| 1 |
# Copyright (c) NumFOCUS. |
| 2 |
# Distributed under the terms of the Modified BSD License. |
| 3 |
|
| 4 |
import datetime |
| 5 |
import decimal |
| 6 |
import warnings |
| 7 |
from collections.abc import Iterator |
| 8 |
from copy import deepcopy |
| 9 |
from math import floor |
| 10 |
|
| 11 |
import numpy as np |
| 12 |
import pandas as pd |
| 13 |
from bqplot.traits import array_from_json, array_to_json |
| 14 |
from ipywidgets import CallbackDispatcher, DOMWidget, widget_serialization |
| 15 |
from traitlets import ( |
| 16 |
Bool, |
| 17 |
Dict, |
| 18 |
Enum, |
| 19 |
Instance, |
| 20 |
Int, |
| 21 |
List, |
| 22 |
Unicode, |
| 23 |
default, |
| 24 |
observe, |
| 25 |
validate, |
| 26 |
) |
| 27 |
|
| 28 |
from ._frontend import module_name, module_version |
| 29 |
from .cellrenderer import BarRenderer, CellRenderer, TextRenderer |
| 30 |
|
| 31 |
|
| 32 |
class SelectionIterator(Iterator): |
| 33 |
def __init__(self, selections): |
| 34 |
self._rect_index = 0 |
| 35 |
self._cell_index = 0 |
| 36 |
self._selections = selections |
| 37 |
|
| 38 |
def __next__(self): |
| 39 |
if self._rect_index >= len(self._selections): |
| 40 |
raise StopIteration |
| 41 |
|
| 42 |
rect = self._selections[self._rect_index] |
| 43 |
row_col = self._index_to_row_col(rect, self._cell_index) |
| 44 |
self._cell_index += 1 |
| 45 |
|
| 46 |
if row_col is None: |
| 47 |
self._rect_index += 1 |
| 48 |
self._cell_index = 0 |
| 49 |
return self.__next__() |
| 50 |
elif self._cell_in_previous_selected_rects(row_col): |
| 51 |
return self.__next__() |
| 52 |
else: |
| 53 |
return row_col |
| 54 |
|
| 55 |
@staticmethod |
| 56 |
def _index_to_row_col(rect, index): |
| 57 |
num_rows = rect["r2"] - rect["r1"] + 1 |
| 58 |
num_cols = rect["c2"] - rect["c1"] + 1 |
| 59 |
if index > (num_rows * num_cols - 1): |
| 60 |
return None |
| 61 |
|
| 62 |
return { |
| 63 |
"r": rect["r1"] + floor(index / num_cols), |
| 64 |
"c": rect["c1"] + index % num_cols, |
| 65 |
} |
| 66 |
|
| 67 |
def _cell_in_previous_selected_rects(self, cell): |
| 68 |
return any( |
| 69 |
self._cell_in_rect(cell, self._selections[i]) |
| 70 |
for i in range(0, self._rect_index) |
| 71 |
) |
| 72 |
|
| 73 |
@staticmethod |
| 74 |
def _cell_in_rect(cell, rect): |
| 75 |
return ( |
| 76 |
rect["r1"] <= cell["r"] <= rect["r2"] |
| 77 |
and rect["c1"] <= cell["c"] <= rect["c2"] |
| 78 |
) |
| 79 |
|
| 80 |
|
| 81 |
class SelectionHelper: |
| 82 |
"""A Helper Class for processing selections. Provides an iterator |
| 83 |
to traverse selected cells. |
| 84 |
""" |
| 85 |
|
| 86 |
def __init__(self, data, selections, selection_mode, **kwargs): |
| 87 |
super().__init__(**kwargs) |
| 88 |
self._data = data |
| 89 |
self._selections = selections |
| 90 |
self._selection_mode = selection_mode |
| 91 |
self._num_columns = -1 |
| 92 |
self._num_rows = -1 |
| 93 |
|
| 94 |
def __iter__(self): |
| 95 |
selections = [ |
| 96 |
self._transform_rect_for_selection_mode(rect) |
| 97 |
for rect in self._selections |
| 98 |
] |
| 99 |
return SelectionIterator(selections) |
| 100 |
|
| 101 |
def __len__(self): |
| 102 |
return sum(1 for _ in self) |
| 103 |
|
| 104 |
def all(self): |
| 105 |
""" |
| 106 |
Returns all selected cells as a list. Each cell is |
| 107 |
represented as a dictionary |
| 108 |
with keys 'r': row and 'c': column |
| 109 |
""" |
| 110 |
return list(self) |
| 111 |
|
| 112 |
def all_values(self): |
| 113 |
""" |
| 114 |
Returns values for all selected cells as a list. |
| 115 |
""" |
| 116 |
return [ |
| 117 |
DataGrid._get_cell_value_by_numerical_index( |
| 118 |
self._data, cell["c"], cell["r"] |
| 119 |
) |
| 120 |
for cell in self |
| 121 |
] |
| 122 |
|
| 123 |
def _transform_rect_for_selection_mode(self, rect): |
| 124 |
selection_mode = self._selection_mode |
| 125 |
if selection_mode == "row": |
| 126 |
return { |
| 127 |
"r1": rect["r1"], |
| 128 |
"c1": 0, |
| 129 |
"r2": rect["r2"], |
| 130 |
"c2": self._get_num_columns() - 1, |
| 131 |
} |
| 132 |
elif selection_mode == "column": |
| 133 |
return { |
| 134 |
"r1": 0, |
| 135 |
"c1": rect["c1"], |
| 136 |
"r2": self._get_num_rows() - 1, |
| 137 |
"c2": rect["c2"], |
| 138 |
} |
| 139 |
else: |
| 140 |
return rect |
| 141 |
|
| 142 |
def _get_num_columns(self): |
| 143 |
if self._num_columns != -1: |
| 144 |
return self._num_columns |
| 145 |
|
| 146 |
data = self._data |
| 147 |
primary_keys = ( |
| 148 |
[] |
| 149 |
if "primaryKey" not in data["schema"] |
| 150 |
else data["schema"]["primaryKey"] |
| 151 |
) |
| 152 |
col_headers = [ |
| 153 |
field["name"] |
| 154 |
for field in data["schema"]["fields"] |
| 155 |
if field["name"] not in primary_keys |
| 156 |
] |
| 157 |
self._num_columns = len(col_headers) |
| 158 |
return self._num_columns |
| 159 |
|
| 160 |
def _get_num_rows(self): |
| 161 |
if self._num_rows != -1: |
| 162 |
return self._num_rows |
| 163 |
|
| 164 |
data = self._data |
| 165 |
self._num_rows = 0 if "data" not in data else len(data["data"]) |
| 166 |
return self._num_rows |
| 167 |
|
| 168 |
|
| 169 |
def _data_to_json(x, _): |
| 170 |
if isinstance(x, dict): |
| 171 |
return {str(k): _data_to_json(v, _) for k, v in x.items()} |
| 172 |
if isinstance(x, np.ndarray): |
| 173 |
return _data_to_json(x.tolist(), _) |
| 174 |
if isinstance(x, (list, tuple)): |
| 175 |
return [_data_to_json(v, _) for v in x] |
| 176 |
if isinstance(x, int): |
| 177 |
return x |
| 178 |
if isinstance(x, float): |
| 179 |
if np.isnan(x): |
| 180 |
return "$NaN$" |
| 181 |
if np.isposinf(x): |
| 182 |
return "$Infinity$" |
| 183 |
if np.isneginf(x): |
| 184 |
return "$NegInfinity$" |
| 185 |
return x |
| 186 |
if isinstance(x, decimal.Decimal): |
| 187 |
return str(x) |
| 188 |
if isinstance(x, (datetime.datetime, datetime.date)): |
| 189 |
return x.isoformat() |
| 190 |
if x is pd.NaT: |
| 191 |
return "$NaT$" |
| 192 |
if pd.isna(x): |
| 193 |
return "$NaN$" |
| 194 |
return str(x) |
| 195 |
|
| 196 |
|
| 197 |
def _data_serialization_impl(data, _): |
| 198 |
if not data: |
| 199 |
return {} |
| 200 |
|
| 201 |
serialized_data = {} |
| 202 |
for column, value in data["data"].items(): |
| 203 |
arr = value.to_numpy() |
| 204 |
if arr.size == 0: |
| 205 |
serialized_data[str(column)] = { |
| 206 |
"value": [], |
| 207 |
"dtype": str(arr.dtype), |
| 208 |
"shape": arr.shape, |
| 209 |
"type": None, |
| 210 |
} |
| 211 |
continue |
| 212 |
try: |
| 213 |
serialized_data[str(column)] = array_to_json(arr) |
| 214 |
except ValueError: |
| 215 |
# Column is most likely heterogeneous, sending the column raw |
| 216 |
serialized_data[str(column)] = { |
| 217 |
"value": _data_to_json(arr, _), |
| 218 |
"type": "raw", |
| 219 |
} |
| 220 |
|
| 221 |
return { |
| 222 |
"data": serialized_data, |
| 223 |
"schema": data["schema"], |
| 224 |
"fields": _data_to_json(data["fields"], _), |
| 225 |
} |
| 226 |
|
| 227 |
|
| 228 |
def _data_deserialization_impl(data, _): # noqa: U101 |
| 229 |
if not data: |
| 230 |
return {} |
| 231 |
|
| 232 |
deserialized_data = {} |
| 233 |
for column, value in data["data"].items(): |
| 234 |
deserialized_data[column] = array_from_json(value.to_numpy()) |
| 235 |
|
| 236 |
return { |
| 237 |
"data": deserialized_data, |
| 238 |
"schema": data["schema"], |
| 239 |
"fields": data["fields"], |
| 240 |
} |
| 241 |
|
| 242 |
|
| 243 |
_data_serialization = { |
| 244 |
"from_json": _data_deserialization_impl, |
| 245 |
"to_json": _data_serialization_impl, |
| 246 |
} |
| 247 |
|
| 248 |
|
| 249 |
def _widgets_dict_to_json(x, obj): |
| 250 |
return { |
| 251 |
str(k): widget_serialization["to_json"](v, obj) for k, v in x.items() |
| 252 |
} |
| 253 |
|
| 254 |
|
| 255 |
_widgets_dict_serialization = { |
| 256 |
"from_json": widget_serialization["from_json"], |
| 257 |
"to_json": _widgets_dict_to_json, |
| 258 |
} |
| 259 |
|
| 260 |
|
| 261 |
class DataGrid(DOMWidget): |
| 262 |
"""A Grid Widget with filter, sort and selection capabilities. |
| 263 |
|
| 264 |
Attributes |
| 265 |
---------- |
| 266 |
base_row_size : int (default: 20) |
| 267 |
Default row height |
| 268 |
base_column_size : int (default: 64) |
| 269 |
Default column width |
| 270 |
base_row_header_size : int (default: 64) |
| 271 |
Default row header width |
| 272 |
base_column_header_size : int (default: 20) |
| 273 |
Default column header height |
| 274 |
header_visibility : {'all', 'row', 'column', 'none'} (default: 'all') |
| 275 |
Header visibility mode |
| 276 |
'all': both row and column headers visible |
| 277 |
'row': only row headers visible |
| 278 |
'column': only column headers visible |
| 279 |
'none': neither row and column headers visible |
| 280 |
dataframe : pandas dataframe |
| 281 |
Data to display on Data Grid. |
| 282 |
renderers : dict |
| 283 |
Custom renderers to use for cell rendering. Keys of dictionary specify |
| 284 |
column name, and value specifies the renderer |
| 285 |
default_renderer : CellRenderer (default: TextRenderer) |
| 286 |
Default renderer to use for cell rendering |
| 287 |
header_renderer : CellRenderer (default: TextRenderer) |
| 288 |
Renderer to use for header cell rendering |
| 289 |
corner_renderer : CellRenderer (default: TextRenderer) |
| 290 |
Renderer to use for corner header cell rendering |
| 291 |
selection_mode : {'row', 'column', 'cell', 'none'} (default: 'none') |
| 292 |
Selection mode used when user clicks on grid or makes selections |
| 293 |
programmatically. |
| 294 |
'row': Selecting a cell will select all the cells on the same row |
| 295 |
'column': Selecting a cell will select all the cells on the same column |
| 296 |
'cell': Individual cell selection |
| 297 |
'none': Selection disabled |
| 298 |
selections : list of dict |
| 299 |
List of all selections. Selections are represented as rectangular |
| 300 |
regions. Rectangles are defined as dictionaries with keys: |
| 301 |
'r1': start row, 'c1': start column, 'r2': end row, 'c2': end column. |
| 302 |
Start of rectangle is top-left corner and end is bottom-right corner |
| 303 |
editable : boolean (default: false) |
| 304 |
Boolean indicating whether cell grid can be directly edited |
| 305 |
column_widths : Dict of strings to int (default: {}) |
| 306 |
Dict to specify custom column sizes |
| 307 |
The keys (strings) indicate the names of the columns |
| 308 |
The values (integers) indicate the widths |
| 309 |
auto_fit_columns : Bool (default: False) |
| 310 |
Specify whether column width should automatically be |
| 311 |
determined by the grid |
| 312 |
auto_fit_params : Dict. Specify column auto fit parameters. |
| 313 |
Supported parameters: |
| 314 |
1) area: where to resize column widths - 'row-header', |
| 315 |
'body' or 'all' (default) |
| 316 |
2) padding: add padding to resized column widths (15 pixels by default) |
| 317 |
3) numCols: cap the number of columns to be resized (None) |
| 318 |
grid_style : Dict of {propertyName: string | VegaExpr | Dict} |
| 319 |
Dict to specify global grid styles. |
| 320 |
The keys (strings) indicate the styling property |
| 321 |
The values (css color properties or Vega Expression) indicate the values |
| 322 |
See below for all supported styling properties |
| 323 |
index_name : str (default: "key") |
| 324 |
String to specify the index column name. **Only set when the grid |
| 325 |
is constructed and is not an observable traitlet** |
| 326 |
horizontal_stripes : bool (default: False) |
| 327 |
Enable themed coloring of alternate grid rows |
| 328 |
vertical_stripes : bool (default: False) |
| 329 |
Enable themed coloring of alternate grid columns |
| 330 |
|
| 331 |
Accessors (not observable traitlets) |
| 332 |
--------- |
| 333 |
selected_cells : list of dict |
| 334 |
List of selected cells. Each cell is represented as a dictionary |
| 335 |
with keys 'r': row and 'c': column |
| 336 |
selected_cell_values : list |
| 337 |
List of values for all selected cells. |
| 338 |
selected_cell_iterator : iterator |
| 339 |
An iterator to traverse selected cells one by one. |
| 340 |
|
| 341 |
Supported styling properties: |
| 342 |
void_color : color of the area where the grid is not painted |
| 343 |
on the canvas |
| 344 |
background_color : background color for all body cells |
| 345 |
row_background_color : row-wise background color (can take |
| 346 |
a string or Vega Expression) |
| 347 |
column_background_color : column-wise background color (can take a |
| 348 |
string of Vega Expression) |
| 349 |
grid_line_color : color of both vertical and horizontal grid lines |
| 350 |
vertical_grid_line_color : vertical grid line color |
| 351 |
horizontal_grid_line_color : horizontal grid line color |
| 352 |
header_background_color : background color for all non-body cells |
| 353 |
(index and columns) |
| 354 |
header_grid_line_color : grid line color for all non-body |
| 355 |
cells (index and columns) |
| 356 |
header_vertical_grid_line_color : vertical grid line color |
| 357 |
for all non-body cells |
| 358 |
header_horizontal_grid_line_color : horizontal grid line color |
| 359 |
for all non-body cells |
| 360 |
selection_fill_color : fill color of selected area |
| 361 |
selection_border_color : border color of selected area |
| 362 |
header_selection_fill_color : fill color of headers intersecting with |
| 363 |
selected area at column or row |
| 364 |
header_selection_border_color : border color of headers |
| 365 |
intersecting with selected area at column or row |
| 366 |
cursor_fill_color : fill color of cursor |
| 367 |
cursor_border_color : border color of cursor |
| 368 |
scroll_shadow : Dict of color parameters for scroll shadow (vertical and |
| 369 |
horizontal). Takes three parameters: |
| 370 |
size : size of shadow in pixels |
| 371 |
color1 : gradient color 1 |
| 372 |
color2 : gradient color 2 |
| 373 |
color3 : gradient color 3 |
| 374 |
""" |
| 375 |
|
| 376 |
_model_name = Unicode("DataGridModel").tag(sync=True) |
| 377 |
_model_module = Unicode(module_name).tag(sync=True) |
| 378 |
_model_module_version = Unicode(module_version).tag(sync=True) |
| 379 |
_view_name = Unicode("DataGridView").tag(sync=True) |
| 380 |
_view_module = Unicode(module_name).tag(sync=True) |
| 381 |
_view_module_version = Unicode(module_version).tag(sync=True) |
| 382 |
|
| 383 |
base_row_size = Int(20).tag(sync=True) |
| 384 |
base_column_size = Int(64).tag(sync=True) |
| 385 |
base_row_header_size = Int(64).tag(sync=True) |
| 386 |
base_column_header_size = Int(20).tag(sync=True) |
| 387 |
|
| 388 |
header_visibility = Enum( |
| 389 |
default_value="all", values=["all", "row", "column", "none"] |
| 390 |
).tag(sync=True) |
| 391 |
|
| 392 |
_transforms = List(Dict()).tag(sync=True, **widget_serialization) |
| 393 |
_visible_rows = List(Int()).tag(sync=True) |
| 394 |
_data = Dict().tag(sync=True, **_data_serialization) |
| 395 |
|
| 396 |
renderers = Dict(Instance(CellRenderer)).tag( |
| 397 |
sync=True, **_widgets_dict_serialization |
| 398 |
) |
| 399 |
default_renderer = Instance(CellRenderer).tag( |
| 400 |
sync=True, **widget_serialization |
| 401 |
) |
| 402 |
header_renderer = Instance(CellRenderer, allow_none=True).tag( |
| 403 |
sync=True, **widget_serialization |
| 404 |
) |
| 405 |
corner_renderer = Instance(CellRenderer, allow_none=True).tag( |
| 406 |
sync=True, **widget_serialization |
| 407 |
) |
| 408 |
selection_mode = Enum( |
| 409 |
default_value="none", values=["row", "column", "cell", "none"] |
| 410 |
).tag(sync=True) |
| 411 |
selections = List(Dict()).tag(sync=True) |
| 412 |
editable = Bool(False).tag(sync=True) |
| 413 |
column_widths = Dict({}).tag(sync=True, to_json=_data_to_json) |
| 414 |
grid_style = Dict(allow_none=True).tag( |
| 415 |
sync=True, **_widgets_dict_serialization |
| 416 |
) |
| 417 |
auto_fit_columns = Bool(False).tag(sync=True) |
| 418 |
auto_fit_params = Dict( |
| 419 |
{"area": "all", "padding": 30, "numCols": None}, allow_none=False |
| 420 |
).tag(sync=True) |
| 421 |
horizontal_stripes = Bool(False).tag(sync=True) |
| 422 |
vertical_stripes = Bool(False).tag(sync=True) |
| 423 |
|
| 424 |
def __init__(self, dataframe, index_name=None, **kwargs): |
| 425 |
# Setting default index name if not explicitly |
| 426 |
# set by the user. |
| 427 |
self._index_name = index_name |
| 428 |
self.data = dataframe |
| 429 |
super().__init__(**kwargs) |
| 430 |
self._cell_click_handlers = CallbackDispatcher() |
| 431 |
self._cell_change_handlers = CallbackDispatcher() |
| 432 |
self.on_msg(self.__handle_custom_msg) |
| 433 |
self._set_renderer_defaults() |
| 434 |
|
| 435 |
def __handle_custom_msg(self, _, content, buffers): # noqa: U101,U100 |
| 436 |
if content["event_type"] == "cell-changed": |
| 437 |
row = content["row"] |
| 438 |
column = self._column_index_to_name( |
| 439 |
self._data, content["column_index"] |
| 440 |
) |
| 441 |
value = content["value"] |
| 442 |
# update data on kernel |
| 443 |
self._data["data"].loc[row, column] = value |
| 444 |
# notify python listeners |
| 445 |
self._cell_change_handlers( |
| 446 |
{ |
| 447 |
"row": row, |
| 448 |
"column": column, |
| 449 |
"column_index": content["column_index"], |
| 450 |
"value": value, |
| 451 |
} |
| 452 |
) |
| 453 |
elif content["event_type"] == "cell-click": |
| 454 |
# notify python listeners |
| 455 |
self._cell_click_handlers( |
| 456 |
{ |
| 457 |
"region": content["region"], |
| 458 |
"column": content["column"], |
| 459 |
"column_index": content["column_index"], |
| 460 |
"row": content["row"], |
| 461 |
"primary_key_row": content["primary_key_row"], |
| 462 |
"cell_value": content["cell_value"], |
| 463 |
} |
| 464 |
) |
| 465 |
|
| 466 |
@property |
| 467 |
def data(self): |
| 468 |
trimmed_primary_key = self._data["schema"]["primaryKey"][:-1] |
| 469 |
if "data" in self._data: |
| 470 |
df = pd.DataFrame(self._data["data"]) |
| 471 |
else: |
| 472 |
df = pd.DataFrame( |
| 473 |
{value["name"]: [] for value in self._data["schema"]["fields"]} |
| 474 |
) |
| 475 |
final_df = df.set_index(trimmed_primary_key) |
| 476 |
final_df = final_df[final_df.columns[:-1]] |
| 477 |
final_df.drop(columns=["ipydguuid"], inplace=True, errors="ignore") |
| 478 |
return final_df |
| 479 |
|
| 480 |
@data.setter |
| 481 |
def data(self, dataframe): |
| 482 |
# Reference for the original frame column and index names |
| 483 |
# This is used to when returning the view data model |
| 484 |
self.__dataframe_reference_index_names = dataframe.index.names |
| 485 |
self.__dataframe_reference_columns = dataframe.columns |
| 486 |
dataframe = dataframe.copy() |
| 487 |
|
| 488 |
# Primary key used |
| 489 |
index_key = self.get_dataframe_index(dataframe) |
| 490 |
|
| 491 |
self._data = self.generate_data_object( |
| 492 |
dataframe, "ipydguuid", index_key |
| 493 |
) |
| 494 |
|
| 495 |
@staticmethod |
| 496 |
def generate_data_object(dataframe, guid_key="ipydguuid", index_name="key"): |
| 497 |
dataframe[guid_key] = pd.RangeIndex(0, dataframe.shape[0]) |
| 498 |
|
| 499 |
# Renaming default index name from 'index' to 'key' on |
| 500 |
# single index DataFrames. This allows users to use |
| 501 |
# 'index' as a column name. If 'key' exists, we add _x |
| 502 |
# suffix to id, where { x | 0 <= x < inf } |
| 503 |
if not isinstance(dataframe.index, pd.MultiIndex): |
| 504 |
if index_name in dataframe.columns: |
| 505 |
index = 0 |
| 506 |
new_index_name = f"{index_name}_{index}" |
| 507 |
while new_index_name in dataframe.columns: |
| 508 |
index += 1 |
| 509 |
new_index_name = f"{index_name}_{index}" |
| 510 |
dataframe = dataframe.rename_axis(new_index_name) |
| 511 |
else: |
| 512 |
dataframe = dataframe.rename_axis(index_name) |
| 513 |
|
| 514 |
schema = pd.io.json.build_table_schema(dataframe) |
| 515 |
reset_index_dataframe = dataframe.reset_index() |
| 516 |
data = reset_index_dataframe |
| 517 |
|
| 518 |
# Check for multiple primary keys |
| 519 |
key = reset_index_dataframe.columns[: dataframe.index.nlevels].tolist() |
| 520 |
|
| 521 |
num_index_levels = len(key) if isinstance(key, list) else 1 |
| 522 |
|
| 523 |
# Check for nested columns in schema, if so, we need to update the |
| 524 |
# schema to represent the actual column name values |
| 525 |
if isinstance(schema["fields"][-1]["name"], tuple): |
| 526 |
num_column_levels = len(dataframe.columns.levels) |
| 527 |
primary_key = key.copy() |
| 528 |
|
| 529 |
for i in range(num_index_levels): |
| 530 |
new_name = [""] * num_column_levels |
| 531 |
new_name[0] = schema["fields"][i]["name"] |
| 532 |
schema["fields"][i]["name"] = tuple(new_name) |
| 533 |
primary_key[i] = tuple(new_name) |
| 534 |
|
| 535 |
schema["primaryKey"] = primary_key |
| 536 |
uuid_pk = list(key[-1]) |
| 537 |
uuid_pk[0] = guid_key |
| 538 |
schema["primaryKey"].append(tuple(uuid_pk)) |
| 539 |
|
| 540 |
else: |
| 541 |
schema["primaryKey"] = key |
| 542 |
schema["primaryKey"].append(guid_key) |
| 543 |
|
| 544 |
schema["primaryKeyUuid"] = guid_key |
| 545 |
|
| 546 |
return { |
| 547 |
"data": data, |
| 548 |
"schema": schema, |
| 549 |
"fields": [{field["name"]: None} for field in schema["fields"]], |
| 550 |
} |
| 551 |
|
| 552 |
def get_dataframe_index(self, dataframe): |
| 553 |
"""Returns a primary key to be used in ipydatagrid's |
| 554 |
view of the passed DataFrame""" |
| 555 |
|
| 556 |
# Passed index_name takes highest priority |
| 557 |
if self._index_name is not None: |
| 558 |
return self._index_name |
| 559 |
|
| 560 |
# Dataframe with names index used by default |
| 561 |
if dataframe.index.name is not None: |
| 562 |
return dataframe.index.name |
| 563 |
|
| 564 |
# If no index_name param, nor named-index DataFrame |
| 565 |
# have been passed, revert to default "key" |
| 566 |
return "key" |
| 567 |
|
| 568 |
def get_cell_value(self, column_name, primary_key_value): |
| 569 |
"""Gets the value for a single or multiple cells by column name and |
| 570 |
index name. |
| 571 |
|
| 572 |
Tuples should be used to index into multi-index columns.""" |
| 573 |
row_indices = self._get_row_index_of_primary_key(primary_key_value) |
| 574 |
|
| 575 |
if isinstance(column_name, list): |
| 576 |
column_name = tuple(column_name) |
| 577 |
|
| 578 |
return [self._data["data"][column_name][row] for row in row_indices] |
| 579 |
|
| 580 |
def set_cell_value(self, column_name, primary_key_value, new_value): |
| 581 |
"""Sets the value for a single cell by column name and primary key. |
| 582 |
|
| 583 |
Note: This method returns a boolean to indicate if the operation |
| 584 |
was successful. |
| 585 |
""" |
| 586 |
row_indices = self._get_row_index_of_primary_key(primary_key_value) |
| 587 |
# Bail early if key could not be found |
| 588 |
if not row_indices: |
| 589 |
return False |
| 590 |
|
| 591 |
if isinstance(column_name, list): |
| 592 |
column_name = tuple(column_name) |
| 593 |
|
| 594 |
# Iterate over all indices |
| 595 |
outcome = True |
| 596 |
for row_index in row_indices: |
| 597 |
has_column = column_name in self._data["data"] |
| 598 |
if has_column and row_index is not None: |
| 599 |
self._data["data"].loc[row_index, column_name] = new_value |
| 600 |
self._notify_cell_change(row_index, column_name, new_value) |
| 601 |
else: |
| 602 |
outcome = False |
| 603 |
return outcome |
| 604 |
|
| 605 |
def set_row_value(self, primary_key_value, new_value): |
| 606 |
"""Sets the value for a row by and primary key. |
| 607 |
|
| 608 |
Note: This method returns a boolean to indicate if the operation |
| 609 |
was successful. |
| 610 |
""" |
| 611 |
row_indices = self._get_row_index_of_primary_key(primary_key_value) |
| 612 |
# Bail early if key could not be found |
| 613 |
if not row_indices: |
| 614 |
return False |
| 615 |
|
| 616 |
# Iterate over all indices |
| 617 |
for row_index in row_indices: |
| 618 |
column_index = 0 |
| 619 |
column = DataGrid._column_index_to_name(self._data, column_index) |
| 620 |
while column is not None: |
| 621 |
self._data["data"].loc[row_index, column] = new_value[ |
| 622 |
column_index |
| 623 |
] |
| 624 |
|
| 625 |
column_index = column_index + 1 |
| 626 |
column = DataGrid._column_index_to_name( |
| 627 |
self._data, column_index |
| 628 |
) |
| 629 |
|
| 630 |
self._notify_row_change(row_index, new_value) |
| 631 |
return True |
| 632 |
|
| 633 |
def get_cell_value_by_index(self, column_name, row_index): |
| 634 |
"""Gets the value for a single cell by column name and row index.""" |
| 635 |
return self._data["data"][column_name][row_index] |
| 636 |
|
| 637 |
def set_cell_value_by_index(self, column_name, row_index, new_value): |
| 638 |
"""Sets the value for a single cell by column name and row index. |
| 639 |
|
| 640 |
Note: This method returns a boolean to indicate if the operation |
| 641 |
was successful. |
| 642 |
""" |
| 643 |
has_column = column_name in self._data["data"] |
| 644 |
if has_column and 0 <= row_index < len(self._data["data"][column_name]): |
| 645 |
self._data["data"].loc[row_index, column_name] = new_value |
| 646 |
self._notify_cell_change(row_index, column_name, new_value) |
| 647 |
return True |
| 648 |
return False |
| 649 |
|
| 650 |
def _notify_cell_change(self, row, column, value): |
| 651 |
column_index = self._column_name_to_index(column) |
| 652 |
# notify python listeners |
| 653 |
self._cell_change_handlers( |
| 654 |
{ |
| 655 |
"row": row, |
| 656 |
"column": column, |
| 657 |
"column_index": column_index, |
| 658 |
"value": value, |
| 659 |
} |
| 660 |
) |
| 661 |
# notify front-end |
| 662 |
self.comm.send( |
| 663 |
data={ |
| 664 |
"method": "custom", |
| 665 |
"content": { |
| 666 |
"event_type": "cell-changed", |
| 667 |
"row": row, |
| 668 |
"column": column, |
| 669 |
"column_index": column_index, |
| 670 |
"value": value, |
| 671 |
}, |
| 672 |
} |
| 673 |
) |
| 674 |
|
| 675 |
def _notify_row_change(self, row, value): |
| 676 |
# notify front-end |
| 677 |
self.comm.send( |
| 678 |
data={ |
| 679 |
"method": "custom", |
| 680 |
"content": { |
| 681 |
"event_type": "row-changed", |
| 682 |
"row": row, |
| 683 |
"value": value, |
| 684 |
}, |
| 685 |
} |
| 686 |
) |
| 687 |
|
| 688 |
def get_visible_data(self): |
| 689 |
"""Returns a dataframe of the current View.""" |
| 690 |
data = deepcopy(self._data) |
| 691 |
if self._visible_rows: |
| 692 |
data["data"] = data["data"].reindex(self._visible_rows) |
| 693 |
|
| 694 |
at = self._data["schema"]["primaryKey"] |
| 695 |
return_df = pd.DataFrame(data["data"]).set_index(at) |
| 696 |
return_df.index = return_df.index.droplevel(return_df.index.nlevels - 1) |
| 697 |
return_df.index.names = self.__dataframe_reference_index_names |
| 698 |
return_df.columns = self.__dataframe_reference_columns |
| 699 |
return return_df |
| 700 |
|
| 701 |
def transform(self, transforms): |
| 702 |
"""Apply a list of transformation to this DataGrid.""" |
| 703 |
# TODO: Validate this input, or let it fail on view side? |
| 704 |
self._transforms = transforms |
| 705 |
|
| 706 |
def revert(self): |
| 707 |
"""Revert all transformations.""" |
| 708 |
self._transforms = [] |
| 709 |
|
| 710 |
@default("default_renderer") |
| 711 |
def _default_renderer(self): |
| 712 |
return TextRenderer() |
| 713 |
|
| 714 |
def clear_selection(self): |
| 715 |
"""Clears all selections.""" |
| 716 |
self.selections.clear() |
| 717 |
self.send_state("selections") |
| 718 |
|
| 719 |
def select(self, row1, column1, row2=None, column2=None, clear_mode="none"): |
| 720 |
""" |
| 721 |
Select an individual cell or rectangular cell region. |
| 722 |
Parameters |
| 723 |
---------- |
| 724 |
row1 : int |
| 725 |
Row index for individual cell selection or |
| 726 |
start row index for rectangular region selection. |
| 727 |
column1 : int |
| 728 |
Column index for individual cell selection or |
| 729 |
start column index for rectangular region selection. |
| 730 |
row2 : int or None, optional (default: None) |
| 731 |
End row index for rectangular region selection. |
| 732 |
column2 : int or None, optional (default: None) |
| 733 |
End column index for rectangular region selection. |
| 734 |
clear_mode : string, optional, {'all', 'current', 'none'} |
| 735 |
(default: 'none') |
| 736 |
Clear mode to use when there are pre-existing selections. |
| 737 |
'all' removes all pre-existing selections |
| 738 |
'current' removes last pre-existing selection |
| 739 |
'none' keeps pre-existing selections |
| 740 |
""" |
| 741 |
if row2 is None or column2 is None: |
| 742 |
row2, column2 = row1, column1 |
| 743 |
|
| 744 |
if clear_mode == "all": |
| 745 |
self.selections.clear() |
| 746 |
elif clear_mode == "current" and len(self.selections) > 0: |
| 747 |
self.selections.pop() |
| 748 |
|
| 749 |
self.selections.append( |
| 750 |
{ |
| 751 |
"r1": min(row1, row2), |
| 752 |
"c1": min(column1, column2), |
| 753 |
"r2": max(row1, row2), |
| 754 |
"c2": max(column1, column2), |
| 755 |
} |
| 756 |
) |
| 757 |
self.send_state("selections") |
| 758 |
|
| 759 |
@property |
| 760 |
def selected_cells(self): |
| 761 |
""" |
| 762 |
List of selected cells. Each cell is represented as a dictionary |
| 763 |
with keys 'r': row and 'c': column |
| 764 |
""" |
| 765 |
return SelectionHelper( |
| 766 |
self._data, self.selections, self.selection_mode |
| 767 |
).all() |
| 768 |
|
| 769 |
@property |
| 770 |
def selected_visible_cell_iterator(self): |
| 771 |
""" |
| 772 |
An iterator to traverse selected visible cells one by one. |
| 773 |
""" |
| 774 |
# Copy of the front-end data model |
| 775 |
view_data = self.get_visible_data() |
| 776 |
|
| 777 |
# Get primary key from dataframe |
| 778 |
index_key = self.get_dataframe_index(view_data) |
| 779 |
|
| 780 |
# Serielize to JSON table schema |
| 781 |
view_data_object = self.generate_data_object( |
| 782 |
view_data, "ipydguuid", index_key |
| 783 |
) |
| 784 |
return SelectionHelper( |
| 785 |
view_data_object, self.selections, self.selection_mode |
| 786 |
) |
| 787 |
|
| 788 |
@property |
| 789 |
def selected_cell_values(self): |
| 790 |
""" |
| 791 |
List of values for all selected cells. |
| 792 |
""" |
| 793 |
return self.selected_visible_cell_iterator.all_values() |
| 794 |
|
| 795 |
@property |
| 796 |
def selected_cell_iterator(self): |
| 797 |
""" |
| 798 |
An iterator to traverse selected cells one by one. |
| 799 |
""" |
| 800 |
return SelectionHelper(self._data, self.selections, self.selection_mode) |
| 801 |
|
| 802 |
@validate("selections") |
| 803 |
def _validate_selections(self, proposal): |
| 804 |
selections = proposal["value"] |
| 805 |
|
| 806 |
for rectangle in selections: |
| 807 |
r1 = min(rectangle["r1"], rectangle["r2"]) |
| 808 |
c1 = min(rectangle["c1"], rectangle["c2"]) |
| 809 |
r2 = max(rectangle["r1"], rectangle["r2"]) |
| 810 |
c2 = max(rectangle["c1"], rectangle["c2"]) |
| 811 |
rectangle["r1"] = r1 |
| 812 |
rectangle["c1"] = c1 |
| 813 |
rectangle["r2"] = r2 |
| 814 |
rectangle["c2"] = c2 |
| 815 |
|
| 816 |
return selections |
| 817 |
|
| 818 |
@validate("editable") |
| 819 |
def _validate_editable(self, proposal): |
| 820 |
value = proposal["value"] |
| 821 |
if value and self.selection_mode == "none": |
| 822 |
self.selection_mode = "cell" |
| 823 |
return value |
| 824 |
|
| 825 |
@validate("_transforms") |
| 826 |
def _validate_transforms(self, proposal): |
| 827 |
transforms = proposal["value"] |
| 828 |
field_len = len(self._data["schema"]["fields"]) |
| 829 |
for transform in transforms: |
| 830 |
if "columnIndex" in transform: |
| 831 |
warnings.warn( |
| 832 |
"Applying transforms on columnIndex is deprecated, " |
| 833 |
"please provide the column name instead", |
| 834 |
DeprecationWarning, |
| 835 |
stacklevel=4, |
| 836 |
) |
| 837 |
|
| 838 |
if "column" not in transform: |
| 839 |
transform["column"] = self._data["schema"]["fields"][ |
| 840 |
transform["columnIndex"] |
| 841 |
]["name"] |
| 842 |
|
| 843 |
if transform["columnIndex"] > field_len: |
| 844 |
raise ValueError("Column index is out of bounds.") |
| 845 |
|
| 846 |
return transforms |
| 847 |
|
| 848 |
@validate("_data") |
| 849 |
def _validate_data(self, proposal): |
| 850 |
table_schema = proposal["value"] |
| 851 |
column_list = [f["name"] for f in table_schema["schema"]["fields"]] |
| 852 |
if len(column_list) != len(set(column_list)): |
| 853 |
msg = "The dataframe must not contain duplicate column names." |
| 854 |
raise ValueError(msg) |
| 855 |
return table_schema |
| 856 |
|
| 857 |
def on_cell_change(self, callback, remove=False): |
| 858 |
"""Register a callback to execute when a cell value changed. |
| 859 |
|
| 860 |
The callback will be called with one argument, the dictionary |
| 861 |
containing cell information with keys |
| 862 |
"row", "column", "column_index", "value". |
| 863 |
|
| 864 |
Parameters |
| 865 |
---------- |
| 866 |
remove: bool (optional) |
| 867 |
Set to true to remove the callback from the list of callbacks. |
| 868 |
""" |
| 869 |
self._cell_change_handlers.register_callback(callback, remove=remove) |
| 870 |
|
| 871 |
def on_cell_click(self, callback, remove=False): |
| 872 |
"""Register a callback to execute when a cell is clicked. |
| 873 |
|
| 874 |
The callback will be called with one argument, the dictionary |
| 875 |
containing cell information with following keys: |
| 876 |
"region", "column", "column_index", "row", "primary_key_row", |
| 877 |
"cell_value" |
| 878 |
|
| 879 |
Parameters |
| 880 |
---------- |
| 881 |
remove: bool (optional) |
| 882 |
Set to true to remove the callback from the list of callbacks. |
| 883 |
""" |
| 884 |
self._cell_click_handlers.register_callback(callback, remove=remove) |
| 885 |
|
| 886 |
@staticmethod |
| 887 |
def _column_index_to_name(data, column_index): |
| 888 |
if "schema" not in data or "fields" not in data["schema"]: |
| 889 |
return None |
| 890 |
col_headers = DataGrid._get_col_headers(data) |
| 891 |
return ( |
| 892 |
None |
| 893 |
if len(col_headers) <= column_index |
| 894 |
else col_headers[column_index] |
| 895 |
) |
| 896 |
|
| 897 |
@staticmethod |
| 898 |
def _get_col_headers(data): |
| 899 |
primary_keys = ( |
| 900 |
[] |
| 901 |
if "primaryKey" not in data["schema"] |
| 902 |
else data["schema"]["primaryKey"] |
| 903 |
) |
| 904 |
col_headers = [ |
| 905 |
field["name"] |
| 906 |
for field in data["schema"]["fields"] |
| 907 |
if field["name"] not in primary_keys |
| 908 |
] |
| 909 |
return col_headers |
| 910 |
|
| 911 |
def _column_name_to_index(self, column_name): |
| 912 |
if "schema" not in self._data or "fields" not in self._data["schema"]: |
| 913 |
return None |
| 914 |
col_headers = self._get_col_headers(self._data) |
| 915 |
try: |
| 916 |
return col_headers.index(column_name) |
| 917 |
except ValueError: |
| 918 |
pass |
| 919 |
|
| 920 |
def _get_row_index_of_primary_key(self, value): |
| 921 |
value = value if isinstance(value, list) else [value] |
| 922 |
schema = self._data["schema"] |
| 923 |
key = schema["primaryKey"][:-1] # Omitting ipydguuid |
| 924 |
if len(value) != len(key): |
| 925 |
raise ValueError( |
| 926 |
"The provided primary key value must be the same length " |
| 927 |
"as the primary key." |
| 928 |
) |
| 929 |
|
| 930 |
df = self._data["data"] |
| 931 |
return pd.RangeIndex(len(df))[ |
| 932 |
(df[key] == value).all(axis="columns") |
| 933 |
].to_list() |
| 934 |
|
| 935 |
@staticmethod |
| 936 |
def _get_cell_value_by_numerical_index(data, column_index, row_index): |
| 937 |
"""Gets the value for a single cell by column index and row index.""" |
| 938 |
# TODO This is really not efficient, we should speed it up |
| 939 |
column = DataGrid._column_index_to_name(data, column_index) |
| 940 |
if column is None: |
| 941 |
return None |
| 942 |
return data["data"].loc[row_index, column] |
| 943 |
|
| 944 |
def _set_renderer_defaults(self): |
| 945 |
# Set sensible default values for renderers that are not completely |
| 946 |
# specified, such as missing a min or max value. |
| 947 |
|
| 948 |
data = None # Only read data once, and only if necessary. |
| 949 |
|
| 950 |
for name, renderer in self.renderers.items(): |
| 951 |
if isinstance(renderer, BarRenderer): |
| 952 |
from bqplot import DateScale, LinearScale, Scale |
| 953 |
|
| 954 |
if renderer.bar_value is None: |
| 955 |
# If BarRenderer.bar_value is not specified, create an |
| 956 |
# appropriate Scale based on the column data type. |
| 957 |
col_schema = next( |
| 958 |
filter( |
| 959 |
lambda x: x["name"] == name, |
| 960 |
self._data["schema"]["fields"], |
| 961 |
) |
| 962 |
) |
| 963 |
is_date = col_schema["type"] in ("date", "time", "datetime") |
| 964 |
if is_date: |
| 965 |
renderer.bar_value = DateScale() |
| 966 |
else: |
| 967 |
renderer.bar_value = LinearScale() |
| 968 |
|
| 969 |
scale = renderer.bar_value |
| 970 |
if ( |
| 971 |
isinstance(scale, Scale) |
| 972 |
and scale.has_trait("min") |
| 973 |
and scale.has_trait("max") |
| 974 |
and (scale.min is None or scale.max is None) |
| 975 |
): |
| 976 |
# Set min and/or max from column data. |
| 977 |
if data is None: |
| 978 |
data = self.data # Only want to get the data once |
| 979 |
column_data = data[name] |
| 980 |
is_date = isinstance(scale, DateScale) |
| 981 |
if scale.min is None: |
| 982 |
min = column_data.min() |
| 983 |
scale.min = min if is_date else float(min) |
| 984 |
if scale.max is None: |
| 985 |
max = column_data.max() |
| 986 |
scale.max = max if is_date else float(max) |
| 987 |
|
| 988 |
|
| 989 |
class StreamingDataGrid(DataGrid): |
| 990 |
"""A blazingly fast Grid Widget. |
| 991 |
This widget needs a live kernel for working |
| 992 |
(does not work when embedded in static HTML) |
| 993 |
""" |
| 994 |
|
| 995 |
_model_name = Unicode("StreamingDataGridModel").tag(sync=True) |
| 996 |
_view_name = Unicode("StreamingDataGridView").tag(sync=True) |
| 997 |
|
| 998 |
_row_count = Int(0).tag(sync=True) |
| 999 |
_debounce_delay = Int(160).tag(sync=True) |
| 1000 |
|
| 1001 |
def __init__(self, *args, debounce_delay=160, **kwargs): |
| 1002 |
super().__init__(*args, **kwargs) |
| 1003 |
|
| 1004 |
self._debounce_delay = debounce_delay |
| 1005 |
self._transformed_data = None |
| 1006 |
|
| 1007 |
self.on_msg(self._handle_comm_msg) |
| 1008 |
|
| 1009 |
@property |
| 1010 |
def data(self): |
| 1011 |
return super().data |
| 1012 |
|
| 1013 |
@data.setter |
| 1014 |
def data(self, dataframe): |
| 1015 |
self.__dataframe_reference_index_names = dataframe.index.names |
| 1016 |
self.__dataframe_reference_columns = dataframe.columns |
| 1017 |
# Not making a copy in the streaming grid |
| 1018 |
self.__dataframe_reference = dataframe |
| 1019 |
|
| 1020 |
# TODO support column indices being integers |
| 1021 |
dataframe.columns = dataframe.columns.map(str) |
| 1022 |
|
| 1023 |
# Primary key used |
| 1024 |
index_key = self.get_dataframe_index(dataframe) |
| 1025 |
|
| 1026 |
self._data_object = self.generate_data_object( |
| 1027 |
dataframe, "ipydguuid", index_key |
| 1028 |
) |
| 1029 |
|
| 1030 |
self._row_count = len(self._data_object["data"]) |
| 1031 |
|
| 1032 |
self._data = { |
| 1033 |
"data": {}, |
| 1034 |
"schema": self._data_object["schema"], |
| 1035 |
"fields": self._data_object["fields"], |
| 1036 |
} |
| 1037 |
|
| 1038 |
def tick(self): |
| 1039 |
"""Notify that the underlying dataframe has changed.""" |
| 1040 |
self.send({"event_type": "tick"}) |
| 1041 |
|
| 1042 |
def _apply_frontend_transforms(self, frontend_transforms, dataframe): |
| 1043 |
for transform in frontend_transforms: |
| 1044 |
if transform["type"] == "sort": |
| 1045 |
column = transform["column"] |
| 1046 |
desc = transform.get("desc", False) |
| 1047 |
dataframe = dataframe.sort_values( |
| 1048 |
by=[column], ascending=not desc |
| 1049 |
) |
| 1050 |
elif transform["type"] == "filter": |
| 1051 |
column = transform["column"] |
| 1052 |
operator = transform["operator"] |
| 1053 |
value = transform["value"] |
| 1054 |
if operator == "<": |
| 1055 |
dataframe = dataframe[dataframe[column] < value] |
| 1056 |
elif operator == ">": |
| 1057 |
dataframe = dataframe[dataframe[column] > value] |
| 1058 |
elif operator == "=": |
| 1059 |
dataframe = dataframe[dataframe[column] == value] |
| 1060 |
elif operator == "<=": |
| 1061 |
dataframe = dataframe[dataframe[column] <= value] |
| 1062 |
elif operator == ">=": |
| 1063 |
dataframe = dataframe[dataframe[column] >= value] |
| 1064 |
elif operator == "!=": |
| 1065 |
dataframe = dataframe[dataframe[column] != value] |
| 1066 |
elif operator == "empty": |
| 1067 |
dataframe = dataframe[dataframe[column].isna()] |
| 1068 |
elif operator == "notempty": |
| 1069 |
dataframe = dataframe[dataframe[column].notna()] |
| 1070 |
elif operator == "in": |
| 1071 |
value = pd.Series(value, dtype=dataframe[column].dtype) |
| 1072 |
dataframe = dataframe[dataframe[column].isin(value)] |
| 1073 |
elif operator == "between": |
| 1074 |
dataframe = dataframe[ |
| 1075 |
dataframe[column].between(value[0], value[1]) |
| 1076 |
] |
| 1077 |
elif operator == "startswith": |
| 1078 |
dataframe = dataframe[ |
| 1079 |
dataframe[column].str.startswith(value) |
| 1080 |
] |
| 1081 |
elif operator == "endswith": |
| 1082 |
dataframe = dataframe[dataframe[column].str.endswith(value)] |
| 1083 |
elif operator == "stringContains": |
| 1084 |
dataframe = dataframe[dataframe[column].str.contains(value)] |
| 1085 |
elif operator == "contains": |
| 1086 |
dataframe = dataframe[dataframe[column].str.contains(value)] |
| 1087 |
elif operator == "!contains": |
| 1088 |
dataframe = dataframe[ |
| 1089 |
not dataframe[column].str.contains(value) |
| 1090 |
] |
| 1091 |
elif operator == "isOnSameDay": |
| 1092 |
value = pd.to_datetime(value).date() |
| 1093 |
dataframe = dataframe[ |
| 1094 |
pd.to_datetime(dataframe[column]).dt.date == value |
| 1095 |
] |
| 1096 |
else: |
| 1097 |
raise RuntimeError( |
| 1098 |
f"Unrecognised filter operator '{operator}'" |
| 1099 |
) |
| 1100 |
|
| 1101 |
return dataframe |
| 1102 |
|
| 1103 |
def _handle_comm_msg(self, _, content, _buffs): |
| 1104 |
event_type = content.get("type", "") |
| 1105 |
|
| 1106 |
if event_type == "data-request": |
| 1107 |
r1 = content.get("r1") |
| 1108 |
r2 = content.get("r2") |
| 1109 |
c1 = content.get("c1") |
| 1110 |
c2 = content.get("c2") |
| 1111 |
|
| 1112 |
# Filter/sort whole dataset before selecting rows/cols of interest |
| 1113 |
if self._transformed_data is not None: |
| 1114 |
# Use existing transformed data. |
| 1115 |
value = self._transformed_data |
| 1116 |
else: |
| 1117 |
value = self.__dataframe_reference |
| 1118 |
|
| 1119 |
value = value.iloc[r1 : r2 + 1, c1 : c2 + 1] |
| 1120 |
|
| 1121 |
serialized, buffers = self._serialize_helper(value) |
| 1122 |
|
| 1123 |
answer = { |
| 1124 |
"event_type": "data-reply", |
| 1125 |
"value": serialized, |
| 1126 |
"r1": r1, |
| 1127 |
"r2": r2, |
| 1128 |
"c1": c1, |
| 1129 |
"c2": c2, |
| 1130 |
} |
| 1131 |
|
| 1132 |
self.send(answer, buffers) |
| 1133 |
|
| 1134 |
elif event_type == "unique-values-request": |
| 1135 |
column = content.get("column") |
| 1136 |
original = self.__dataframe_reference[column].drop_duplicates() |
| 1137 |
serialized, buffers = self._serialize_helper(pd.DataFrame(original)) |
| 1138 |
|
| 1139 |
answer = { |
| 1140 |
"event_type": "unique-values-reply", |
| 1141 |
"column": column, |
| 1142 |
"value": serialized, |
| 1143 |
} |
| 1144 |
self.send(answer, buffers) |
| 1145 |
|
| 1146 |
@observe("_transforms") |
| 1147 |
def _on_transforms_changed(self, change): |
| 1148 |
# Transforms is an array of dicts. |
| 1149 |
frontend_transforms = change["new"] |
| 1150 |
|
| 1151 |
self._transformed_data = None |
| 1152 |
data = self.__dataframe_reference |
| 1153 |
|
| 1154 |
if frontend_transforms: |
| 1155 |
self._transformed_data = self._apply_frontend_transforms( |
| 1156 |
frontend_transforms, data |
| 1157 |
) |
| 1158 |
data = self._transformed_data |
| 1159 |
|
| 1160 |
self._row_count = len(data) # Sync to frontend. |
| 1161 |
|
| 1162 |
# Should only request a tick if the transforms have changed. |
| 1163 |
self.tick() |
| 1164 |
|
| 1165 |
def _serialize_helper(self, dataframe): |
| 1166 |
# Primary key used |
| 1167 |
index_key = self.get_dataframe_index(dataframe) |
| 1168 |
|
| 1169 |
serialized = _data_serialization_impl( |
| 1170 |
self.generate_data_object(dataframe, "ipydguuid", index_key), None |
| 1171 |
) |
| 1172 |
|
| 1173 |
# Extract all buffers |
| 1174 |
buffers = [] |
| 1175 |
for column in serialized["data"].keys(): |
| 1176 |
col = serialized["data"][column] |
| 1177 |
if not isinstance(col, list) and col["type"] != "raw": |
| 1178 |
buffers.append(col["value"]) |
| 1179 |
col["value"] = len(buffers) - 1 |
| 1180 |
|
| 1181 |
return serialized, buffers |