93 lines
2.8 KiB
Python
93 lines
2.8 KiB
Python
import os
|
|
import sys
|
|
from typing import Any, Dict
|
|
from matplotlib import pyplot as plt
|
|
import seaborn as sns
|
|
import torch
|
|
import ast
|
|
import pandas as pd
|
|
|
|
PROJECT_ROOT = os.path.realpath(os.path.join(os.path.dirname(__file__), "..", ".."))
|
|
MODULE_ROOT = os.path.join(PROJECT_ROOT, "modules")
|
|
sys.path.append(MODULE_ROOT)
|
|
from seaborn_wrapper.common import common_input_types_y_only_bin_pt
|
|
from seaborn_wrapper.util import plot_post_steps
|
|
from seaborn_wrapper.util import set_up_sns_style_palette
|
|
|
|
|
|
class SNSHistogramPt:
|
|
"""
|
|
SNS Histogram Pt:
|
|
Generates a histogram from a PyTorch tensor using Seaborn.
|
|
|
|
category: Plot
|
|
"""
|
|
|
|
@classmethod
|
|
def INPUT_TYPES(cls) -> Dict[str, Any]:
|
|
"""
|
|
Defines the input types for the function.
|
|
|
|
Returns:
|
|
Dict[str, Any]: A dictionary specifying required input types.
|
|
"""
|
|
return common_input_types_y_only_bin_pt()
|
|
|
|
RETURN_TYPES: tuple = ("IMAGE",)
|
|
FUNCTION: str = "f"
|
|
CATEGORY: str = "Data Analysis"
|
|
|
|
def f(self,
|
|
tens: torch.Tensor,
|
|
y_axis_dims: str,
|
|
title: str,
|
|
y_axis_label: str,
|
|
# legend_label: str,
|
|
bins: int,
|
|
style: str=None,
|
|
palette: str=None) -> tuple:
|
|
"""
|
|
Generates a histogram from a PyTorch tensor using Seaborn.
|
|
|
|
Args:
|
|
tens (torch.Tensor): The tensor.
|
|
y_axix_dims (str): The indices in tensor for the y-axis.
|
|
y_column_name (str): The column name for the y-axis.
|
|
title (str): The title of the plot.
|
|
y_axis_label (str): The label for the y-axis data.
|
|
bins (int): Number of bins.
|
|
style (str): Style of the plot
|
|
palette (str): Palette of the plot
|
|
Returns:
|
|
tuple: A tuple containing the image tensor representation of the plot.
|
|
"""
|
|
y_axes = ast.literal_eval(y_axis_dims)
|
|
if isinstance(y_axes, int):
|
|
y_axes = [y_axes]
|
|
|
|
if tens.dim() > 2:
|
|
raise ValueError("Only rank 1 or 2 tensors are supported.")
|
|
|
|
if tens.dim() == 1:
|
|
tens2 = torch.unsqueeze(tens, 1)
|
|
else:
|
|
tens2 = tens
|
|
|
|
labels = [""] * len(y_axes)
|
|
set_up_sns_style_palette(style, palette)
|
|
|
|
# Create the plot
|
|
fig, ax = plt.subplots()
|
|
for i in range(len(y_axes)):
|
|
sns.histplot(tens2[:, y_axes[i]].detach().cpu().numpy(),
|
|
bins=bins, ax=ax, label=labels[i])
|
|
break # TODO: Remove this if we decide support multiple plots
|
|
|
|
return plot_post_steps(
|
|
fig,
|
|
ax,
|
|
plt,
|
|
title,
|
|
None,
|
|
y_axis_label,
|
|
None) |