Skip to content

nirvana.DataFrame

nirvana.dataframe.frame.DataFrame(data: dict | pd.DataFrame = None, *args, **kwargs)

Bases: LineageMixin

Source code in nirvana/dataframe/frame.py 
32
33
34
35
36
37
38
39
40
41
42
def __init__(
    self,
    data: dict | pd.DataFrame = None,
    *args,
    **kwargs
):
    self._data = pd.DataFrame(data) if isinstance(data, dict) else data
    for col in self._data.columns:
        converted_array, inferred_dtype = infer_and_convert_dtype(self._data[col])
        self._data[col] = converted_array
    self.initialize()

Attributes

columns property writable
primary_key property
nrows property

Functions

__len__()
Source code in nirvana/dataframe/frame.py 
44
45
46
def __len__(self):
    _len = self.nrows
    return _len
__contains__(item)
Source code in nirvana/dataframe/frame.py 
48
49
def __contains__(self, item):
    return self.columns.__contains__(item)
__repr__() -> str
Source code in nirvana/dataframe/frame.py 
51
52
def __repr__(self) -> str:
    return f"{self.__class__.__name__}(nrows={self.nrows}, ncols={len(self.columns)})"
to_pandas() -> pd.DataFrame
Source code in nirvana/dataframe/frame.py 
70
71
def to_pandas(self) -> pd.DataFrame:
    return self._data.copy()
from_external_file(path: str, sep=',', **kwargs) classmethod
Source code in nirvana/dataframe/frame.py 
73
74
75
76
@classmethod
def from_external_file(cls, path: str, sep=',', **kwargs):
    df = pd.read_table(path, sep=sep, **kwargs)
    return cls(df)
head(n=5)
Source code in nirvana/dataframe/frame.py 
78
79
def head(self, n=5):
    return self._data.head(n)
tail(n=5)
Source code in nirvana/dataframe/frame.py 
81
82
def tail(self, n=5):
    return self._data.tail(n)
__getitem__(posidx)
Source code in nirvana/dataframe/frame.py 
111
112
def __getitem__(self, posidx):
    return self._get(posidx)
semantic_map(user_instruction: str, input_columns: list[str], output_columns: list[str], context: list[dict] | str | None = None, model: str | None = None, func: Callable | None = None, strategy: Literal['plain', 'fewshot', 'self-refine'] = 'plain', limit: int | None = None, rate_limit: int = 16, assertions: list[Callable] | None = [])
Source code in nirvana/dataframe/frame.py 
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
def semantic_map(
    self,
    user_instruction: str,
    input_columns: list[str],
    output_columns: list[str],
    context: list[dict] | str | None = None,
    model: str | None = None,
    func: Callable | None = None,
    strategy: Literal["plain", "fewshot", "self-refine"] = "plain",
    limit: int | None = None,
    rate_limit: int = 16,
    assertions: list[Callable] | None = [],
):
    op_kwargs = {
        "user_instruction": user_instruction,
        "input_columns": input_columns,
        "output_columns": output_columns,
        "context": context,
        "model": model,
        "tool": func,
        "strategy": strategy,
        "limit": limit,
        "rate_limit": rate_limit,
        "assertions": assertions,
    }
    data_kwargs = {
        "left_input_fields": self.leaf_node.node_fields.output_fields,
        "right_input_fields": [],
        "output_fields": self.leaf_node.node_fields.left_input_fields + output_columns,
    }
    self.add_operator(op_name="map",
                      op_kwargs=op_kwargs,
                      data_kwargs=data_kwargs,
                      rate_limit=rate_limit)
semantic_filter(user_instruction: str, input_columns: list[str], func: Callable | None = None, context: list[dict] | str | None = None, model: str | None = None, strategy: Literal['plain', 'fewshot', 'self-refine'] = 'plain', limit: int | None = None, rate_limit: int = 16, assertions: list[Callable] | None = [])
Source code in nirvana/dataframe/frame.py 
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
def semantic_filter(
    self,
    user_instruction: str,
    input_columns: list[str],
    func: Callable | None = None,
    context: list[dict] | str | None = None,
    model: str | None = None,
    strategy: Literal["plain", "fewshot", "self-refine"] = "plain",
    limit: int | None = None,
    rate_limit: int = 16,
    assertions: list[Callable] | None = [],
):
    op_kwargs = {
        "user_instruction": user_instruction,
        "input_columns": input_columns,
        "context": context,
        "model": model,
        "tool": func,
        "strategy": strategy,
        "limit": limit,
        "rate_limit": rate_limit,
        "assertions": assertions,
    }
    data_kwargs = {
        "left_input_fields": self.leaf_node.node_fields.output_fields,
        "right_input_fields": [],
        "output_fields": self.leaf_node.node_fields.output_fields,
    }
    self.add_operator(op_name="filter",
                      op_kwargs=op_kwargs,
                      data_kwargs=data_kwargs,
                      rate_limit=rate_limit)
semantic_reduce(user_instruction: str, input_column: str, context: list[dict] | str | None = None, model: str | None = None, func: Callable | None = None, strategy: Literal['plain'] = 'plain', rate_limit: int = 16, assertions: list[Callable] | None = [])
Source code in nirvana/dataframe/frame.py 
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
def semantic_reduce(
    self,
    user_instruction: str,
    input_column: str,
    context: list[dict] | str | None = None,
    model: str | None = None,
    func: Callable | None = None,
    strategy: Literal["plain"] = "plain",
    rate_limit: int = 16,
    assertions: list[Callable] | None = [],
):
    op_kwargs = {
        "user_instruction": user_instruction,
        "input_columns": [input_column],
        "context": context,
        "model": model,
        "tool": func,
        "strategy": strategy,
        "rate_limit": rate_limit,
        "assertions": assertions,
    }
    data_kwargs = {
        "left_input_fields": self.leaf_node.node_fields.output_fields,
        "right_input_fields": [],
        "output_fields": []
    }
    self.add_operator(op_name="reduce",
                      op_kwargs=op_kwargs,
                      data_kwargs=data_kwargs,
                      rate_limit=rate_limit)
semantic_join(other: DataFrame, user_instruction: str, left_on: str, right_on: str, how: Literal['inner', 'left', 'right'] = 'inner', context: list[dict] | str | None = None, model: str | None = None, func: Callable | None = None, strategy: Literal['nest', 'block'] = 'nest', limit: int | None = None, rate_limit: int = 16, assertions: list[Callable] | None = [], batch_size: int = 5)
Source code in nirvana/dataframe/frame.py 
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
def semantic_join(
    self,
    other: "DataFrame",
    user_instruction: str,
    left_on: str,
    right_on: str,
    how: Literal["inner", "left", "right"] = "inner",
    context: list[dict] | str | None = None,
    model: str | None = None,
    func: Callable | None = None,
    strategy: Literal["nest", "block"] = "nest",
    limit: int | None = None,
    rate_limit: int = 16,
    assertions: list[Callable] | None = [],
    batch_size: int = 5,
):
    union_fields = (
        list(set(self.leaf_node.node_fields.output_fields) | set(other.leaf_node.node_fields.output_fields))
    )
    op_kwargs = {
        "user_instruction": user_instruction,
        "left_on": [left_on],
        "right_on": [right_on],
        "how": how,
        "context": context,
        "model": model,
        "tool": func,
        "strategy": strategy,
        "limit": limit,
        "rate_limit": rate_limit,
        "assertions": assertions,
        "batch_size": batch_size,
    }
    data_kwargs = {
        "input_left_fields": self.leaf_node.node_fields.output_fields,
        "input_right_fields": other.leaf_node.node_fields.output_fields,
        "output_fields": union_fields
    }
    self.add_operator(op_name="join",
                      op_kwargs=op_kwargs,
                      data_kwargs=data_kwargs,
                      other=other,
                      rate_limit=rate_limit)
semantic_rank(user_instruction: str, input_column: str, descend: bool = True, context: list[dict] | str | None = None, model: str | None = None, func: Callable | None = None, strategy: Literal['plain'] = 'plain', limit: int | None = None, rate_limit: int = 16, assertions: list[Callable] | None = [])
Source code in nirvana/dataframe/frame.py 
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
def semantic_rank(
    self,
    user_instruction: str,
    input_column: str,
    descend: bool = True,
    context: list[dict] | str | None = None,
    model: str | None = None,
    func: Callable | None = None,
    strategy: Literal["plain"] = "plain",
    limit: int | None = None,
    rate_limit: int = 16,
    assertions: list[Callable] | None = [],
):
    op_kwargs = {
        "user_instruction": user_instruction,
        "input_columns": [input_column],
        "descend": descend,
        "context": context,
        "model": model,
        "tool": func,
        "strategy": strategy,
        "limit": limit,
        "rate_limit": rate_limit,
        "assertions": assertions,
    }
    data_kwargs = {
        "left_input_fields": self.leaf_node.node_fields.output_fields,
        "right_input_fields": [],
        "output_fields": self.leaf_node.node_fields.output_fields,
    }
    self.add_operator(op_name="rank",
                      op_kwargs=op_kwargs,
                      data_kwargs=data_kwargs,
                      rate_limit=rate_limit)
optimize_and_execute(optim_config=None)
Source code in nirvana/dataframe/frame.py 
292
293
294
295
296
297
298
299
300
301
302
def optimize_and_execute(self, optim_config = None):
    self.create_plan_optimizer(optim_config)
    if self.optimizer.config.do_logical_optimization:
        self.leaf_node = self.optimizer.optimize_logical_plan(self.leaf_node)
    if self.optimizer.config.do_physical_optimization:
        output, cost, runtime = self.optimizer.optimize_physical_plan(
            self.leaf_node,
        )
    else:
        output, cost, runtime = self.execute()
    return output, cost, runtime