def get_results(filter_expression="", row_types=['runattr', 'itervar', 'config', 'scalar', 'vector', 'statistic', 'histogram', 'param', 'attr'], omit_unused_columns=True, start_time=-inf, end_time=inf):
"""
Returns a filtered set of results and metadata in CSV-like format.
The items can be any type, even mixed together in a single `DataFrame`.
They are selected from the complete set of data referenced by the analysis file (`.anf`),
including only those for which the given `filter_expression` evaluates to `True`.
# Parameters
- **filter_expression** *(string)*: The filter expression to select the desired items. Example: `module =~ "*host*" AND name =~ "numPacket*"`
- **row_types**: Optional. When given, filters the returned rows by type. Should be a unique list, containing any number of these strings:
`"runattr"`, `"itervar"`, `"config"`, `"scalar"`, `"vector"`, `"statistic"`, `"histogram"`, `"param"`, `"attr"`
- **omit_unused_columns** *(bool)*: Optional. If `True`, all columns that would only contain `None` are removed from the returned DataFrame
- **start_time**, **end_time** *(double)*: Optional time limits to trim the data of vector type results.
The unit is seconds, both the `vectime` and `vecvalue` arrays will be affected, the interval is left-closed, right-open.
# Columns of the returned DataFrame
- **runID** *(string)*: Identifies the simulation run
- **type** *(string)*: Row type, one of the following: scalar, vector, statistics, histogram, runattr, itervar, param, attr
- **module** *(string)*: Hierarchical name (a.k.a. full path) of the module that recorded the result item
- **name** *(string)*: Name of the result item (scalar, statistic, histogram or vector)
- **attrname** *(string)*: Name of the run attribute or result item attribute (in the latter case, the module and name columns identify the result item the attribute belongs to)
- **attrvalue** *(string)*: Value of run and result item attributes, iteration variables, saved ini param settings (runattr, attr, itervar, param)
- **value** *(double or string)*: Output scalar or attribute value
- **count**, **sumweights**, **mean**, **min**, **max**, **stddev** *(double)*: Fields of the statistics or histogram
- **binedges**, **binvalues** *(np.array)*: Histogram bin edges and bin values, as space-separated lists. `len(binedges)==len(binvalues)+1`
- **underflows**, **overflows** *(double)*: Sum of weights (or counts) of underflown and overflown samples of histograms
- **vectime**, **vecvalue** *(np.array)*: Output vector time and value arrays, as space-separated lists
"""
return impl.get_results(**locals())
def get_results(filter_or_dataframe="",
row_types=None,
omit_unused_columns=True,
include_fields_as_scalars=False,
start_time=-inf,
end_time=inf):
"""
Returns a filtered set of results and metadata in a Pandas `DataFrame`. The
items can be any type, even mixed together in a single `DataFrame`. They are
selected from the complete set of data referenced by the analysis file
(`.anf`), including only those for which the given `filter_or_dataframe`
evaluates to `True`.
Parameters:
- `filter_or_dataframe` (string or dataframe): The filter expression to
select the desired items from the inputs, or a dataframe in the "raw"
format. Example: `module =~ "*host*" AND name =~ "numPacket*"`
- `row_types`: Optional. When given, filters the returned rows by type.
Should be a unique list, containing any number of these strings:
`"runattr"`, `"itervar"`, `"config"`, `"scalar"`, `"vector"`,
`"statistic"`, `"histogram"`, `"param"`, `"attr"`
- `omit_unused_columns` (bool): Optional. If `True`, all columns that would
only contain `None` are removed from the returned DataFrame
- `include_fields_as_scalars` (bool): Optional. If `True`, the fields of
statistics and histograms (`:min`, `:mean`, etc.) are also returned as
synthetic scalars.
- `start_time`, `end_time` (double): Optional time limits to trim the data
of vector type results. The unit is seconds, both the `vectime` and
`vecvalue` arrays will be affected, the interval is left-closed,
right-open.
Returns: a `DataFrame` in the "raw" format (see the corresponding section of
the module documentation for details).
"""
if type(filter_or_dataframe) is str:
filter_expression = filter_or_dataframe
del filter_or_dataframe
return impl.get_results(**locals())
else:
if include_fields_as_scalars:
raise ValueError(
"include_fields_as_scalars is not supported when filter_or_dataframe is a dataframe"
)
df = filter_or_dataframe
if row_types is not None:
df = df[df["type"].isin(row_types)]
if omit_unused_columns:
df = df.dropna(axis='columns', how='all')
df.reset_index(inplace=True, drop=True)
if start_time != -inf or end_time != inf:
def crop(row):
t = row['vectime']
v = row['vecvalue']
from_index = np.searchsorted(t, start_time, 'left')
to_index = np.searchsorted(t, end_time, 'left')
row['vectime'] = t[from_index:to_index]
row['vecvalue'] = v[from_index:to_index]
return row
df = df.transform(crop, axis='columns')
return df