

Raises:

Type	Description
`PatternError`	If `body` is not a normal callable, if a timeout/require failure occurs, or if execution cannot infer a scan axis.

`MatchRecords(start, end, duration, status, captures)` ¶

Bases: Sequence[MatchRecord]

Columnar batch of pattern match records.

start and end are point waveforms: .value stores waveform-array sample indices, .clock stores absolute cycle numbers, and .time stores simulation timestamps. duration.value is end.value - start.value + 1.

`failed` `property` ¶

Return a boolean result-row mask for failed matches.

Returns:

Type	Description
`Waveform`	One-bit waveform aligned to result rows. `value` is true where `status` is not `MatchStatus.OK`.

`ok` `property` ¶

Return a boolean result-row mask for successful matches.

Returns:

Type	Description
`Waveform`	One-bit waveform aligned to result rows. `value` is true where `status` is `MatchStatus.OK`.

`filter_failed()` ¶

Return records whose status is not MatchStatus.OK.

Returns:

Type	Description
`MatchRecords`	A row-masked batch with all fields and captures filtered together.

`filter_ok()` ¶

Return records whose status is MatchStatus.OK.

Returns:

Type	Description
`MatchRecords`	A row-masked batch with all fields and captures filtered together.

`filter_status(status)` ¶

Return records matching a status class.

Parameters:

Name	Type	Description	Default
`status`	`type[MatchStatusValue]`	Status class such as `MatchStatus.Timeout`.	required

Returns:

Type	Description
`MatchRecords`	A row-masked batch with all fields and captures filtered together.

`MatchRecord(start, end, status, captures)` `dataclass` ¶

One pattern match record.

Attributes:

Name	Type	Description
`start, end`		Inclusive match boundary points. `MatchPoint.index` is the waveform-array sample index, `MatchPoint.cycle` is the absolute clock cycle, and `MatchPoint.time` is the simulation timestamp.
`status`	`MatchStatusValue`	Terminal status object: `MatchStatus.OK()`, `MatchStatus.Timeout(...)`, or `MatchStatus.RequireViolated(...)`.
`captures`	`dict[str, Any]`	Per-record captured Python values.

`duration` `property` ¶

Return the inclusive duration in sampled cycles.

Returns:

Type	Description
`int`	`end.index - start.index + 1`.

`MatchPoint(index, cycle, time)` `dataclass` ¶

One match boundary point.

index is the waveform-array sample index; cycle and time are the corresponding absolute cycle number and simulation timestamp.

`MatchStatus` ¶

Terminal status namespace for pattern match records.

`OK()` `dataclass` ¶

Bases: MatchStatusValue

Successful completion of a pattern candidate.

`RequireViolated(message=None)` `dataclass` ¶

Bases: MatchStatusValue

A non-blocking require check failed.

`Timeout(message=None)` `dataclass` ¶

Bases: MatchStatusValue

A pattern candidate did not complete within its allowed duration.

`Channel` ¶

Identity object for explicit consume ownership.

A Channel represents a logical event stream from which at most one pattern instance may consume per cycle. Plain wait steps are observational and do not consume channels.

`PatternError` ¶

Bases: Exception

Raised for pattern definition or runtime errors.

信号层次结构和查询¶

这些对象表示波形文件中的层次结构、信号、范围以及查询结果中的 capture。

`Node(base_name, parent)` `dataclass` ¶

Bases: ABC

An immutable node in a waveform-file hierarchy.

`children` `abstractmethod` `property` ¶

Return this node's direct children.

`full_name` `cached` `property` ¶

Return this node's fully-qualified real hierarchy name.

`is_range_selectable` `property` ¶

Return whether this node supports a trailing bit-range selection.

`name` `property` ¶

Return this node's local name, including a signal range when present.

`get_matched_nodes(path)` ¶

Return matching descendant nodes keyed by binding captures.

`get_matched_scopes(path)` ¶

Return matching descendant scopes keyed by binding captures.

`get_matched_signals(path)` ¶

Return matching descendant signals keyed by binding captures.

`Scope(base_name, parent)` `dataclass` ¶

Bases: Node

A real hierarchy scope.

`children` `abstractmethod` `property` ¶

Return this node's direct children.

`full_name` `cached` `property` ¶

Return this node's fully-qualified real hierarchy name.

`is_range_selectable` `property` ¶

Return whether this node supports a trailing bit-range selection.

`name` `property` ¶

Return this node's local name, including a signal range when present.

`get_matched_nodes(path)` ¶

Return matching descendant nodes keyed by binding captures.

`get_matched_scopes(path)` ¶

Return matching descendant scopes keyed by binding captures.

`get_matched_signals(path)` ¶

Return matching descendant signals keyed by binding captures.

`Signal(base_name, parent, range, composite_type=None, native_range=None)` `dataclass` ¶

Bases: Node

An immutable signal view, optionally narrowed by a selection range.

`children` `abstractmethod` `property` ¶

Return this node's direct children.

`full_name` `cached` `property` ¶

Return this node's fully-qualified real hierarchy name.

`is_leaf` `property` ¶

Return whether this signal has no composite children.

`is_range_selectable` `property` ¶

Return whether this signal supports a trailing bit-range selection.

`name` `property` ¶

Return this node's local name, including a signal range when present.

`native_width` `cached` `property` ¶

Return the width of the complete signal before range selection.

`width` `cached` `property` ¶

Return the width of the current selected signal view.

`get_matched_nodes(path)` ¶

Return matching descendant nodes keyed by binding captures.

`get_matched_scopes(path)` ¶

Return matching descendant scopes keyed by binding captures.

`get_matched_signals(path)` ¶

Return matching descendant signals keyed by binding captures.

`with_range(selected_range)` ¶

Return a view with selected_range, or restore the native range for None.

`Range(start, end)` `dataclass` ¶

An HDL index range whose direction is preserved as start:end.

`SignalCompositeType` ¶

Bases: Enum

Composite signal type as reported by a waveform backend.

`Capture(anchor_node=None, node=None, definition=None)` `dataclass` ¶

Base class for typed bindings returned in a capture tuple.

Matcher-specific code may create a partial Capture with no node context. Matcher.match() completes it before it can escape the matcher layer.

`finalize(*, anchor_node, node, definition=None)` ¶

Complete a partial Capture with its matched node context.

`with_anchor_node(anchor_node)` ¶

Return a complete Capture rebased to anchor_node.

`ExactCapture(anchor_node=None, node=None, definition=None)` `dataclass` ¶

Bases: Capture

Exact-name binding; only module-definition matches are public.

`finalize(*, anchor_node, node, definition=None)` ¶

Complete a partial Capture with its matched node context.

`with_anchor_node(anchor_node)` ¶

Return a complete Capture rebased to anchor_node.

`BraceCapture(anchor_node=None, node=None, definition=None, groups=())` `dataclass` ¶

Bases: Capture

Binding produced by brace expansion; groups stores brace values.

`finalize(*, anchor_node, node, definition=None)` ¶

Complete a partial Capture with its matched node context.

`with_anchor_node(anchor_node)` ¶

Return a complete Capture rebased to anchor_node.

`RegexCapture(anchor_node=None, node=None, definition=None, groups=())` `dataclass` ¶

Bases: Capture

Binding produced by regex matching; groups stores regex groups.

`finalize(*, anchor_node, node, definition=None)` ¶

Complete a partial Capture with its matched node context.

`with_anchor_node(anchor_node)` ¶

Return a complete Capture rebased to anchor_node.

`WildcardCapture(anchor_node=None, node=None, definition=None)` `dataclass` ¶

Bases: Capture

Binding produced by * or ** wildcard matching.

`finalize(*, anchor_node, node, definition=None)` ¶

Complete a partial Capture with its matched node context.

`with_anchor_node(anchor_node)` ¶

Return a complete Capture rebased to anchor_node.

API 参考¶

Waveform¶

Waveform(value, clock, time, width=None, signed=False, signal=None) ¶

data property ¶

ahead(n=1, pad='repeat', pad_value=None) ¶

any_edge() ¶

as_signed() ¶

as_unsigned() ¶

back(n=1, pad='repeat', pad_value=None) ¶

bit_count() ¶

changed() ¶

compress() ¶

concatenate(waves) staticmethod ¶

copy() ¶

cycle_slice(begin_cycle=None, end_cycle=None, include_end=False) ¶

downsample(chunk_size, func=np.mean) ¶

falling_edge() ¶

filter(condition) ¶

map(func, width=None, signed=None) ¶

mask(mask) ¶

merge(waves, func, width, signed) staticmethod ¶

relative(offset, pad='repeat', pad_value=None) ¶

rising_edge() ¶

slice(begin_idx, end_idx, include_end=False) ¶

split_bits(bit_group_size, padding=False) ¶

take(indices) ¶

time_slice(begin_time=None, end_time=None, include_end=False) ¶

unique_consecutive() ¶

vectorized_filter(func) ¶

vectorized_map(func, width=None, signed=None) ¶

Reader¶

VcdReader(file) ¶

begin_time property ¶

end_time property ¶

top_scopes cached property ¶

close() ¶

eval(expr, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, mode='single', root_scope=None) ¶

get_matched_scopes(path, root_scope=None) ¶

get_matched_signals(path, root_scope=None) ¶

get_scope(path, root_scope=None) ¶

get_signal(path, root_scope=None) ¶

load_matched_unknown_masks(signal_path, clock_path, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None) ¶

load_matched_waveforms(signal_path, clock_path, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None) ¶

load_unknown_mask(signal, clock, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None) ¶

load_waveform(signal, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None) ¶

FstReader(file) ¶

begin_time cached property ¶

end_time cached property ¶

top_scopes property ¶

close() ¶

eval(expr, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, mode='single', root_scope=None) ¶

get_matched_scopes(path, root_scope=None) ¶

get_matched_signals(path, root_scope=None) ¶

get_scope(path, root_scope=None) ¶

get_signal(path, root_scope=None) ¶

load_matched_unknown_masks(signal_path, clock_path, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None) ¶

load_matched_waveforms(signal_path, clock_path, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None) ¶

load_unknown_mask(signal, clock, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None) ¶

load_waveform(signal, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None) ¶

FsdbReader(file, *, quiet=True) ¶

begin_time property ¶

end_time property ¶

top_scopes cached property ¶

close() ¶

eval(expr, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, mode='single', root_scope=None) ¶

get_matched_scopes(path, root_scope=None) ¶

get_matched_signals(path, root_scope=None) ¶

get_scope(path, root_scope=None) ¶

get_signal(path, root_scope=None) ¶

load_matched_unknown_masks(signal_path, clock_path, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None) ¶

load_matched_waveforms(signal_path, clock_path, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None) ¶

load_unknown_mask(signal, clock, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None) ¶

load_waveform(signal, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None) ¶

has_fsdb_support() ¶

模式匹配¶

Pattern() ¶

branch(cond, true_body=None, false_body=None) ¶

capture(name, signal, *, mode='last') ¶

consume(cond, channel=None, *, require=None, require_message=None) ¶

delay(n, *, require=None, require_message=None) ¶

`Waveform(value, clock, time, width=None, signed=False, signal=None)` ¶

`data` `property` ¶

`ahead(n=1, pad='repeat', pad_value=None)` ¶

`any_edge()` ¶

`as_signed()` ¶

`as_unsigned()` ¶

`back(n=1, pad='repeat', pad_value=None)` ¶

`bit_count()` ¶

`changed()` ¶

`compress()` ¶

`concatenate(waves)` `staticmethod` ¶

`copy()` ¶

`cycle_slice(begin_cycle=None, end_cycle=None, include_end=False)` ¶

`downsample(chunk_size, func=np.mean)` ¶

`falling_edge()` ¶

`filter(condition)` ¶

`map(func, width=None, signed=None)` ¶

`mask(mask)` ¶

`merge(waves, func, width, signed)` `staticmethod` ¶

`relative(offset, pad='repeat', pad_value=None)` ¶

`rising_edge()` ¶

`slice(begin_idx, end_idx, include_end=False)` ¶

`split_bits(bit_group_size, padding=False)` ¶

`take(indices)` ¶

`time_slice(begin_time=None, end_time=None, include_end=False)` ¶

`unique_consecutive()` ¶

`vectorized_filter(func)` ¶

`vectorized_map(func, width=None, signed=None)` ¶

`VcdReader(file)` ¶

`begin_time` `property` ¶

`end_time` `property` ¶

`top_scopes` `cached` `property` ¶

`close()` ¶

`eval(expr, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, mode='single', root_scope=None)` ¶

`get_matched_scopes(path, root_scope=None)` ¶

`get_matched_signals(path, root_scope=None)` ¶

`get_scope(path, root_scope=None)` ¶

`get_signal(path, root_scope=None)` ¶

`load_matched_unknown_masks(signal_path, clock_path, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

`load_matched_waveforms(signal_path, clock_path, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

`load_unknown_mask(signal, clock, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

`load_waveform(signal, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

`FstReader(file)` ¶

`begin_time` `cached` `property` ¶

`end_time` `cached` `property` ¶

`top_scopes` `property` ¶

`close()` ¶

`eval(expr, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, mode='single', root_scope=None)` ¶

`get_matched_scopes(path, root_scope=None)` ¶

`get_matched_signals(path, root_scope=None)` ¶

`get_scope(path, root_scope=None)` ¶

`get_signal(path, root_scope=None)` ¶

`load_matched_unknown_masks(signal_path, clock_path, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

`load_matched_waveforms(signal_path, clock_path, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

`load_unknown_mask(signal, clock, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

`load_waveform(signal, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

`FsdbReader(file, *, quiet=True)` ¶

`begin_time` `property` ¶

`end_time` `property` ¶

`top_scopes` `cached` `property` ¶

`close()` ¶

`eval(expr, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, mode='single', root_scope=None)` ¶

`get_matched_scopes(path, root_scope=None)` ¶

`get_matched_signals(path, root_scope=None)` ¶

`get_scope(path, root_scope=None)` ¶

`get_signal(path, root_scope=None)` ¶

`load_matched_unknown_masks(signal_path, clock_path, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

`load_matched_waveforms(signal_path, clock_path, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None, root_scope=None)` ¶

`load_unknown_mask(signal, clock, include_x=True, include_z=True, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

`load_waveform(signal, clock, xz_value=0, signed=False, sample_on_posedge=False, begin_time=None, end_time=None, begin_cycle=None, end_cycle=None)` ¶

`has_fsdb_support()` ¶

`Pattern()` ¶

`branch(cond, true_body=None, false_body=None)` ¶

`capture(name, signal, *, mode='last')` ¶

`consume(cond, channel=None, *, require=None, require_message=None)` ¶

`delay(n, *, require=None, require_message=None)` ¶

`loop(body, *, until=None, when=None)` ¶

`repeat(body, n)` ¶

`require(cond, *, message=None)` ¶

`wait(cond, *, require=None, require_message=None)` ¶