(self, start, end, state_data)
| 157 | self.kwargs = kwargs |
| 158 | |
| 159 | def __call__(self, start, end, state_data): |
| 160 | dtypes = self.dtypes |
| 161 | columns = self.columns |
| 162 | freq = self.freq |
| 163 | kwargs = self.kwargs |
| 164 | state = np.random.RandomState(state_data) |
| 165 | index = pd.date_range(start=start, end=end, freq=freq, name="timestamp") |
| 166 | data = {} |
| 167 | for k, dt in dtypes.items(): |
| 168 | kws = { |
| 169 | kk.rsplit("_", 1)[1]: v |
| 170 | for kk, v in kwargs.items() |
| 171 | if kk.rsplit("_", 1)[0] == k |
| 172 | } |
| 173 | # Note: we compute data for all dtypes in order, not just those in the output |
| 174 | # columns. This ensures the same output given the same state_data, regardless |
| 175 | # of whether there is any column projection. |
| 176 | # cf. https://github.com/dask/dask/pull/9538#issuecomment-1267461887 |
| 177 | result = make[dt](len(index), state, **kws) |
| 178 | if k in columns: |
| 179 | data[k] = result |
| 180 | df = pd.DataFrame(data, index=index, columns=columns) |
| 181 | if df.index[-1] == end: |
| 182 | df = df.iloc[:-1] |
| 183 | return df |
| 184 | |
| 185 | |
| 186 | def timeseries( |
nothing calls this directly
no test coverage detected